Copy disabled (too large)
Download .txt
Showing preview only (24,223K chars total). Download the full file to get everything.
Repository: bupticybee/ChineseAiDungeon
Branch: main
Commit: be9a8eb1e955
Files: 108
Total size: 78.0 MB
Directory structure:
gitextract_o5tgzk64/
├── 3rd/
│ └── gdown.pl/
│ ├── LICENSE.txt
│ ├── README.md
│ └── gdown.pl
├── CPM-Generate/
│ ├── .gitignore
│ ├── LICENSE
│ ├── README.md
│ ├── arguments.py
│ ├── bpe_3w_new/
│ │ ├── chinese_vocab.model
│ │ ├── chinese_vocab.vocab
│ │ ├── merges.txt
│ │ └── vocab.json
│ ├── configure_data.py
│ ├── data/
│ │ ├── Makefile
│ │ ├── __init__.py
│ │ ├── bert_dataset.py
│ │ ├── dataset_utils.py
│ │ ├── gpt2_dataset.py
│ │ ├── helpers.cpp
│ │ ├── indexed_dataset.py
│ │ ├── samplers.py
│ │ └── test/
│ │ ├── test_indexed_dataset.py
│ │ └── test_preprocess_data.sh
│ ├── data_utils/
│ │ ├── __init__.py
│ │ ├── corpora.py
│ │ ├── datasets.py
│ │ ├── file_utils.py
│ │ ├── lazy_loader.py
│ │ ├── samplers.py
│ │ ├── tf_dl.py
│ │ ├── tokenization.py
│ │ ├── tokenization_gpt2.py
│ │ └── wordpiece.py
│ ├── fp16/
│ │ ├── __init__.py
│ │ ├── fp16.py
│ │ ├── fp16util.py
│ │ └── loss_scaler.py
│ ├── generate_samples.py
│ ├── model/
│ │ ├── __init__.py
│ │ ├── distributed.py
│ │ ├── gpt2_modeling.py
│ │ ├── model.py
│ │ └── modeling.py
│ ├── mpu/
│ │ ├── __init__.py
│ │ ├── cross_entropy.py
│ │ ├── data.py
│ │ ├── grads.py
│ │ ├── initialize.py
│ │ ├── layers.py
│ │ ├── mappings.py
│ │ ├── random.py
│ │ ├── tests/
│ │ │ ├── __init__.py
│ │ │ ├── commons.py
│ │ │ ├── test_cross_entropy.py
│ │ │ ├── test_data.py
│ │ │ ├── test_initialize.py
│ │ │ ├── test_layers.py
│ │ │ └── test_random.py
│ │ ├── transformer.py
│ │ └── utils.py
│ ├── requirements.txt
│ ├── scripts/
│ │ └── generate_text.sh
│ └── utils.py
├── ChineseAiDungeonColabDemo.ipynb
├── LICENSE
├── finetune.ipynb
├── labeled_data/
│ ├── advanture_translated/
│ │ ├── process_data.ipynb
│ │ ├── processed_translated_story.txt
│ │ ├── processed_translated_story_valid.txt
│ │ ├── text_advanture_trans.txt
│ │ ├── train_adventures_valid_translated.txt
│ │ ├── truncated_advanture_train.json
│ │ ├── truncated_advanture_valid.json
│ │ └── valid_advantures_valid_translated.txt
│ ├── raw_data/
│ │ ├── Jinyong/
│ │ │ ├── 书剑恩仇录.txt
│ │ │ ├── 侠客行.txt
│ │ │ ├── 倚天屠龙记.txt
│ │ │ ├── 天龙八部.txt
│ │ │ ├── 射雕英雄传.txt
│ │ │ ├── 白马啸西风.txt
│ │ │ ├── 碧血剑.txt
│ │ │ ├── 神雕侠侣.txt
│ │ │ ├── 笑傲江湖.txt
│ │ │ ├── 越女剑.txt
│ │ │ ├── 连城诀.txt
│ │ │ ├── 雪山飞狐.txt
│ │ │ ├── 飞狐外传.txt
│ │ │ ├── 鸳鸯刀.txt
│ │ │ └── 鹿鼎记.txt
│ │ └── 邪气凛然.txt
│ ├── 一次冒险.txt
│ ├── 我从来都不主动.txt
│ ├── 盗墓实录.txt
│ └── 邪气凛然.txt
├── readme.md
├── requirements.txt
├── tf2gpt/
│ ├── .gitignore
│ ├── README.md
│ ├── loading.ipynb
│ ├── model.py
│ ├── predict.ipynb
│ ├── test.ipynb
│ ├── train.ipynb
│ └── vocab.txt
├── utils/
│ ├── gpt2_tokenizer.py
│ ├── progress_bar.py
│ ├── story_helper.py
│ └── story_util.py
└── 标注.ipynb
================================================
FILE CONTENTS
================================================
================================================
FILE: 3rd/gdown.pl/LICENSE.txt
================================================
GNU GENERAL PUBLIC LICENSE
Version 3, 29 June 2007
Copyright (C) 2007 Free Software Foundation, Inc. <http://fsf.org/>
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
Preamble
The GNU General Public License is a free, copyleft license for
software and other kinds of works.
The licenses for most software and other practical works are designed
to take away your freedom to share and change the works. By contrast,
the GNU General Public License is intended to guarantee your freedom to
share and change all versions of a program--to make sure it remains free
software for all its users. We, the Free Software Foundation, use the
GNU General Public License for most of our software; it applies also to
any other work released this way by its authors. You can apply it to
your programs, too.
When we speak of free software, we are referring to freedom, not
price. Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
them if you wish), that you receive source code or can get it if you
want it, that you can change the software or use pieces of it in new
free programs, and that you know you can do these things.
To protect your rights, we need to prevent others from denying you
these rights or asking you to surrender the rights. Therefore, you have
certain responsibilities if you distribute copies of the software, or if
you modify it: responsibilities to respect the freedom of others.
For example, if you distribute copies of such a program, whether
gratis or for a fee, you must pass on to the recipients the same
freedoms that you received. You must make sure that they, too, receive
or can get the source code. And you must show them these terms so they
know their rights.
Developers that use the GNU GPL protect your rights with two steps:
(1) assert copyright on the software, and (2) offer you this License
giving you legal permission to copy, distribute and/or modify it.
For the developers' and authors' protection, the GPL clearly explains
that there is no warranty for this free software. For both users' and
authors' sake, the GPL requires that modified versions be marked as
changed, so that their problems will not be attributed erroneously to
authors of previous versions.
Some devices are designed to deny users access to install or run
modified versions of the software inside them, although the manufacturer
can do so. This is fundamentally incompatible with the aim of
protecting users' freedom to change the software. The systematic
pattern of such abuse occurs in the area of products for individuals to
use, which is precisely where it is most unacceptable. Therefore, we
have designed this version of the GPL to prohibit the practice for those
products. If such problems arise substantially in other domains, we
stand ready to extend this provision to those domains in future versions
of the GPL, as needed to protect the freedom of users.
Finally, every program is threatened constantly by software patents.
States should not allow patents to restrict development and use of
software on general-purpose computers, but in those that do, we wish to
avoid the special danger that patents applied to a free program could
make it effectively proprietary. To prevent this, the GPL assures that
patents cannot be used to render the program non-free.
The precise terms and conditions for copying, distribution and
modification follow.
TERMS AND CONDITIONS
0. Definitions.
"This License" refers to version 3 of the GNU General Public License.
"Copyright" also means copyright-like laws that apply to other kinds of
works, such as semiconductor masks.
"The Program" refers to any copyrightable work licensed under this
License. Each licensee is addressed as "you". "Licensees" and
"recipients" may be individuals or organizations.
To "modify" a work means to copy from or adapt all or part of the work
in a fashion requiring copyright permission, other than the making of an
exact copy. The resulting work is called a "modified version" of the
earlier work or a work "based on" the earlier work.
A "covered work" means either the unmodified Program or a work based
on the Program.
To "propagate" a work means to do anything with it that, without
permission, would make you directly or secondarily liable for
infringement under applicable copyright law, except executing it on a
computer or modifying a private copy. Propagation includes copying,
distribution (with or without modification), making available to the
public, and in some countries other activities as well.
To "convey" a work means any kind of propagation that enables other
parties to make or receive copies. Mere interaction with a user through
a computer network, with no transfer of a copy, is not conveying.
An interactive user interface displays "Appropriate Legal Notices"
to the extent that it includes a convenient and prominently visible
feature that (1) displays an appropriate copyright notice, and (2)
tells the user that there is no warranty for the work (except to the
extent that warranties are provided), that licensees may convey the
work under this License, and how to view a copy of this License. If
the interface presents a list of user commands or options, such as a
menu, a prominent item in the list meets this criterion.
1. Source Code.
The "source code" for a work means the preferred form of the work
for making modifications to it. "Object code" means any non-source
form of a work.
A "Standard Interface" means an interface that either is an official
standard defined by a recognized standards body, or, in the case of
interfaces specified for a particular programming language, one that
is widely used among developers working in that language.
The "System Libraries" of an executable work include anything, other
than the work as a whole, that (a) is included in the normal form of
packaging a Major Component, but which is not part of that Major
Component, and (b) serves only to enable use of the work with that
Major Component, or to implement a Standard Interface for which an
implementation is available to the public in source code form. A
"Major Component", in this context, means a major essential component
(kernel, window system, and so on) of the specific operating system
(if any) on which the executable work runs, or a compiler used to
produce the work, or an object code interpreter used to run it.
The "Corresponding Source" for a work in object code form means all
the source code needed to generate, install, and (for an executable
work) run the object code and to modify the work, including scripts to
control those activities. However, it does not include the work's
System Libraries, or general-purpose tools or generally available free
programs which are used unmodified in performing those activities but
which are not part of the work. For example, Corresponding Source
includes interface definition files associated with source files for
the work, and the source code for shared libraries and dynamically
linked subprograms that the work is specifically designed to require,
such as by intimate data communication or control flow between those
subprograms and other parts of the work.
The Corresponding Source need not include anything that users
can regenerate automatically from other parts of the Corresponding
Source.
The Corresponding Source for a work in source code form is that
same work.
2. Basic Permissions.
All rights granted under this License are granted for the term of
copyright on the Program, and are irrevocable provided the stated
conditions are met. This License explicitly affirms your unlimited
permission to run the unmodified Program. The output from running a
covered work is covered by this License only if the output, given its
content, constitutes a covered work. This License acknowledges your
rights of fair use or other equivalent, as provided by copyright law.
You may make, run and propagate covered works that you do not
convey, without conditions so long as your license otherwise remains
in force. You may convey covered works to others for the sole purpose
of having them make modifications exclusively for you, or provide you
with facilities for running those works, provided that you comply with
the terms of this License in conveying all material for which you do
not control copyright. Those thus making or running the covered works
for you must do so exclusively on your behalf, under your direction
and control, on terms that prohibit them from making any copies of
your copyrighted material outside their relationship with you.
Conveying under any other circumstances is permitted solely under
the conditions stated below. Sublicensing is not allowed; section 10
makes it unnecessary.
3. Protecting Users' Legal Rights From Anti-Circumvention Law.
No covered work shall be deemed part of an effective technological
measure under any applicable law fulfilling obligations under article
11 of the WIPO copyright treaty adopted on 20 December 1996, or
similar laws prohibiting or restricting circumvention of such
measures.
When you convey a covered work, you waive any legal power to forbid
circumvention of technological measures to the extent such circumvention
is effected by exercising rights under this License with respect to
the covered work, and you disclaim any intention to limit operation or
modification of the work as a means of enforcing, against the work's
users, your or third parties' legal rights to forbid circumvention of
technological measures.
4. Conveying Verbatim Copies.
You may convey verbatim copies of the Program's source code as you
receive it, in any medium, provided that you conspicuously and
appropriately publish on each copy an appropriate copyright notice;
keep intact all notices stating that this License and any
non-permissive terms added in accord with section 7 apply to the code;
keep intact all notices of the absence of any warranty; and give all
recipients a copy of this License along with the Program.
You may charge any price or no price for each copy that you convey,
and you may offer support or warranty protection for a fee.
5. Conveying Modified Source Versions.
You may convey a work based on the Program, or the modifications to
produce it from the Program, in the form of source code under the
terms of section 4, provided that you also meet all of these conditions:
a) The work must carry prominent notices stating that you modified
it, and giving a relevant date.
b) The work must carry prominent notices stating that it is
released under this License and any conditions added under section
7. This requirement modifies the requirement in section 4 to
"keep intact all notices".
c) You must license the entire work, as a whole, under this
License to anyone who comes into possession of a copy. This
License will therefore apply, along with any applicable section 7
additional terms, to the whole of the work, and all its parts,
regardless of how they are packaged. This License gives no
permission to license the work in any other way, but it does not
invalidate such permission if you have separately received it.
d) If the work has interactive user interfaces, each must display
Appropriate Legal Notices; however, if the Program has interactive
interfaces that do not display Appropriate Legal Notices, your
work need not make them do so.
A compilation of a covered work with other separate and independent
works, which are not by their nature extensions of the covered work,
and which are not combined with it such as to form a larger program,
in or on a volume of a storage or distribution medium, is called an
"aggregate" if the compilation and its resulting copyright are not
used to limit the access or legal rights of the compilation's users
beyond what the individual works permit. Inclusion of a covered work
in an aggregate does not cause this License to apply to the other
parts of the aggregate.
6. Conveying Non-Source Forms.
You may convey a covered work in object code form under the terms
of sections 4 and 5, provided that you also convey the
machine-readable Corresponding Source under the terms of this License,
in one of these ways:
a) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by the
Corresponding Source fixed on a durable physical medium
customarily used for software interchange.
b) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by a
written offer, valid for at least three years and valid for as
long as you offer spare parts or customer support for that product
model, to give anyone who possesses the object code either (1) a
copy of the Corresponding Source for all the software in the
product that is covered by this License, on a durable physical
medium customarily used for software interchange, for a price no
more than your reasonable cost of physically performing this
conveying of source, or (2) access to copy the
Corresponding Source from a network server at no charge.
c) Convey individual copies of the object code with a copy of the
written offer to provide the Corresponding Source. This
alternative is allowed only occasionally and noncommercially, and
only if you received the object code with such an offer, in accord
with subsection 6b.
d) Convey the object code by offering access from a designated
place (gratis or for a charge), and offer equivalent access to the
Corresponding Source in the same way through the same place at no
further charge. You need not require recipients to copy the
Corresponding Source along with the object code. If the place to
copy the object code is a network server, the Corresponding Source
may be on a different server (operated by you or a third party)
that supports equivalent copying facilities, provided you maintain
clear directions next to the object code saying where to find the
Corresponding Source. Regardless of what server hosts the
Corresponding Source, you remain obligated to ensure that it is
available for as long as needed to satisfy these requirements.
e) Convey the object code using peer-to-peer transmission, provided
you inform other peers where the object code and Corresponding
Source of the work are being offered to the general public at no
charge under subsection 6d.
A separable portion of the object code, whose source code is excluded
from the Corresponding Source as a System Library, need not be
included in conveying the object code work.
A "User Product" is either (1) a "consumer product", which means any
tangible personal property which is normally used for personal, family,
or household purposes, or (2) anything designed or sold for incorporation
into a dwelling. In determining whether a product is a consumer product,
doubtful cases shall be resolved in favor of coverage. For a particular
product received by a particular user, "normally used" refers to a
typical or common use of that class of product, regardless of the status
of the particular user or of the way in which the particular user
actually uses, or expects or is expected to use, the product. A product
is a consumer product regardless of whether the product has substantial
commercial, industrial or non-consumer uses, unless such uses represent
the only significant mode of use of the product.
"Installation Information" for a User Product means any methods,
procedures, authorization keys, or other information required to install
and execute modified versions of a covered work in that User Product from
a modified version of its Corresponding Source. The information must
suffice to ensure that the continued functioning of the modified object
code is in no case prevented or interfered with solely because
modification has been made.
If you convey an object code work under this section in, or with, or
specifically for use in, a User Product, and the conveying occurs as
part of a transaction in which the right of possession and use of the
User Product is transferred to the recipient in perpetuity or for a
fixed term (regardless of how the transaction is characterized), the
Corresponding Source conveyed under this section must be accompanied
by the Installation Information. But this requirement does not apply
if neither you nor any third party retains the ability to install
modified object code on the User Product (for example, the work has
been installed in ROM).
The requirement to provide Installation Information does not include a
requirement to continue to provide support service, warranty, or updates
for a work that has been modified or installed by the recipient, or for
the User Product in which it has been modified or installed. Access to a
network may be denied when the modification itself materially and
adversely affects the operation of the network or violates the rules and
protocols for communication across the network.
Corresponding Source conveyed, and Installation Information provided,
in accord with this section must be in a format that is publicly
documented (and with an implementation available to the public in
source code form), and must require no special password or key for
unpacking, reading or copying.
7. Additional Terms.
"Additional permissions" are terms that supplement the terms of this
License by making exceptions from one or more of its conditions.
Additional permissions that are applicable to the entire Program shall
be treated as though they were included in this License, to the extent
that they are valid under applicable law. If additional permissions
apply only to part of the Program, that part may be used separately
under those permissions, but the entire Program remains governed by
this License without regard to the additional permissions.
When you convey a copy of a covered work, you may at your option
remove any additional permissions from that copy, or from any part of
it. (Additional permissions may be written to require their own
removal in certain cases when you modify the work.) You may place
additional permissions on material, added by you to a covered work,
for which you have or can give appropriate copyright permission.
Notwithstanding any other provision of this License, for material you
add to a covered work, you may (if authorized by the copyright holders of
that material) supplement the terms of this License with terms:
a) Disclaiming warranty or limiting liability differently from the
terms of sections 15 and 16 of this License; or
b) Requiring preservation of specified reasonable legal notices or
author attributions in that material or in the Appropriate Legal
Notices displayed by works containing it; or
c) Prohibiting misrepresentation of the origin of that material, or
requiring that modified versions of such material be marked in
reasonable ways as different from the original version; or
d) Limiting the use for publicity purposes of names of licensors or
authors of the material; or
e) Declining to grant rights under trademark law for use of some
trade names, trademarks, or service marks; or
f) Requiring indemnification of licensors and authors of that
material by anyone who conveys the material (or modified versions of
it) with contractual assumptions of liability to the recipient, for
any liability that these contractual assumptions directly impose on
those licensors and authors.
All other non-permissive additional terms are considered "further
restrictions" within the meaning of section 10. If the Program as you
received it, or any part of it, contains a notice stating that it is
governed by this License along with a term that is a further
restriction, you may remove that term. If a license document contains
a further restriction but permits relicensing or conveying under this
License, you may add to a covered work material governed by the terms
of that license document, provided that the further restriction does
not survive such relicensing or conveying.
If you add terms to a covered work in accord with this section, you
must place, in the relevant source files, a statement of the
additional terms that apply to those files, or a notice indicating
where to find the applicable terms.
Additional terms, permissive or non-permissive, may be stated in the
form of a separately written license, or stated as exceptions;
the above requirements apply either way.
8. Termination.
You may not propagate or modify a covered work except as expressly
provided under this License. Any attempt otherwise to propagate or
modify it is void, and will automatically terminate your rights under
this License (including any patent licenses granted under the third
paragraph of section 11).
However, if you cease all violation of this License, then your
license from a particular copyright holder is reinstated (a)
provisionally, unless and until the copyright holder explicitly and
finally terminates your license, and (b) permanently, if the copyright
holder fails to notify you of the violation by some reasonable means
prior to 60 days after the cessation.
Moreover, your license from a particular copyright holder is
reinstated permanently if the copyright holder notifies you of the
violation by some reasonable means, this is the first time you have
received notice of violation of this License (for any work) from that
copyright holder, and you cure the violation prior to 30 days after
your receipt of the notice.
Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License. If your rights have been terminated and not permanently
reinstated, you do not qualify to receive new licenses for the same
material under section 10.
9. Acceptance Not Required for Having Copies.
You are not required to accept this License in order to receive or
run a copy of the Program. Ancillary propagation of a covered work
occurring solely as a consequence of using peer-to-peer transmission
to receive a copy likewise does not require acceptance. However,
nothing other than this License grants you permission to propagate or
modify any covered work. These actions infringe copyright if you do
not accept this License. Therefore, by modifying or propagating a
covered work, you indicate your acceptance of this License to do so.
10. Automatic Licensing of Downstream Recipients.
Each time you convey a covered work, the recipient automatically
receives a license from the original licensors, to run, modify and
propagate that work, subject to this License. You are not responsible
for enforcing compliance by third parties with this License.
An "entity transaction" is a transaction transferring control of an
organization, or substantially all assets of one, or subdividing an
organization, or merging organizations. If propagation of a covered
work results from an entity transaction, each party to that
transaction who receives a copy of the work also receives whatever
licenses to the work the party's predecessor in interest had or could
give under the previous paragraph, plus a right to possession of the
Corresponding Source of the work from the predecessor in interest, if
the predecessor has it or can get it with reasonable efforts.
You may not impose any further restrictions on the exercise of the
rights granted or affirmed under this License. For example, you may
not impose a license fee, royalty, or other charge for exercise of
rights granted under this License, and you may not initiate litigation
(including a cross-claim or counterclaim in a lawsuit) alleging that
any patent claim is infringed by making, using, selling, offering for
sale, or importing the Program or any portion of it.
11. Patents.
A "contributor" is a copyright holder who authorizes use under this
License of the Program or a work on which the Program is based. The
work thus licensed is called the contributor's "contributor version".
A contributor's "essential patent claims" are all patent claims
owned or controlled by the contributor, whether already acquired or
hereafter acquired, that would be infringed by some manner, permitted
by this License, of making, using, or selling its contributor version,
but do not include claims that would be infringed only as a
consequence of further modification of the contributor version. For
purposes of this definition, "control" includes the right to grant
patent sublicenses in a manner consistent with the requirements of
this License.
Each contributor grants you a non-exclusive, worldwide, royalty-free
patent license under the contributor's essential patent claims, to
make, use, sell, offer for sale, import and otherwise run, modify and
propagate the contents of its contributor version.
In the following three paragraphs, a "patent license" is any express
agreement or commitment, however denominated, not to enforce a patent
(such as an express permission to practice a patent or covenant not to
sue for patent infringement). To "grant" such a patent license to a
party means to make such an agreement or commitment not to enforce a
patent against the party.
If you convey a covered work, knowingly relying on a patent license,
and the Corresponding Source of the work is not available for anyone
to copy, free of charge and under the terms of this License, through a
publicly available network server or other readily accessible means,
then you must either (1) cause the Corresponding Source to be so
available, or (2) arrange to deprive yourself of the benefit of the
patent license for this particular work, or (3) arrange, in a manner
consistent with the requirements of this License, to extend the patent
license to downstream recipients. "Knowingly relying" means you have
actual knowledge that, but for the patent license, your conveying the
covered work in a country, or your recipient's use of the covered work
in a country, would infringe one or more identifiable patents in that
country that you have reason to believe are valid.
If, pursuant to or in connection with a single transaction or
arrangement, you convey, or propagate by procuring conveyance of, a
covered work, and grant a patent license to some of the parties
receiving the covered work authorizing them to use, propagate, modify
or convey a specific copy of the covered work, then the patent license
you grant is automatically extended to all recipients of the covered
work and works based on it.
A patent license is "discriminatory" if it does not include within
the scope of its coverage, prohibits the exercise of, or is
conditioned on the non-exercise of one or more of the rights that are
specifically granted under this License. You may not convey a covered
work if you are a party to an arrangement with a third party that is
in the business of distributing software, under which you make payment
to the third party based on the extent of your activity of conveying
the work, and under which the third party grants, to any of the
parties who would receive the covered work from you, a discriminatory
patent license (a) in connection with copies of the covered work
conveyed by you (or copies made from those copies), or (b) primarily
for and in connection with specific products or compilations that
contain the covered work, unless you entered into that arrangement,
or that patent license was granted, prior to 28 March 2007.
Nothing in this License shall be construed as excluding or limiting
any implied license or other defenses to infringement that may
otherwise be available to you under applicable patent law.
12. No Surrender of Others' Freedom.
If conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License. If you cannot convey a
covered work so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you may
not convey it at all. For example, if you agree to terms that obligate you
to collect a royalty for further conveying from those to whom you convey
the Program, the only way you could satisfy both those terms and this
License would be to refrain entirely from conveying the Program.
13. Use with the GNU Affero General Public License.
Notwithstanding any other provision of this License, you have
permission to link or combine any covered work with a work licensed
under version 3 of the GNU Affero General Public License into a single
combined work, and to convey the resulting work. The terms of this
License will continue to apply to the part which is the covered work,
but the special requirements of the GNU Affero General Public License,
section 13, concerning interaction through a network will apply to the
combination as such.
14. Revised Versions of this License.
The Free Software Foundation may publish revised and/or new versions of
the GNU General Public License from time to time. Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.
Each version is given a distinguishing version number. If the
Program specifies that a certain numbered version of the GNU General
Public License "or any later version" applies to it, you have the
option of following the terms and conditions either of that numbered
version or of any later version published by the Free Software
Foundation. If the Program does not specify a version number of the
GNU General Public License, you may choose any version ever published
by the Free Software Foundation.
If the Program specifies that a proxy can decide which future
versions of the GNU General Public License can be used, that proxy's
public statement of acceptance of a version permanently authorizes you
to choose that version for the Program.
Later license versions may give you additional or different
permissions. However, no additional obligations are imposed on any
author or copyright holder as a result of your choosing to follow a
later version.
15. Disclaimer of Warranty.
THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
16. Limitation of Liability.
IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
SUCH DAMAGES.
17. Interpretation of Sections 15 and 16.
If the disclaimer of warranty and limitation of liability provided
above cannot be given local legal effect according to their terms,
reviewing courts shall apply local law that most closely approximates
an absolute waiver of all civil liability in connection with the
Program, unless a warranty or assumption of liability accompanies a
copy of the Program in return for a fee.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Programs
If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest
to attach them to the start of each source file to most effectively
state the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.
<one line to give the program's name and a brief idea of what it does.>
Copyright (C) <year> <name of author>
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
Also add information on how to contact you by electronic and paper mail.
If the program does terminal interaction, make it output a short
notice like this when it starts in an interactive mode:
<program> Copyright (C) <year> <name of author>
This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
This is free software, and you are welcome to redistribute it
under certain conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the appropriate
parts of the General Public License. Of course, your program's commands
might be different; for a GUI interface, you would use an "about box".
You should also get your employer (if you work as a programmer) or school,
if any, to sign a "copyright disclaimer" for the program, if necessary.
For more information on this, and how to apply and follow the GNU GPL, see
<http://www.gnu.org/licenses/>.
The GNU General Public License does not permit incorporating your program
into proprietary programs. If your program is a subroutine library, you
may consider it more useful to permit linking proprietary applications with
the library. If this is what you want to do, use the GNU Lesser General
Public License instead of this License. But first, please read
<http://www.gnu.org/philosophy/why-not-lgpl.html>.
================================================
FILE: 3rd/gdown.pl/README.md
================================================
gdown.pl
========
Google Drive direct download of big files
Requirements
============
*wget* and *Perl* must be in the PATH.
**Windows** and **linux** compatible.
Usage
=====
Use Google Drive shareable links, viewable by anyone:
$ ./gdown.pl 'gdrive file url' ['desired file name']
Example
=======
For example, to download [this video](https://drive.google.com/file/d/0B1L_hFrWJfRhLUJZdXdSdTdfSWs/edit) from my [axolotl project](https://circulosmeos.wordpress.com/2015/03/04/axolotl-a-simple-plain-text-documentation-system/), just copy the url, and give a file name if desired:
$ ./gdown.pl https://drive.google.com/file/d/0B1L_hFrWJfRhLUJZdXdSdTdfSWs/edit axolotl.mp4
Resuming a download
===================
If you need to resume a download, please, use [**gdown.pl v2.0** here](https://github.com/circulosmeos/gdown.pl/tree/with-resume).
As long as a file name is indicated as second parameter, *gdown.pl v2.0* **will try to resume the partially downloaded file** if a local incomplete file with that name already exists.
Version
=======
This version is **v1.4**.
### Warning
Please, note that v1.2 (available between days 12 to 31 of Jan/2019) **should not be used**, as it contains a bug that could result in unusable downloaded files. Proceed to overwrite with v1.3 in case you have it.
Docker
======
A simple Docker file is provided, to build a simple Docker image with gdown.pl.
This has been used for pre-pulling data from a Google Drive to Kubernetes persistent volumes. Thanks @anton-khodak
License
=======
Distributed [under GPL 3](http://www.gnu.org/licenses/gpl-3.0.html)
Disclaimer
==========
This software is provided "as is", without warranty of any kind, express or implied.
More info
=========
[https://circulosmeos.wordpress.com/2014/04/12/google-drive-direct-download-of-big-files](https://circulosmeos.wordpress.com/2014/04/12/google-drive-direct-download-of-big-files)
Contact
=======
by [circulosmeos](loopidle@gmail.com)
================================================
FILE: 3rd/gdown.pl/gdown.pl
================================================
#!/usr/bin/env perl
#
# Google Drive direct download of big files
# ./gdown.pl 'gdrive file url' ['desired file name']
#
# v1.0 by circulosmeos 04-2014.
# v1.1 by circulosmeos 01-2017.
# v1.2, 2.0 by circulosmeos 01-2019.
# v2.1 by circulosmeos 12-2020.
# //circulosmeos.wordpress.com/2014/04/12/google-drive-direct-download-of-big-files
# Distributed under GPL 3 (//www.gnu.org/licenses/gpl-3.0.html)
#
use strict;
use POSIX;
my $TEMP='gdown.cookie.temp';
my $COMMAND;
my $confirm;
my $check;
sub execute_command();
my $URL=shift;
die "\n./gdown.pl 'gdrive file url' [desired file name]\n\n" if $URL eq '';
my $FILENAME=shift;
my $TEMP_FILENAME='gdown.'.strftime("%Y%m%d%H%M%S", localtime).'.'.substr(rand,2);
if ($URL=~m#^https?://drive.google.com/file/d/([^/]+)#) {
$URL="https://docs.google.com/uc?id=$1&export=download";
}
elsif ($URL=~m#^https?://drive.google.com/open\?id=([^/]+)#) {
$URL="https://docs.google.com/uc?id=$1&export=download";
}
execute_command();
while (-s $TEMP_FILENAME < 100000) { # only if the file isn't the download yet
open fFILENAME, '<', $TEMP_FILENAME;
$check=0;
foreach (<fFILENAME>) {
if (/href="(\/uc\?export=download[^"]+)/) {
$URL='https://docs.google.com'.$1;
$URL=~s/&/&/g;
$confirm='';
$check=1;
last;
}
if (/confirm=([^;&]+)/) {
$confirm=$1;
$check=1;
last;
}
if (/"downloadUrl":"([^"]+)/) {
$URL=$1;
$URL=~s/\\u003d/=/g;
$URL=~s/\\u0026/&/g;
$confirm='';
$check=1;
last;
}
}
close fFILENAME;
die "Couldn't download the file :-(\n" if ($check==0);
$URL=~s/confirm=([^;&]+)/confirm=$confirm/ if $confirm ne '';
execute_command();
}
unlink $TEMP;
sub execute_command() {
my $OUTPUT_FILENAME = $TEMP_FILENAME;
my $CONTINUE = '';
# check contents before download & if a $FILENAME has been indicated resume on content download
# please, note that for this to work, wget must correctly provide --spider with --server-response (-S)
if ( length($FILENAME) > 0 ) {
$COMMAND="wget -q -S --no-check-certificate --spider --load-cookie $TEMP --save-cookie $TEMP \"$URL\" 2>&1";
my @HEADERS=`$COMMAND`;
foreach my $header (@HEADERS) {
if ( ( $header =~ /Content-Type: (.+)/ && $1 !~ 'text/html' ) ||
$header =~ 'HTTP/1.1 405 Method Not Allowed'
) {
$OUTPUT_FILENAME = $FILENAME;
$CONTINUE = '-c';
last;
}
}
}
$COMMAND="wget $CONTINUE --progress=dot:giga --no-check-certificate --load-cookie $TEMP --save-cookie $TEMP \"$URL\"";
$COMMAND.=" -O \"$OUTPUT_FILENAME\"";
my $OUTPUT = system( $COMMAND );
if ( $OUTPUT == 2 ) { # do a clean exit with Ctrl+C
unlink $TEMP;
die "\nDownloading interrupted by user\n\n";
} elsif ( $OUTPUT == 0 && length($CONTINUE)>0 ) { # do a clean exit with $FILENAME provided
unlink $TEMP;
die "\nDownloading complete\n\n";
}
return 1;
}
================================================
FILE: CPM-Generate/.gitignore
================================================
# Byte-compiled / optimized / DLL files
__pycache__/
*.py[cod]
*$py.class
# C extensions
*.so
# Distribution / packaging
.Python
build/
develop-eggs/
dist/
downloads/
eggs/
.eggs/
lib/
lib64/
parts/
sdist/
var/
wheels/
pip-wheel-metadata/
share/python-wheels/
*.egg-info/
.installed.cfg
*.egg
MANIFEST
# PyInstaller
# Usually these files are written by a python script from a template
# before PyInstaller builds the exe, so as to inject date/other infos into it.
*.manifest
*.spec
# Installer logs
pip-log.txt
pip-delete-this-directory.txt
# Unit test / coverage reports
htmlcov/
.tox/
.nox/
.coverage
.coverage.*
.cache
nosetests.xml
coverage.xml
*.cover
*.py,cover
.hypothesis/
.pytest_cache/
# Translations
*.mo
*.pot
# Django stuff:
*.log
local_settings.py
db.sqlite3
db.sqlite3-journal
# Flask stuff:
instance/
.webassets-cache
# Scrapy stuff:
.scrapy
# Sphinx documentation
docs/_build/
# PyBuilder
target/
# Jupyter Notebook
.ipynb_checkpoints
# IPython
profile_default/
ipython_config.py
# pyenv
.python-version
# pipenv
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
# However, in case of collaboration, if having platform-specific dependencies or dependencies
# having no cross-platform support, pipenv may install dependencies that don't work, or not
# install all needed dependencies.
#Pipfile.lock
# PEP 582; used by e.g. github.com/David-OConnor/pyflow
__pypackages__/
# Celery stuff
celerybeat-schedule
celerybeat.pid
# SageMath parsed files
*.sage.py
# Environments
.env
.venv
env/
venv/
ENV/
env.bak/
venv.bak/
# Spyder project settings
.spyderproject
.spyproject
# Rope project settings
.ropeproject
# mkdocs documentation
/site
# mypy
.mypy_cache/
.dmypy.json
dmypy.json
# Pyre type checker
.pyre/
================================================
FILE: CPM-Generate/LICENSE
================================================
MIT License
Copyright (c) 2020 THU-PLM
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
================================================
FILE: CPM-Generate/README.md
================================================
# CPM-Generate
为了促进中文自然语言处理研究的发展，本项目提供了 **CPM-LM** (2.6B) 模型的文本生成代码，可用于文本生成的本地测试，并以此为基础进一步研究零次学习/少次学习等场景。[[项目首页](https://cpm.baai.ac.cn)] [[模型下载](https://cpm.baai.ac.cn/download.html)]
## 安装
首先安装pytorch等基础依赖，再安装[APEX](https://github.com/NVIDIA/apex#quick-start)以支持fp16：
```
pip install -r requirements.txt
git clone https://github.com/NVIDIA/apex
cd apex
pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" ./
```
## 模型
模型下载后文件夹的目录结构需设置如下：
```
.
├── 80000
│ ├── mp_rank_00_model_states.pt
│ └── mp_rank_01_model_states.pt
└── latest_checkpointed_iteration.txt
```
为保证下载文件的正确性，文件的checksum如下：
```
SHA1
7a8461098dae374a836f55bf2002bf21cba05964 80000/mp_rank_00_model_states.pt
6ab97bd06979c1707080b33cfddff4f08c9af675 80000/mp_rank_01_model_states.pt
MD5
ffb39d8e477aedf9f8714cc467f34eae 80000/mp_rank_00_model_states.pt
8af02c1aa85e76a430d45f3ad073edc5 80000/mp_rank_01_model_states.pt
```
## 使用
提供了命令行交互式生成：
```
bash scripts/generate_text.sh /path/to/CPM
```
运行该脚本需要两块GPU，每张卡的GPU内存占用约为7GB。该项目主要基于 [Megatron-LM](https://github.com/NVIDIA/Megatron-LM) 进行修改。
## 引用
```
@article{cpm-v1,
title={CPM: A Large-scale Generative Chinese Pre-trained Language Model},
author={Zhang, Zhengyan and Han, Xu, and Zhou, Hao, and Ke, Pei, and Gu, Yuxian and Ye, Deming and Qin, Yujia and Su, Yusheng and Ji, Haozhe and Guan, Jian and Qi, Fanchao and Wang, Xiaozhi and Zheng, Yanan and Cao, Jiannan and Zeng, Guoyang and Cao, Huanqi and Chen, Shengqi and Li, Daixuan and Sun, Zhenbo and Liu, Zhiyuan and Huang, Minlie and Han, Wentao and Tang, Jie and Li, Juanzi and Sun, Maosong},
year={2020}
}
```
================================================
FILE: CPM-Generate/arguments.py
================================================
# coding=utf-8
# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""argparser configuration"""
import argparse
import os
import torch
def add_model_config_args(parser):
"""Model arguments"""
group = parser.add_argument_group('model', 'model configuration')
group.add_argument('--pretrained-bert', action='store_true',
help='use a pretrained bert-large-uncased model instead'
'of initializing from scratch. See '
'--tokenizer-model-type to specify which pretrained '
'BERT model to use')
group.add_argument('--attention-dropout', type=float, default=0.1,
help='dropout probability for attention weights')
group.add_argument('--num-attention-heads', type=int, default=16,
help='num of transformer attention heads')
group.add_argument('--hidden-size', type=int, default=1024,
help='tansformer hidden size')
group.add_argument('--intermediate-size', type=int, default=None,
help='transformer embedding dimension for FFN'
'set to 4*`--hidden-size` if it is None')
group.add_argument('--num-layers', type=int, default=24,
help='num decoder layers')
group.add_argument('--layernorm-epsilon', type=float, default=1e-5,
help='layer norm epsilon')
group.add_argument('--hidden-dropout', type=float, default=0.1,
help='dropout probability for hidden state transformer')
group.add_argument('--max-position-embeddings', type=int, default=512,
help='maximum number of position embeddings to use')
group.add_argument('--vocab-size', type=int, default=30522,
help='vocab size to use for non-character-level '
'tokenization. This value will only be used when '
'creating a tokenizer')
group.add_argument('--deep-init', action='store_true',
help='initialize bert model similar to gpt2 model.'
'scales initialization of projection layers by a '
'factor of 1/sqrt(2N). Necessary to train bert '
'models larger than BERT-Large.')
group.add_argument('--make-vocab-size-divisible-by', type=int, default=128,
help='Pad the vocab size to be divisible by this value.'
'This is added for computational efficieny reasons.')
group.add_argument('--cpu-optimizer', action='store_true',
help='Run optimizer on CPU')
group.add_argument('--cpu_torch_adam', action='store_true',
help='Use Torch Adam as optimizer on CPU.')
return parser
def add_fp16_config_args(parser):
"""Mixed precision arguments."""
group = parser.add_argument_group('fp16', 'fp16 configurations')
group.add_argument('--fp16', action='store_true',
help='Run model in fp16 mode')
group.add_argument('--fp32-embedding', action='store_true',
help='embedding in fp32')
group.add_argument('--fp32-layernorm', action='store_true',
help='layer norm in fp32')
group.add_argument('--fp32-tokentypes', action='store_true',
help='embedding token types in fp32')
group.add_argument('--fp32-allreduce', action='store_true',
help='all-reduce in fp32')
group.add_argument('--hysteresis', type=int, default=2,
help='hysteresis for dynamic loss scaling')
group.add_argument('--loss-scale', type=float, default=None,
help='Static loss scaling, positive power of 2 '
'values can improve fp16 convergence. If None, dynamic'
'loss scaling is used.')
group.add_argument('--loss-scale-window', type=float, default=1000,
help='Window over which to raise/lower dynamic scale')
group.add_argument('--min-scale', type=float, default=1,
help='Minimum loss scale for dynamic loss scale')
return parser
def add_training_args(parser):
"""Training arguments."""
group = parser.add_argument_group('train', 'training configurations')
group.add_argument('--batch-size', type=int, default=4,
help='Data Loader batch size')
group.add_argument('--weight-decay', type=float, default=0.01,
help='weight decay coefficient for L2 regularization')
group.add_argument('--checkpoint-activations', action='store_true',
help='checkpoint activation to allow for training '
'with larger models and sequences')
group.add_argument('--checkpoint-num-layers', type=int, default=1,
help='chunk size (number of layers) for checkpointing')
group.add_argument('--clip-grad', type=float, default=1.0,
help='gradient clipping')
group.add_argument('--train-iters', type=int, default=1000000,
help='total number of iterations to train over all training runs')
group.add_argument('--log-interval', type=int, default=100,
help='report interval')
group.add_argument('--exit-interval', type=int, default=None,
help='Exit the program after this many new iterations.')
group.add_argument('--seed', type=int, default=1234,
help='random seed')
# Batch prodecuer arguments
group.add_argument('--reset-position-ids', action='store_true',
help='Reset posistion ids after end-of-document token.')
group.add_argument('--reset-attention-mask', action='store_true',
help='Reset self attention maske after '
'end-of-document token.')
# Learning rate.
group.add_argument('--lr-decay-iters', type=int, default=None,
help='number of iterations to decay LR over,'
' If None defaults to `--train-iters`*`--epochs`')
group.add_argument('--lr-decay-style', type=str, default='linear',
choices=['constant', 'linear', 'cosine', 'exponential'],
help='learning rate decay function')
group.add_argument('--lr', type=float, default=1.0e-4,
help='initial learning rate')
group.add_argument('--warmup', type=float, default=0.01,
help='percentage of data to warmup on (.01 = 1% of all '
'training iters). Default 0.01')
# model checkpointing
group.add_argument('--save', type=str, default=None,
help='Output directory to save checkpoints to.')
group.add_argument('--save-interval', type=int, default=5000,
help='number of iterations between saves')
group.add_argument('--no-save-optim', action='store_true',
help='Do not save current optimizer.')
group.add_argument('--no-save-rng', action='store_true',
help='Do not save current rng state.')
group.add_argument('--load', type=str, default=None,
help='Path to a directory containing a model checkpoint.')
group.add_argument('--no-load-optim', action='store_true',
help='Do not load optimizer when loading checkpoint.')
group.add_argument('--no-load-rng', action='store_true',
help='Do not load rng state when loading checkpoint.')
group.add_argument('--finetune', action='store_true',
help='Load model for finetuning. Do not load optimizer '
'or rng state from checkpoint and set iteration to 0. '
'Assumed when loading a release checkpoint.')
group.add_argument('--resume-dataloader', action='store_true',
help='Resume the dataloader when resuming training. '
'Does not apply to tfrecords dataloader, try resuming'
'with a different seed in this case.')
# distributed training args
group.add_argument('--distributed-backend', default='nccl',
help='which backend to use for distributed '
'training. One of [gloo, nccl]')
group.add_argument('--local_rank', type=int, default=None,
help='local rank passed from distributed launcher')
return parser
def add_evaluation_args(parser):
"""Evaluation arguments."""
group = parser.add_argument_group('validation', 'validation configurations')
group.add_argument('--eval-batch-size', type=int, default=None,
help='Data Loader batch size for evaluation datasets.'
'Defaults to `--batch-size`')
group.add_argument('--eval-iters', type=int, default=100,
help='number of iterations to run for evaluation'
'validation/test for')
group.add_argument('--eval-interval', type=int, default=1000,
help='interval between running evaluation on validation set')
group.add_argument('--eval-seq-length', type=int, default=None,
help='Maximum sequence length to process for '
'evaluation. Defaults to `--seq-length`')
group.add_argument('--eval-max-preds-per-seq', type=int, default=None,
help='Maximum number of predictions to use for '
'evaluation. Defaults to '
'math.ceil(`--eval-seq-length`*.15/10)*10')
group.add_argument('--overlapping-eval', type=int, default=32,
help='sliding window for overlapping eval ')
group.add_argument('--cloze-eval', action='store_true',
help='Evaluation dataset from `--valid-data` is a cloze task')
group.add_argument('--eval-hf', action='store_true',
help='perform evaluation with huggingface openai model.'
'use `--load` to specify weights path to be loaded')
group.add_argument('--load-openai', action='store_true',
help='load openai weights into our model. Use `--load` '
'to specify weights path to be loaded')
return parser
def add_text_generate_args(parser):
"""Text generate arguments."""
group = parser.add_argument_group('Text generation', 'configurations')
group.add_argument("--temperature", type=float, default=1.0)
group.add_argument("--top_p", type=float, default=0.0)
group.add_argument("--top_k", type=int, default=0)
group.add_argument("--out-seq-length", type=int, default=256)
return parser
def add_data_args(parser):
"""Train/valid/test data arguments."""
group = parser.add_argument_group('data', 'data configurations')
group.add_argument('--data-impl', type=str, default='infer',
choices=['lazy', 'cached', 'mmap', 'infer'],
help='Implementation of indexed datasets.')
group.add_argument('--mmap-warmup', action='store_true',
help='Warm up mmap files.')
group.add_argument('--model-parallel-size', type=int, default=1,
help='size of the model parallel.')
group.add_argument('--shuffle', action='store_true',
help='Shuffle data. Shuffling is deterministic '
'based on seed and current epoch.')
#group.add_argument('--train-data', nargs='+', default=None,
# help='Whitespace separated filenames or corpora names '
# 'for training.')
group.add_argument('--data-path', type=str, default=None,
help='Path to combined dataset to split.')
group.add_argument('--use-npy-data-loader', action='store_true',
help='Use the numpy data loader. If set, then'
'train-data-path, val-data-path, and test-data-path'
'should also be provided.')
group.add_argument('--train-data-path', type=str, default='',
help='path to the training data')
group.add_argument('--val-data-path', type=str, default='',
help='path to the validation data')
group.add_argument('--test-data-path', type=str, default='',
help='path to the test data')
group.add_argument('--input-data-sizes-file', type=str, default='sizes.txt',
help='the filename containing all the shards sizes')
group.add_argument('--delim', default=',',
help='delimiter used to parse csv data files')
group.add_argument('--text-key', default='sentence',
help='key to use to extract text from json/csv')
group.add_argument('--eval-text-key', default=None,
help='key to use to extract text from '
'json/csv evaluation datasets')
group.add_argument('--valid-data', nargs='*', default=None,
help="""Filename for validation data.""")
group.add_argument('--split', default='1000,1,1',
help='comma-separated list of proportions for training,'
' validation, and test split')
group.add_argument('--test-data', nargs='*', default=None,
help="""Filename for testing""")
group.add_argument('--lazy-loader', action='store_true',
help='whether to lazy read the data set')
group.add_argument('--loose-json', action='store_true',
help='Use loose json (one json-formatted string per '
'newline), instead of tight json (data file is one '
'json string)')
group.add_argument('--presplit-sentences', action='store_true',
help='Dataset content consists of documents where '
'each document consists of newline separated sentences')
group.add_argument('--num-workers', type=int, default=2,
help="""Number of workers to use for dataloading""")
group.add_argument('--tokenizer-model-type', type=str,
default='bert-large-uncased',
help="Model type to use for sentencepiece tokenization \
(one of ['bpe', 'char', 'unigram', 'word']) or \
bert vocab to use for BertWordPieceTokenizer (one of \
['bert-large-uncased', 'bert-large-cased', etc.])")
group.add_argument('--tokenizer-path', type=str, default='tokenizer.model',
help='path used to save/load sentencepiece tokenization '
'models')
group.add_argument('--tokenizer-type', type=str,
default='BertWordPieceTokenizer',
choices=['CharacterLevelTokenizer',
'SentencePieceTokenizer',
'BertWordPieceTokenizer',
'GPT2BPETokenizer'],
help='what type of tokenizer to use')
group.add_argument("--cache-dir", default=None, type=str,
help="Where to store pre-trained BERT downloads")
group.add_argument('--use-tfrecords', action='store_true',
help='load `--train-data`, `--valid-data`, '
'`--test-data` from BERT tf records instead of '
'normal data pipeline')
group.add_argument('--seq-length', type=int, default=512,
help="Maximum sequence length to process")
group.add_argument('--max-preds-per-seq', type=int, default=None,
help='Maximum number of predictions to use per sequence.'
'Defaults to math.ceil(`--seq-length`*.15/10)*10.'
'MUST BE SPECIFIED IF `--use-tfrecords` is True.')
return parser
def get_args():
"""Parse all the args."""
parser = argparse.ArgumentParser(description='PyTorch BERT Model')
parser = add_model_config_args(parser)
parser = add_fp16_config_args(parser)
parser = add_training_args(parser)
parser = add_evaluation_args(parser)
parser = add_text_generate_args(parser)
parser = add_data_args(parser)
args = parser.parse_args()
if not args.data_path and not args.train_data_path:
print('WARNING: No training data specified')
args.cuda = torch.cuda.is_available()
args.rank = int(os.getenv('RANK', '0'))
args.world_size = int(os.getenv("WORLD_SIZE", '1'))
if os.getenv('OMPI_COMM_WORLD_LOCAL_RANK'):
# We are using (OpenMPI) mpirun for launching distributed data parallel processes
local_rank = int(os.getenv('OMPI_COMM_WORLD_LOCAL_RANK'))
local_size = int(os.getenv('OMPI_COMM_WORLD_LOCAL_SIZE'))
# Possibly running with Slurm
num_nodes = int(os.getenv('SLURM_JOB_NUM_NODES', '1'))
nodeid = int(os.getenv('SLURM_NODEID', '0'))
args.local_rank = local_rank
args.rank = nodeid*local_size + local_rank
args.world_size = num_nodes*local_size
args.model_parallel_size = min(args.model_parallel_size, args.world_size)
if args.rank == 0:
print('using world size: {} and model-parallel size: {} '.format(
args.world_size, args.model_parallel_size))
args.dynamic_loss_scale = False
if args.loss_scale is None:
args.dynamic_loss_scale = True
if args.rank == 0:
print(' > using dynamic loss scaling')
# The args fp32_* or fp16_* meant to be active when the
# args fp16 is set. So the default behaviour should all
# be false.
if not args.fp16:
args.fp32_embedding = False
args.fp32_tokentypes = False
args.fp32_layernorm = False
return args
================================================
FILE: CPM-Generate/bpe_3w_new/chinese_vocab.vocab
================================================
<unk> 0
<s> 0
</s> 0
<cls> 0
<sep> 0
<pad> 0
<mask> 0
<eod> 0
▁ -1.72313
, -3.12488
▂ -3.39822
▁的 -3.48588
。 -3.96942
、 -4.75109
▁了 -4.9576
▁是 -4.96831
▁我 -5.03877
: -5.27007
▁在 -5.28669
▁有 -5.47599
▁? -5.52095
▁和 -5.54784
) -5.59145
▁( -5.60208
▁不 -5.68255
▁这 -5.78432
▁也 -5.78973
▁你 -5.80415
▁他 -5.94112
▁就 -5.94249
▁都 -5.9429
▁人 -5.94672
“ -6.02459
个 -6.02884
” -6.03092
是 -6.03102
▁. -6.05749
▁为 -6.16243
▁年 -6.20417
▁对 -6.21288
▁一个 -6.23571
— -6.28548
▁上 -6.33959
▁说 -6.35651
▁一 -6.36115
▁! -6.36247
▁很 -6.36465
; -6.38016
▁- -6.38203
▁看 -6.40957
》 -6.41369
▁但 -6.41523
《 -6.41608
▁1 -6.42653
▁自己 -6.43153
▁中 -6.46683
▁与 -6.47977
▁还 -6.48873
▁... -6.52697
▁而 -6.55111
▁之 -6.55349
▁会 -6.55881
人 -6.5685
▁好 -6.58159
▁可以 -6.58698
于 -6.63929
/ -6.66273
▁多 -6.67332
到 -6.68012
▁大 -6.69993
种 -6.71742
▁要 -6.72093
▁等 -6.72914
有 -6.7353
为 -6.74028
▁2 -6.74393
▁被 -6.74523
▁没有 -6.74655
▁什么 -6.74768
们 -6.79986
▁能 -6.80145
▁到 -6.82032
▁那 -6.82276
▁做 -6.82435
▁想 -6.85469
▁最 -6.86205
▁她 -6.86774
▁只 -6.87663
▁就是 -6.88267
子 -6.89051
道 -6.89532
▁去 -6.90234
来 -6.9327
▁3 -6.93732
▁问题 -6.9569
▁月 -6.96071
= -6.96547
▁并 -6.98468
▁中国 -7.00578
▁下 -7.00865
性 -7.01976
▁我们 -7.0325
▁小 -7.05666
▁当 -7.06192
▁以 -7.0644
▁更 -7.0952
▁没 -7.10579
▁让 -7.11392
▁用 -7.12722
▁后 -7.13505
因为 -7.1643
▁工作 -7.16739
▁可 -7.18634
家 -7.19207
▁从 -7.19224
▁如果 -7.1993
▁给 -7.20088
▁又 -7.20542
生 -7.20692
▁时候 -7.21321
还是 -7.22769
▁知 -7.22802
▁觉得 -7.23545
然 -7.24014
▁不是 -7.24724
者 -7.26277
▁把 -7.27058
会 -7.2716
▁4 -7.2789
▁现在 -7.28441
▁过 -7.31155
▁或 -7.31632
章 -7.3231
▁吧 -7.32894
什么 -7.33723
▁已 -7.33861
经 -7.35717
要 -7.35861
▁时 -7.36604
了 -7.3735
▁怎么 -7.3784
▁出 -7.38278
法 -7.38746
▁喜 -7.38958
好 -7.39481
上 -7.39812
点 -7.40685
▁5 -7.40774
▁所以 -7.4253
▁啊 -7.43476
▁吗 -7.44093
欢 -7.44564
后 -7.44802
▁高 -7.4532
里 -7.45956
▁本 -7.46725
话 -7.4695
▁着 -7.48661
▁很多 -7.48834
▁应 -7.48988
且 -7.49306
▁来 -7.49625
▁三 -7.50099
▁这样 -7.50728
▁地 -7.51797
▁将 -7.53163
大 -7.53241
▁如何 -7.53331
▁发展 -7.53855
出 -7.54207
▁新 -7.5506
▁再 -7.55146
▁及 -7.55461
· -7.57562
▁可能 -7.58248
▁由 -7.58621
该 -7.5876
面 -7.59767
时 -7.60193
▁日 -7.61322
▁总 -7.62151
▁每 -7.62319
次 -7.6245
] -7.62615
[ -7.62748
式 -7.62814
长 -7.64121
▁第 -7.66112
外 -7.66374
▁公司 -7.66769
▁第一 -7.66803
▁管理 -7.67631
▁然后 -7.67711
▁时间 -7.68062
区 -7.68119
▁开始 -7.69191
▁其 -7.69403
着 -7.6947
▁10 -7.69805
化 -7.70869
▁呢 -7.70909
些 -7.71201
▁6 -7.72152
▁需要 -7.72571
▁生活 -7.7294
间 -7.72968
▁或者 -7.73142
▁跟 -7.7316
▁打 -7.73607
% -7.73615
部 -7.73616
成 -7.737
▁感觉 -7.73764
▁它 -7.74581
书 -7.74603
下 -7.74995
▁技术 -7.7586
▁主要 -7.76137
前 -7.77794
▁里 -7.78913
▁比较 -7.79195
▁作 -7.79235
节 -7.79633
▁真的 -7.79636
▁大家 -7.8005
去 -7.8159
▁不能 -7.81808
▁专业 -7.81904
▁所 -7.81956
▁一些 -7.82171
▁基本 -7.82486
▁那么 -7.82556
▁得 -7.83407
▁研究 -7.83531
▁不会 -7.83545
▁企业 -7.83945
力 -7.84313
行 -7.84567
▁吃 -7.85665
▁比 -7.86217
▁国家 -7.87584
▁前 -7.87836
机 -7.87869
▁产品 -7.88315
▁二 -7.88326
解 -7.88467
▁才 -7.8863
过 -7.88708
线 -7.89074
方 -7.89802
▁进行 -7.90733
~ -7.91162
▁信息 -7.91928
▁其他 -7.91941
▁太 -7.92056
▁无 -7.92059
▁游戏 -7.92279
说 -7.92318
名 -7.92837
天 -7.9287
▁使用 -7.93038
中 -7.93372
▁服务 -7.9391
▁买 -7.94109
▁这些 -7.94726
少 -7.95389
▁8 -7.95424
▁7 -7.95771
▁如 -7.96206
▁学校 -7.97144
▁其实 -7.97555
度 -7.97668
头 -7.97818
器 -7.98839
▁不要 -7.99375
▁一直 -7.99572
▁一下 -7.99744
▁开 -8.00033
▁世界 -8.00702
▁该 -8.01514
▁学习 -8.02077
类 -8.02195
▁一样 -8.02244
样 -8.0315
近 -8.03179
室 -8.03454
▁却 -8.03482
市 -8.03764
▁男 -8.04081
s -8.04191
▁个人 -8.04427
▁设计 -8.04585
" -8.05094
▁发现 -8.05217
▁情况 -8.05693
▁于 -8.06297
量 -8.06499
▁全 -8.06722
▁通过 -8.06948
▁找 -8.07011
▁社会 -8.07296
▁朋友 -8.07377
▁这么 -8.08182
▁包 -8.08657
～ -8.08742
▁玩 -8.09015
▁第二 -8.09324
▁一定 -8.09461
▁爱 -8.09613
▁选择 -8.09767
▁内 -8.09961
期 -8.10013
得 -8.1029
▁女 -8.10474
位 -8.10663
▁最后 -8.11401
▁经济 -8.11555
友 -8.11603
一起 -8.12069
年 -8.12554
▁重要 -8.12605
▁事 -8.12763
▁东西 -8.13256
▁学 -8.13693
【 -8.14297
▁一般 -8.14677
公司 -8.14865
】 -8.15965
▁走 -8.16156
心 -8.16266
山 -8.16314
▁分 -8.16343
水 -8.16822
▁大学 -8.16927
▁各种 -8.16972
▁请 -8.17373
▁不同 -8.17663
▁虽 -8.17688
地 -8.17755
县 -8.17958
之 -8.18313
▁文化 -8.18599
▁以及 -8.1882
用 -8.19191
▁非常 -8.19451
底 -8.19608
事 -8.20301
关 -8.20466
▁系统 -8.20674
▁教育 -8.20696
2 -8.20997
学 -8.21086
▁希望 -8.2124
▁孩子 -8.21269
▁像 -8.21343
▁影响 -8.21377
▁美国 -8.21459
的 -8.21625
级 -8.2229
▁学生 -8.22639
▁同 -8.2285
▁名 -8.22905
开 -8.23048
▁别人 -8.23104
员 -8.23177
▁成为 -8.23192
▁至 -8.23193
▁回答 -8.2396
▁使 -8.24195
▁内容 -8.24198
▁另 -8.24743
村 -8.24981
场 -8.25077
够 -8.25283
▁9 -8.2531
边 -8.25395
▁认 -8.25499
▁先 -8.25947
起来 -8.26508
▁0 -8.26861
▁四 -8.26866
▁方法 -8.27298
物 -8.27409
师 -8.27436
▁20 -8.27513
▁哪些 -8.27646
立 -8.27833
▁其中 -8.27839
感 -8.27871
▁特别 -8.27913
▁老 -8.28262
▁12 -8.28341
队 -8.28841
▁问 -8.29348
▁两 -8.29835
条 -8.30434
▁各 -8.31077
▁自 -8.31188
身 -8.31636
▁电影 -8.31824
路 -8.32001
文 -8.3204
手 -8.3236
见 -8.32557
想 -8.33227
▁方 -8.33334
▁所有 -8.33421
▁提供 -8.33527
▁成 -8.33642
▁张 -8.34062
▁比如 -8.34078
▁不过 -8.3408
▁两个 -8.34128
▁后来 -8.34489
不 -8.34587
▁写 -8.3477
▁怎样 -8.34887
历 -8.36055
事情 -8.36265
▁哪 -8.36534
而 -8.36688
▁此 -8.36822
▁按 -8.36921
▁一次 -8.37099
先 -8.38051
▁市场 -8.38555
▁水 -8.38556
▁几 -8.3876
▁地方 -8.38781
▁同时 -8.38922
体 -8.39021
品 -8.39024
▁能力 -8.39096
▁你们 -8.39448
▁第三 -8.39455
▁听 -8.39697
国 -8.39872
▁长 -8.39966
▁首 -8.40074
▁相关 -8.40224
▁谁 -8.40236
案 -8.40372
入 -8.41075
朋友 -8.4112
▁带 -8.41246
待 -8.41527
▁是否 -8.41664
率 -8.41777
▁项目 -8.41793
一 -8.41825
积 -8.4217
▁一点 -8.43009
▁目前 -8.43185
▁要求 -8.43316
▁国际 -8.43395
▁即 -8.4354
▁A -8.43851
▁向 -8.44269
卡 -8.44313
儿 -8.44475
▁老师 -8.446
▁日本 -8.44854
▁万 -8.44967
▁生产 -8.45312
▁组织 -8.45477
▁功能 -8.45732
▁历史 -8.45825
▁建设 -8.45935
▁那个 -8.45957
▁变 -8.45978
色 -8.4607
▁国内 -8.46081
▁支持 -8.46238
▁出现 -8.46372
▁基础 -8.46623
▁拿 -8.46774
车 -8.4679
口 -8.46837
▁主 -8.47072
制 -8.47519
局 -8.47534
意 -8.47666
本 -8.47704
▁王 -8.47946
▁11 -8.48008
▁五 -8.48325
1 -8.48525
乎 -8.48584
▁直接 -8.48815
在 -8.48858
▁以后 -8.49218
▁叫 -8.49313
斯 -8.49407
任 -8.49664
不是 -8.49779
▁方面 -8.49836
▁结果 -8.49912
气 -8.502
▁+ -8.50271
▁关系 -8.50943
▁手机 -8.51263
商 -8.51278
▁体验 -8.51309
▁以上 -8.51607
▁钱 -8.51713
米 -8.51886
▁环境 -8.52005
▁天 -8.52088
▁安全 -8.52452
许 -8.52471
▁B -8.52727
▁网络 -8.528
▁正 -8.52825
德 -8.53154
括 -8.53411
处 -8.53574
▁原 -8.53752
来说 -8.53781
▁李 -8.53976
▁知乎 -8.5406
▁北京 -8.5434
尔 -8.54453
起 -8.54518
」 -8.54534
▁放 -8.5485
称 -8.54878
字 -8.55047
城 -8.55093
▁负责 -8.55234
片 -8.55351
加 -8.55432
照 -8.55457
▁完全 -8.55656
▁分析 -8.56185
▁30 -8.56211
▁城市 -8.56235
号 -8.5634
▁活动 -8.56471
▁15 -8.56619
题 -8.56758
能 -8.56851
▁特 -8.56868
* -8.56933
▁低 -8.57137
3 -8.57165
▁章 -8.57304
▁指 -8.57595
钱 -8.57734
大学 -8.57774
图 -8.57942
定 -8.5808
「 -8.5854
▁别 -8.5867
▁元 -8.58708
▁发生 -8.59285
▁_ -8.59382
▁原因 -8.59442
▁毕业 -8.59613
办 -8.60475
业 -8.61143
动 -8.61437
像 -8.61452
▁除 -8.61632
时候 -8.61734
▁应用 -8.61811
型 -8.61997
▁数据 -8.62176
▁过程 -8.62367
▁甚至 -8.62412
▁具有 -8.62555
▁全国 -8.62661
▁C -8.62773
▁回 -8.63001
病 -8.63469
门 -8.63522
▁不好 -8.63641
数 -8.63645
▁投资 -8.64047
▁评价 -8.64197
▁题主 -8.64447
利 -8.64482
▁存在 -8.64636
▁S -8.65183
▁自然 -8.65507
▁属 -8.65676
▁强 -8.65681
▁行业 -8.65905
▁花 -8.65968
▁真 -8.66072
▁完 -8.66123
▁答 -8.66229
▁获得 -8.66587
▁发 -8.66594
▁黑 -8.66625
件 -8.66856
▁有限 -8.67202
▁较 -8.67852
几个 -8.68702
金 -8.6874
小 -8.6877
▁差 -8.68827
▁控制 -8.69046
---------------- -8.69052
▁干 -8.69132
▁副 -8.69317
▁作品 -8.69325
▁曾 -8.69547
▁用户 -8.69556
▁任何 -8.69589
时间 -8.7011
▁故事 -8.70224
发 -8.70345
▁女生 -8.70446
花 -8.70475
内 -8.70981
▁开发 -8.71299
院 -8.71312
光 -8.71379
看 -8.71462
▁根据 -8.71479
军 -8.7189
▁这是 -8.7192
▁推荐 -8.71933
▁金 -8.7209
妈 -8.7238
▁死 -8.72418
服 -8.72519
▁美 -8.72594
▁参加 -8.72621
怕 -8.72685
▁某 -8.72823
▁努力 -8.73008
▁建筑 -8.73076
日 -8.73114
▁点 -8.73383
▁上海 -8.73449
克 -8.7359
▁号 -8.73634
▁品牌 -8.73717
赛 -8.73774
▁父母 -8.73863
▁心 -8.74765
▁理解 -8.74786
爱 -8.74849
▁讲 -8.7488
▁白 -8.74901
▁处理 -8.75023
▁拥有 -8.75643
▁部分 -8.75876
▁16 -8.75939
▁解决 -8.75992
▁每天 -8.76052
▁岁 -8.76079
▁只能 -8.76191
款 -8.76388
▁教学 -8.76542
▁中心 -8.76688
有人 -8.76719
▁同学 -8.76982
▁英语 -8.77163
质 -8.77216
以 -8.773
▁路 -8.77302
▁工程 -8.77723
▁西 -8.77789
▁艺术 -8.77798
▁么 -8.77958
▁建议 -8.78362
元 -8.78511
▁标准 -8.78532
▁简单 -8.78599
▁关于 -8.78644
▁得到 -8.78682
费 -8.78721
风 -8.7882
▁100 -8.78955
▁起 -8.78966
▁算 -8.78973
▁音乐 -8.7898
概 -8.78983
▁理论 -8.79284
明 -8.79414
▁成功 -8.79989
▁重 -8.80023
▁类型 -8.80115
▁介绍 -8.80214
▁无法 -8.80457
念 -8.80762
主 -8.80788
▁加 -8.80946
▁难 -8.81085
▁反 -8.81348
C -8.8137
▁刚 -8.81543
▁挺 -8.81572
科 -8.81708
语 -8.81718
▁有点 -8.81744
▁经 -8.81746
▁考 -8.82105
▁必须 -8.82334
▁进 -8.82499
▁单位 -8.82687
▁最大 -8.82755
月 -8.82794
网 -8.82888
▁则 -8.82983
▁连 -8.83095
给 -8.83268
▁外 -8.83426
▁相 -8.83471
多 -8.83655
▁车 -8.83674
▁容易 -8.83744
▁科学 -8.83784
教 -8.83852
台 -8.84122
▁黄 -8.84578
▁实现 -8.84662
出版社 -8.84723
▁知识 -8.85023
▁提高 -8.85131
▁产生 -8.85304
进 -8.85419
站 -8.85603
▁效果 -8.85939
4 -8.86069
▁面 -8.86103
当 -8.86189
▁清 -8.86192
▁球 -8.86197
▁非 -8.86223
简介 -8.8656
▁18 -8.86893
▁周 -8.87012
视 -8.87191
▁优秀 -8.87315
没有 -8.87666
▁多少 -8.8768
美 -8.8768
▁运动 -8.87709
▁米 -8.87901
建 -8.88018
▁注意 -8.88154
往 -8.88375
▁跑 -8.88402
5 -8.88714
▁T -8.88764
▁称 -8.88875
店 -8.89113
系统 -8.89143
▁经常 -8.89206
作 -8.89364
▁约 -8.89573
海 -8.89612
▁决定 -8.89712
代 -8.8975
▁操作 -8.8978
▁行为 -8.8978
园 -8.8998
权 -8.90125
▁质量 -8.90619
▁现代 -8.90851
▁因 -8.90986
世 -8.91064
▁经验 -8.91158
高 -8.9121
▁选 -8.91329
▁科技 -8.91479
▁越 -8.91859
特 -8.91901
▁红 -8.92289
单 -8.924
▁教 -8.92428
▁作用 -8.92463
▁陈 -8.92577
形 -8.92735
▁里面 -8.92832
▁网站 -8.93338
▁软件 -8.93569
▁东 -8.93597
声 -8.93706
▁南 -8.93789
▁双 -8.93834
▁平台 -8.93869
▁心理 -8.94151
▁啦 -8.94178
▁阿 -8.94211
▁不错 -8.94285
▁职业 -8.94357
▁懂 -8.94441
▁因此 -8.94585
掉 -8.94726
▁人生 -8.94773
州 -8.9488
住 -8.9489
可 -8.95018
▁完成 -8.95028
▁提 -8.95322
流 -8.95467
▁受 -8.95595
真正 -8.95731
界 -8.95827
▁快 -8.95958
情 -8.96066
▁帮助 -8.96143
▁以前 -8.96234
▁告诉 -8.96453
白 -8.96472
问 -8.96472
校 -8.96476
▁作者 -8.96641
▁法律 -8.96694
▁形成 -8.96796
▁刘 -8.9686
快 -8.96911
▁50 -8.96933
▁遇到 -8.97112
王 -8.97252
▁价格 -8.97326
▁医院 -8.97413
▁客户 -8.9757
▁普通 -8.97631
言 -8.97974
▁您 -8.98082
▁站 -8.98188
▁治疗 -8.98194
▁条件 -8.98339
状 -8.98525
▁考虑 -8.9853
▁结构 -8.98773
▁进入 -8.98867
▁代表 -8.99033
▁保护 -8.993
▁理 -8.99433
▁改 -8.99518
▁小时 -8.9963
▁今天 -8.99681
重 -8.99727
民 -8.99743
观 -8.99788
▁一年 -8.99813
▁实际 -8.99837
▁接受 -8.99948
房 -9
▁第四 -9.00054
▁手 -9.00131
▁2010 -9.00234
▁坐 -9.00294
▁正常 -9.00602
▁网 -9.00651
▁省 -9.00742
▁综合 -9.00751
所 -9.00862
▁推 -9.00883
▁家庭 -9.00925
▁才能 -9.01045
▁改变 -9.01078
▁随 -9.01086
▁继续 -9.0118
强 -9.01188
▁帮 -9.01367
▁水平 -9.01463
系 -9.01554
▁合作 -9.01568
▁准备 -9.01872
圈 -9.01881
林 -9.02338
▁建立 -9.02409
▁关注 -9.0242
集 -9.02491
▁离 -9.02727
石 -9.02794
▁男生 -9.02892
▁模式 -9.03012
▁于是 -9.03129
施 -9.0316
▁达到 -9.03337
▁传统 -9.03419
▁制作 -9.03445
▁不断 -9.03593
▁看看 -9.03621
拉 -9.03817
分 -9.03902
▁之一 -9.03927
▁具体 -9.03944
▁培训 -9.04103
角 -9.04219
▁目标 -9.0424
▁左右 -9.04283
▁适合 -9.04299
战 -9.04339
▁坚持 -9.04388
▁表现 -9.04527
▁资源 -9.04597
版 -9.04695
层 -9.04822
▁认识 -9.04864
兵 -9.04949
▁14 -9.04978
▁版 -9.05373
▁地区 -9.05387
▁机构 -9.0551
马 -9.05527
楼 -9.05699
▁一天 -9.05703
▁占 -9.05747
镇 -9.05767
已 -9.05859
▁六 -9.05902
▁期 -9.0602
▁com -9.06107
群 -9.06132
▁转 -9.06172
项 -9.0647
盘 -9.06572
▁精神 -9.0662
▁逼 -9.06671
▁未来 -9.0668
▁肯定 -9.06714
程 -9.06937
华 -9.06959
带 -9.07024
▁等等 -9.07088
知 -9.07105
▁实验 -9.0711
▁换 -9.07127
▁家里 -9.07241
▁突然 -9.07576
届 -9.07582
▁旅游 -9.07645
▁公里 -9.07741
初 -9.0782
▁设备 -9.07928
▁出版 -9.07984
▁重点 -9.08215
▁I -9.08247
▁靠 -9.08407
▁送 -9.08491
▁语言 -9.08546
▁方向 -9.08659
读 -9.08774
▁身 -9.08919
▁马 -9.09046
死 -9.09212
▁神 -9.09214
▁一句 -9.09216
素 -9.09218
龙 -9.09283
▁系列 -9.09521
理 -9.09652
▁见 -9.09698
▁不到 -9.10122
' -9.10126
▁位于 -9.1018
▁领导 -9.10307
▁办 -9.10753
▁采用 -9.10821
奖 -9.10871
▁P -9.10883
▁计划 -9.10948
▁经营 -9.11148
y -9.11167
▁读 -9.11333
▁妈 -9.1141
▁穿 -9.11507
块 -9.11559
▁导致 -9.11617
▁销售 -9.11783
▁x -9.11808
张 -9.1181
▁人类 -9.11949
▁医 -9.12211
▁价值 -9.12378
目 -9.12543
合 -9.12554
球 -9.12718
▁2009 -9.12725
▁视频 -9.1277
▁身体 -9.12803
管 -9.12805
比 -9.12917
▁政治 -9.13009
料 -9.1335
▁确实 -9.13425
▁找到 -9.13591
平 -9.13635
工 -9.13659
▁指导 -9.13745
> -9.13783
▁意义 -9.13812
▁男人 -9.13874
收入 -9.13877
e -9.13878
▁一切 -9.14091
▁毕竟 -9.14255
▁本人 -9.14431
▁不用 -9.1461
▁相信 -9.14618
▁13 -9.14793
▁学院 -9.15054
▁说明 -9.15114
▁2011 -9.15176
▁政府 -9.15292
▁高中 -9.15468
▁谢谢 -9.15553
▁字 -9.15644
▁业务 -9.15804
▁拍 -9.15878
▁状态 -9.15939
▁整个 -9.15985
▁图 -9.16062
▁班 -9.16082
▁附 -9.16128
▁百度 -9.16226
▁the -9.1627
设 -9.16276
▁E -9.16318
▁求 -9.16377
▁2008 -9.1639
▁根本 -9.16738
河 -9.16759
7 -9.1693
▁北 -9.16961
▁合 -9.16983
油 -9.16994
报 -9.1712
宝 -9.1715
板 -9.17249
馆 -9.17257
安 -9.17296
▁第一次 -9.17352
无 -9.17431
▁古 -9.17468
▁利用 -9.1748
具 -9.17676
堂 -9.17691
论 -9.17844
▁规定 -9.1793
▁感受 -9.18076
▁香港 -9.18382
▁D -9.18462
▁未 -9.18515
▁万元 -9.18601
▁人物 -9.18629
0 -9.18769
▁拉 -9.18796
▁行政 -9.18815
▁增加 -9.1887
弃 -9.18919
通 -9.18931
新 -9.19193
▁女性 -9.19265
▁农村 -9.1928
▁简 -9.19311
▁时代 -9.19352
演 -9.19395
份 -9.19453
▁思想 -9.19508
志 -9.19556
段 -9.19563
d -9.19758
▁以为 -9.19763
▁少 -9.19867
▁健康 -9.19982
▁亩 -9.20016
眼 -9.2002
▁共 -9.20024
景 -9.20097
▁喝 -9.20228
述 -9.20575
▁提出 -9.20592
人员 -9.20781
▁人民 -9.20788
音 -9.20869
星 -9.20874
▁有效 -9.20908
达 -9.20914
▁公 -9.21149
▁七 -9.21167
值 -9.21332
夫 -9.21342
▁均 -9.21359
阳 -9.21405
直 -9.21411
▁许多 -9.21473
团 -9.21545
▁搞 -9.21697
火 -9.21799
结 -9.21927
▁越来越 -9.21929
▁各位 -9.21945
同 -9.22141
义 -9.22166
▁仍 -9.22166
▁以下 -9.22221
神 -9.22362
嘛 -9.22374
▁电子 -9.22459
士 -9.22513
公 -9.22615
▁收 -9.22669
▁表示 -9.22707
▁设 -9.22707
▁小学 -9.22745
▁曾经 -9.228
▁规划 -9.2285
▁过去 -9.22885
▁造成 -9.22925
▁M -9.23149
▁机会 -9.23152
▁心里 -9.23301
▁25 -9.23312
味 -9.23386
然而 -9.23528
▁年轻 -9.23735
表 -9.23793
▁现 -9.23926
如 -9.24088
▁需求 -9.24108
▁早 -9.24121
▁材料 -9.24148
▁http -9.24337
6 -9.24358
▁新闻 -9.24654
▁领域 -9.24676
▁海 -9.24999
▁类似 -9.25009
常 -9.25029
g -9.25138
▁空间 -9.25198
▁基 -9.25292
▁自由 -9.25317
症 -9.25359
由 -9.25548
▁考试 -9.2555
西 -9.25567
▁明显 -9.25689
▁产业 -9.25707
▁台湾 -9.25729
取 -9.25772
设计 -9.25856
组 -9.25864
▁单 -9.25913
词 -9.25975
至于 -9.26048
▁程度 -9.26426
▁a -9.26523
▁& -9.26553
史 -9.26575
录 -9.26668
▁获 -9.26687
▁女人 -9.26734
▁啥 -9.2675
▁交 -9.26758
厂 -9.2676
▁卖 -9.26911
▁哪里 -9.26971
货 -9.26994
▁最终 -9.27081
▁广告 -9.27187
▁生命 -9.27218
▁集团 -9.27303
▁2006 -9.27338
▁F -9.27353
及 -9.27674
格 -9.27737
食 -9.2775
班 -9.27778
云 -9.28149
▁偏 -9.28292
▁从事 -9.2836
▁书 -9.28375
▁八 -9.28462
▁事件 -9.28571
▁经过 -9.28592
▁减 -9.28644
▁含 -9.28979
▁17 -9.29115
▁实施 -9.29219
▁电 -9.29274
影 -9.29284
女 -9.29354
▁2007 -9.29453
源 -9.29462
价 -9.29496
厅 -9.29534
▁愿意 -9.29655
▁便 -9.29662
▁画 -9.29753
▁40 -9.29999
▁交流 -9.30026
▁百 -9.30093
乐 -9.30111
药 -9.30255
▁讨论 -9.30543
▁变化 -9.3064
朝 -9.30772
▁加入 -9.30788
变 -9.31117
至 -9.31218
▁热 -9.31324
× -9.31328
▁三个 -9.31435
▁of -9.31479
▁村 -9.31616
▁距离 -9.31702
▁政策 -9.31711
▁全部 -9.3183
信 -9.32323
▁无论 -9.32326
领 -9.32326
从 -9.32457
▁任务 -9.32486
菜 -9.32586
酒 -9.3266
▁范围 -9.32662
▁往 -9.32718
三 -9.32753
座 -9.32913
▁电脑 -9.33031
G -9.33068
▁补充 -9.33127
回 -9.33136
▁型 -9.33202
▁专 -9.33208
▁创新 -9.33312
侧 -9.33393
纳 -9.33456
▁一段 -9.33623
▁比赛 -9.33717
▁相对 -9.33834
▁简介 -9.33851
▁丰富 -9.33907
▁先进 -9.33937
D -9.3397
东 -9.34124
▁独立 -9.34195
▁人口 -9.34219
▁显示 -9.34235
▁目的 -9.34253
▁程序 -9.34353
篇 -9.34414
▁据 -9.34533
伤 -9.34556
▁2000 -9.34615
▁大量 -9.34752
▁小说 -9.34782
▁现实 -9.34856
▁倒 -9.34891
例 -9.35066
▁我国 -9.35097
▁门 -9.35102
二 -9.35122
▁感情 -9.35178
受 -9.35242
▁成绩 -9.35249
么 -9.35286
渐 -9.35351
▁联系 -9.35529
▁N -9.35631
▁参考 -9.35728
久 -9.35765
丝 -9.35933
▁办公 -9.35971
则 -9.36111
▁所谓 -9.36116
干 -9.36401
▁商业 -9.36533
▁纯 -9.36536
▁评论 -9.36544
▁解释 -9.36585
▁网上 -9.36614
包 -9.36712
岁 -9.36728
▁微博 -9.36768
▁图书 -9.36771
害 -9.36912
差 -9.37065
▁晚上 -9.37067
▁镇 -9.37096
▁金融 -9.37187
粉 -9.37261
▁保持 -9.37331
▁课程 -9.37384
▁如此 -9.37423
▁微信 -9.37442
▁半 -9.3762
▁农业 -9.37628
▁大部分 -9.37913
▁市 -9.37929
江 -9.38079
▁表 -9.38092
▁记得 -9.38114
▁培养 -9.38198
正 -9.38357
▁结婚 -9.38466
▁汽车 -9.3852
回来 -9.38544
▁安 -9.38554
▁爱情 -9.38634
▁工业 -9.38732
▁初 -9.38807
▁克 -9.38807
剂 -9.38834
8 -9.38952
▁达 -9.39022
官 -9.39042
▁关键 -9.39103
▁24 -9.39119
课 -9.3952
▁情 -9.39592
▁收入 -9.39647
▁平 -9.39691
▁患 -9.39695
▁想法 -9.39784
▁入 -9.39821
▁交通 -9.39854
▁专家 -9.39877
▁杨 -9.39883
答 -9.39922
▁特色 -9.39925
▁尤其 -9.39978
▁位置 -9.40015
兰 -9.40061
▁资料 -9.40085
▁带来 -9.40106
▁数 -9.40163
相 -9.40253
▁超 -9.4032
▁山 -9.40323
调 -9.40367
▁2012 -9.40422
灯 -9.40449
皮 -9.40498
▁保证 -9.40654
▁似 -9.40658
▁核心 -9.40762
▁人才 -9.4087
▁思考 -9.40917
装 -9.41145
▁远 -9.41146
▁人员 -9.4129
▁制度 -9.41478
▁团队 -9.41576
游戏 -9.41589
▁稳定 -9.41763
▁21 -9.41844
▁亿 -9.41991
▁超过 -9.42
空 -9.42107
▁微 -9.42117
▁一家 -9.42163
▁短 -9.42255
▁余 -9.42262
▁分钟 -9.42397
▁第五 -9.42415
▁严重 -9.4244
▁哭 -9.42545
▁户 -9.42575
▁分享 -9.42743
▁压力 -9.42887
▁参与 -9.42891
▁数学 -9.42925
▁不管 -9.42963
及其 -9.42974
▁即使 -9.43015
▁全面 -9.43195
▁光 -9.43434
▁角度 -9.43497
良 -9.43537
▁明白 -9.43639
▁学术 -9.43661
▁九 -9.43678
▁深 -9.43753
▁申请 -9.43762
▁不少 -9.43902
▁狗 -9.44163
传 -9.44171
▁常 -9.44234
▁显 -9.44238
▁2005 -9.44274
记 -9.44483
▁描述 -9.44587
▁接触 -9.44626
▁优势 -9.44658
▁方便 -9.44658
▁G -9.44661
季 -9.44837
▁满足 -9.44925
▁类 -9.44935
为主 -9.44952
▁60 -9.45194
剧 -9.45204
▁R -9.45223
句 -9.45249
▁实在 -9.45263
▁文学 -9.45336
▁真是 -9.45381
▁部门 -9.45403
▁来自 -9.45449
姐 -9.45465
▁计算机 -9.4555
▁借 -9.45605
▁云 -9.4561
▁教师 -9.45626
红 -9.45683
▁楼 -9.45849
▁接 -9.45903
▁银行 -9.45975
远 -9.46092
▁留 -9.46161
▁受到 -9.46286
▁著名 -9.46372
▁石 -9.46672
▁照片 -9.46732
▁形式 -9.46746
▁自动 -9.46771
▁真实 -9.4679
岛 -9.46815
▁回家 -9.46866
ing -9.4693
▁十分 -9.46962
▁确定 -9.46982
▁习惯 -9.47001
牌 -9.47081
田 -9.47274
▁移动 -9.47296
▁L -9.47298
▁肉 -9.47304
境 -9.47332
▁处 -9.47381
▁绝对 -9.47558
m -9.47607
▁多年 -9.47807
▁活 -9.47835
氏 -9.47929
遍 -9.4803
管理 -9.48031
▁酒店 -9.48171
▁提升 -9.48232
▁成本 -9.48243
▁检查 -9.48258
▁通 -9.48259
适 -9.48272
证 -9.48565
▁H -9.48678
▁定 -9.4877
▁风格 -9.48778
▁结合 -9.48814
编 -9.48825
er -9.48908
▁好好 -9.48979
▁互联网 -9.49034
宗 -9.49075
▁分别 -9.49085
▁风 -9.49091
尼 -9.49209
曲 -9.4925
架 -9.49257
▁民族 -9.49314
▁幸福 -9.49433
▁英国 -9.49685
基 -9.49693
▁今年 -9.49733
▁集 -9.49892
▁速度 -9.49903
鱼 -9.49919
▁装 -9.50007
▁80 -9.50011
▁文件 -9.50022
W -9.50186
▁电话 -9.502
▁衣 -9.50543
▁永远 -9.50601
望 -9.50738
▁19 -9.50743
▁积极 -9.50774
和 -9.50778
▁教授 -9.5083
▁交易 -9.50864
南 -9.51017
▁会员 -9.51041
何 -9.51092
▁轻 -9.51227
★ -9.51239
▁生物 -9.51243
联 -9.51344
▁妹 -9.5145
微 -9.51521
▁皮肤 -9.51568
▁追求 -9.51692
势 -9.51698
▁配 -9.51704
▁m -9.51709
▁任 -9.51758
▁组成 -9.51771
▁平时 -9.51904
▁汉 -9.51914
政府 -9.51995
▁开心 -9.52106
▁面对 -9.52169
▁长期 -9.52201
丽 -9.52244
▁本来 -9.52283
▁乡 -9.52308
▁极 -9.52512
▁发表 -9.5254
▁最高 -9.52606
▁博士 -9.52655
mm -9.52751
▁吴 -9.52897
▁诗 -9.52916
▁追 -9.52957
就算 -9.52959
清 -9.52981
▁委员会 -9.53031
▁笑 -9.53058
▁不可 -9.53181
▁背 -9.53181
▁乐 -9.53202
端 -9.53285
▁贴 -9.53289
其 -9.53329
叶 -9.53364
▁值得 -9.5343
电话 -9.53438
香 -9.53454
▁慢慢 -9.53534
▁明 -9.53677
▁设置 -9.53856
▁十 -9.53909
▁性格 -9.5399
▁阅读 -9.54014
▁房 -9.54088
罗 -9.54256
老 -9.54286
自 -9.54409
全 -9.5443
炎 -9.54448
歌 -9.54479
▁若 -9.54722
乡 -9.54734
▁认真 -9.54947
▁建 -9.55116
▁高考 -9.5514
画 -9.55202
治 -9.55361
▁本科 -9.55371
不多 -9.55395
▁必要 -9.55536
V -9.55659
▁一条 -9.55756
▁毛 -9.55787
▁中学 -9.55848
▁出去 -9.56252
▁劳动 -9.56321
▁训练 -9.56371
木 -9.56405
▁卡 -9.56412
著 -9.56412
▁看着 -9.56443
派 -9.56654
▁宣传 -9.5677
▁2004 -9.56809
▁创业 -9.5691
卷 -9.56932
▁意思 -9.56947
▁足 -9.57144
实 -9.57217
▁引起 -9.57217
▁自我 -9.5731
▁正确 -9.57411
@ -9.57484
▁表达 -9.57637
▁事业 -9.57745
非 -9.57868
人家 -9.57924
▁条 -9.58039
▁背景 -9.58073
峰 -9.58088
▁样子 -9.58135
步 -9.58199
▁执行 -9.58229
▁不仅 -9.58266
▁病 -9.58312
▁概念 -9.58326
▁更新 -9.58374
电 -9.58394
▁科研 -9.58398
客 -9.58436
9 -9.5855
晚 -9.5856
▁接口 -9.58576
▁尺寸 -9.58602
▁谢 -9.58606
▁飞 -9.58706
▁清楚 -9.58786
▁媒体 -9.58847
选 -9.58849
▁研究生 -9.589
▁力量 -9.58919
▁享受 -9.58955
▁商 -9.58993
▁and -9.59037
▁22 -9.59122
学院 -9.59153
都 -9.59163
▁V -9.59186
▁年代 -9.59191
▁吃饭 -9.59206
▁K -9.59298
▁共同 -9.59315
▁哦 -9.59386
▁机械 -9.59419
▁营销 -9.59425
K -9.59435
▁-- -9.59603
波 -9.59711
户 -9.59742
▁动物 -9.59752
▁唯一 -9.5978
S -9.59819
▁区别 -9.59951
a -9.60017
肉 -9.60146
▁发布 -9.60166
▁主任 -9.60219
热 -9.60226
▁全球 -9.60291
▁文章 -9.60347
根 -9.60374
▁不敢 -9.60393
▁医疗 -9.60425
▁卫生 -9.60425
x -9.60432
▁一位 -9.60478
▁高级 -9.6049
▁亦 -9.60556
此 -9.60556
w -9.60563
向 -9.60643
▁脸 -9.60646
▁安装 -9.60722
▁X -9.60732
▁冷 -9.60794
▁调 -9.60827
族 -9.60874
▁不够 -9.60961
▁特殊 -9.61129
▁精 -9.61235
▁科 -9.61266
▁读书 -9.61283
▁员工 -9.61304
▁好多 -9.61395
▁性能 -9.61414
▁父亲 -9.61423
▁说话 -9.61449
▁原来 -9.61528
▁罗 -9.61579
▁图片 -9.61635
▁J -9.61742
▁实践 -9.61821
▁国外 -9.61866
M -9.61881
▁电视 -9.61899
▁责任 -9.61958
▁快速 -9.61966
▁宽 -9.62031
▁技能 -9.62138
走 -9.62221
▁林 -9.62274
} -9.62405
▁兴趣 -9.62488
{ -9.6249
应 -9.62582
▁实 -9.62641
▁成长 -9.6269
▁韩国 -9.62724
▁声音 -9.62775
▁相当 -9.62898
▁农民 -9.63025
弟 -9.63067
▁欧洲 -9.6322
▁住 -9.63252
▁创造 -9.63302
A -9.63444
▁居 -9.63496
打 -9.63576
▁计算 -9.6358
▁90 -9.63642
▁现象 -9.63685
箱 -9.63769
算 -9.63783
▁角色 -9.63875
p -9.64013
▁针对 -9.64027
▁正在 -9.64063
▁竞争 -9.64068
▁德国 -9.64086
酸 -9.64126
▁其它 -9.64159
草 -9.64372
k -9.64468
▁感谢 -9.64492
桥 -9.64498
▁反应 -9.64546
居 -9.64565
▁风险 -9.64599
▁分手 -9.64765
▁200 -9.64832
难 -9.64851
▁火 -9.64916
▁自身 -9.64958
▁证明 -9.65041
▁购买 -9.65059
▁不了 -9.65069
▁体育 -9.65113
命 -9.65127
别 -9.65164
b -9.65279
中心 -9.65471
刻 -9.65474
▁商品 -9.65474
▁工具 -9.65539
▁第六 -9.65567
▁战略 -9.65701
▁协会 -9.65809
▁那样 -9.65818
▁英 -9.65852
放 -9.65937
▁文 -9.65944
▁国 -9.6595
px -9.65998
▁既 -9.66032
▁地图 -9.66093
树 -9.66118
果 -9.66166
▁2003 -9.66201
▁铁 -9.663
▁龙 -9.66301
街 -9.66362
▁结束 -9.66427
▁抱 -9.66557
▁观点 -9.66636
▁原则 -9.66664
遇 -9.66668
对 -9.66696
▁参数 -9.66752
▁因素 -9.66761
活 -9.6677
▁免费 -9.66885
▁杀 -9.67075
关系 -9.67227
压 -9.67443
▁牛 -9.67464
▁区域 -9.67511
玉 -9.6754
n -9.67844
血 -9.6785
▁呀 -9.6786
▁取得 -9.67879
▁翻 -9.67903
▁女孩 -9.67954
▁社区 -9.6801
布 -9.6805
▁平均 -9.68244
终于 -9.68281
液 -9.68307
教育 -9.68314
五 -9.68323
▁组 -9.68343
野 -9.68346
求 -9.68486
▁资金 -9.68491
c -9.68513
▁苏 -9.68531
饭 -9.68625
▁中央 -9.68663
▁23 -9.68731
▁化学 -9.68733
术 -9.68925
▁信 -9.68964
▁气 -9.69
套 -9.69041
▁制造 -9.69068
▁对象 -9.69104
▁正式 -9.69166
▁世纪 -9.6934
社 -9.69349
▁开展 -9.69383
▁套 -9.69399
▁木 -9.6944
▁来源 -9.69513
▁调查 -9.69637
▁骂 -9.69708
考 -9.6977
▁植物 -9.69787
▁赵 -9.69825
况 -9.69826
脚 -9.69841
夜 -9.69911
存 -9.69994
o -9.69996
▁内心 -9.70044
▁不如 -9.70047
▁改革 -9.70091
▁睡 -9.70127
▁从来 -9.70158
池 -9.70195
▁符合 -9.70236
▁今 -9.70256
易 -9.70257
▁喷 -9.70259
▁分类 -9.70318
刀 -9.70396
▁直 -9.70641
▁记 -9.70684
▁论文 -9.70707
落 -9.70777
▁需 -9.70787
▁咨询 -9.70875
▁测试 -9.71018
计 -9.71042
使 -9.71049
▁担 -9.71099
▁管 -9.71109
▁善 -9.71148
衣 -9.712
▁态度 -9.71302
指 -9.71447
径 -9.71559
▁英雄 -9.71584
▁调整 -9.71634
▁当年 -9.71955
▁28 -9.71961
世界 -9.71975
▁详细 -9.71996
托 -9.72036
▁大型 -9.72074
塔 -9.72237
▁蓝 -9.72258
▁加上 -9.7231
▁例 -9.72346
意思 -9.72365
▁错 -9.72397
约 -9.7241
▁青 -9.72412
▁体系 -9.72474
才 -9.72478
令 -9.72519
▁消费 -9.72522
▁食品 -9.72655
▁血 -9.72748
▁唐 -9.72781
▁阶段 -9.72837
▁不太 -9.72864
▁法国 -9.72885
▁主动 -9.72922
ed -9.72955
▁种植 -9.72962
身上 -9.73082
转 -9.73095
镜 -9.73119
▁事实 -9.73153
▁千 -9.73198
止 -9.73226
飞 -9.73292
▁s -9.73309
▁谈 -9.73473
做 -9.73476
骨 -9.73524
▁www -9.73538
▁基金 -9.7374
完 -9.73773
▁伤害 -9.73805
▁2016 -9.7386
杀 -9.73903
愿 -9.73948
距 -9.73959
▁内部 -9.73967
▁反正 -9.73975
▁左 -9.74
轮 -9.74034
▁段 -9.74094
北 -9.74169
▁推广 -9.74245
▁良好 -9.74252
仁 -9.74371
早 -9.74418
脸 -9.74475
▁26 -9.74548
▁一件 -9.74578
▁in -9.74719
▁做好 -9.74743
▁加工 -9.74755
方面 -9.74773
额 -9.74794
养 -9.74863
▁统一 -9.74887
▁利益 -9.7508
▁促进 -9.75102
▁娱乐 -9.7514
▁页 -9.75147
▁to -9.75271
▁菜 -9.75276
▁三年 -9.75279
▁重新 -9.75341
部门 -9.75524
▁稍 -9.75612
▁名称 -9.75659
▁疾病 -9.75711
临 -9.75711
▁大多数 -9.75761
▁例如 -9.75784
庄 -9.75853
▁代 -9.75875
▁周围 -9.75931
思 -9.7594
▁颜色 -9.75959
▁口 -9.75963
▁注册 -9.7601
▁平方 -9.76125
B -9.7613
▁工资 -9.76142
▁最佳 -9.7619
▁假 -9.7632
▁经典 -9.76369
’ -9.76399
纸 -9.76503
佛 -9.76521
▁O -9.76563
▁爸爸 -9.76747
▁c -9.76791
▁主持 -9.7685
主义 -9.76903
▁承担 -9.76943
▁大小 -9.76968
▁文字 -9.77013
首 -9.77026
▁华 -9.77031
▁创作 -9.77104
▁摄影 -9.77296
▁故 -9.77387
▁逐 -9.77445
胜 -9.77452
▁歌 -9.77454
省 -9.77533
\ -9.77541
▁持续 -9.77618
▁猫 -9.77731
▁评 -9.77867
▁01 -9.77885
速 -9.77926
▁成员 -9.77989
▁b -9.77999
▁亲 -9.78053
▁茶 -9.78056
▁仅 -9.78083
▁保险 -9.78092
某些 -9.78135
司 -9.78138
▁大陆 -9.78142
票 -9.78149
▁补 -9.78176
笔 -9.78205
▁记录 -9.78211
▁大神 -9.78257
▁合理 -9.78273
▁刷 -9.78297
党 -9.78302
▁掌握 -9.78333
技术 -9.78391
▁导演 -9.78528
▁苹果 -9.78534
▁宋 -9.78641
▁广州 -9.78709
▁初中 -9.78712
太 -9.7876
屏 -9.7887
▁2002 -9.78916
▁词 -9.78966
▁营养 -9.78998
▁母亲 -9.79011
▁吸引 -9.7904
含 -9.79041
▁估计 -9.79064
笑 -9.7907
▁何 -9.79099
惜 -9.79101
l -9.79121
等奖 -9.79352
梦 -9.79584
▁编辑 -9.79585
▁n -9.79624
▁p -9.79685
亚 -9.79902
▁店 -9.7993
▁淘宝 -9.80011
▁过来 -9.80052
▁县 -9.80098
▁技巧 -9.80225
春 -9.80236
▁支付 -9.80269
▁通常 -9.80377
▁恋爱 -9.80524
▁成熟 -9.80524
▁挂 -9.80672
▁中文 -9.80678
▁变得 -9.80694
▁圣 -9.80725
▁香 -9.80728
甲 -9.80728
▁先后 -9.80758
▁W -9.80776
▁运营 -9.80791
产 -9.80839
青 -9.80847
效 -9.80999
▁几年 -9.81011
满 -9.81096
▁报 -9.81116
▁意识 -9.81196
▁农户 -9.81227
原 -9.81255
▁思维 -9.81351
母 -9.81436
▁难道 -9.81503
▁70 -9.81512
▁不知 -9.81544
森 -9.81585
册 -9.81631
F -9.81662
川 -9.81672
▁来看 -9.8179
汤 -9.81863
▁硕士 -9.8191
▁一部分 -9.81944
▁监督 -9.82037
宫 -9.8206
斤 -9.82067
▁固定 -9.82076
雨 -9.82094
象 -9.82129
工程 -9.82264
▁有个 -9.82319
▁取 -9.82329
▁小孩 -9.82358
▁出生 -9.82443
▁日常 -9.82447
▁500 -9.82493
茶 -9.82578
▁几天 -9.82598
▁两年 -9.82634
等 -9.82652
膜 -9.82673
▁整体 -9.82685
▁目录 -9.82721
▁硬 -9.82733
▁眼睛 -9.82844
▁忘 -9.82886
▁教材 -9.82923
秀 -9.82982
▁聊天 -9.82991
▁数量 -9.83005
▁攻击 -9.83065
▁逻辑 -9.8307
否 -9.831
英 -9.83127
▁特征 -9.83182
▁打开 -9.83327
不到 -9.83346
▁直到 -9.83375
▁美好 -9.8339
▁偶 -9.835
▁道路 -9.83519
▁场 -9.83593
▁情绪 -9.83659
▁切 -9.83705
▁分布 -9.8371
致 -9.83879
▁第七 -9.83902
▁一张 -9.8394
▁夜 -9.83953
▁至今 -9.84001
生活 -9.84008
▁医学 -9.84019
工作 -9.84117
▁会议 -9.84184
▁不再 -9.84243
巴 -9.84247
▁鱼 -9.84254
▁欢迎 -9.84302
▁抓 -9.8431
▁查 -9.84327
▁回到 -9.84383
诗 -9.8439
▁跳 -9.84446
▁顶 -9.84459
黑 -9.84551
< -9.84555
乱 -9.84689
▁财务 -9.84733
▁2001 -9.84734
痛 -9.8477
▁卷 -9.84791
怀 -9.8485
▁赶 -9.84862
不了 -9.84927
后面 -9.84953
▁叶 -9.85046
▁快乐 -9.85071
土 -9.85171
▁1000 -9.85171
▁放在 -9.85183
▁临床 -9.85201
君 -9.85265
▁话题 -9.85265
▁零 -9.85313
▁洗 -9.85388
▁数字 -9.85529
▁公共 -9.85529
▁点击 -9.85535
▁连续 -9.85627
▁拒绝 -9.85677
^ -9.85699
▁g -9.85791
道理 -9.85807
▁当地 -9.8587
▁检测 -9.8589
▁修改 -9.85894
▁概 -9.85896
▁养 -9.85919
▁酒 -9.85923
▁土 -9.85923
▁每年 -9.85947
▁物理 -9.85955
曰 -9.85964
▁实力 -9.86
亮 -9.86011
▁儿童 -9.86084
错 -9.8614
▁年龄 -9.86152
▁唱 -9.86308
▁原理 -9.86388
▁混 -9.86401
▁传 -9.86422
▁版本 -9.86446
▁动作 -9.86517
▁感到 -9.86519
库 -9.86524
▁眼 -9.86539
▁完美 -9.86547
▁规范 -9.86571
▁投 -9.86603
备 -9.86616
湖 -9.867
i -9.86802
▁注 -9.86857
▁记忆 -9.86946
铁 -9.86951
谷 -9.86959
▁27 -9.87017
▁物质 -9.87027
汉 -9.87038
▁手术 -9.87069
注 -9.87129
▁连接 -9.87146
倒 -9.87229
▁一部 -9.87263
灵 -9.87273
▁加强 -9.87321
▁朱 -9.87348
▁党 -9.87384
▁敌 -9.87395
结构 -9.8744
▁油 -9.87446
▁主题 -9.87507
▁傻 -9.87651
▁失败 -9.87754
▁英文 -9.87797
▁依然 -9.87877
产品 -9.87969
万 -9.87983
▁演员 -9.88083
▁多种 -9.88258
雅 -9.88306
▁心情 -9.88321
▁重大 -9.88342
▁输入 -9.88413
▁全村 -9.88512
▁兼 -9.8859
不住 -9.88631
▁优 -9.88655
▁意见 -9.88768
▁工艺 -9.88795
▁中医 -9.88827
▁表面 -9.88861
毛 -9.89035
▁搜索 -9.89062
分钟 -9.89107
▁印象 -9.89116
▁鸡 -9.89167
▁运行 -9.89283
▁形象 -9.89334
武 -9.89344
▁复杂 -9.89378
▁台 -9.89493
单位 -9.89536
善 -9.89559
▁担心 -9.89563
▁沟通 -9.89575
而言 -9.8967
项目 -9.89673
▁男性 -9.89697
▁健身 -9.8977
▁The -9.89792
帝 -9.89833
▁动画 -9.89888
▁儿子 -9.90014
▁适用 -9.90091
▁智能 -9.90146
礼 -9.9031
▁练习 -9.90321
大学生 -9.90401
-9.90423
▁地位 -9.9045
▁食物 -9.90579
▁东北 -9.90602
▁广泛 -9.90691
▁恢复 -9.90804
▁多次 -9.90808
▁理想 -9.90952
一只 -9.9097
恩 -9.90976
牙 -9.91008
▁痛苦 -9.91052
▁药 -9.91202
▁青年 -9.91284
▁重量 -9.91303
▁分为 -9.91311
▁化妆 -9.91352
→ -9.91375
鞋 -9.91403
▁报告 -9.91413
▁t -9.91451
▁心态 -9.91482
▁梦 -9.915
▁体现 -9.91515
地区 -9.91541
▁判断 -9.91615
忙 -9.91632
职 -9.91662
爸 -9.91743
▁依 -9.91813
▁细节 -9.91852
▁胡 -9.91857
X -9.91876
▁就业 -9.9192
▁定义 -9.91965
▁留学 -9.92056
▁制定 -9.9212
▁剧情 -9.92164
▁老板 -9.92184
▁基地 -9.9221
▁增长 -9.92212
离 -9.9222
▁哲学 -9.92239
败 -9.92323
交 -9.92375
▁漂亮 -9.92392
▁定位 -9.92455
消费者 -9.92466
▁结 -9.92761
▁土地 -9.92789
▁避免 -9.92844
▁维护 -9.92878
宁 -9.92905
▁屏幕 -9.92908
▁降低 -9.92915
企业 -9.92918
足 -9.92961
▁归 -9.93038
▁战斗 -9.93093
f -9.93146
奇 -9.93166
▁生态 -9.93235
就 -9.93252
▁苦 -9.93501
贵 -9.93595
限 -9.93612
▁统计 -9.93694
玩 -9.93732
▁超级 -9.93793
▁两种 -9.93802
▁保障 -9.93872
▁发行 -9.9389
▁奖 -9.939
▁想想 -9.93913
▁烧 -9.9392
▁退 -9.94034
▁赚 -9.94114
▁空 -9.94118
听到 -9.94147
▁一生 -9.94158
▁哈哈哈 -9.94349
▁留下 -9.94352
▁拍摄 -9.94378
| -9.94399
▁失 -9.94428
P -9.94485
▁文明 -9.94555
▁增 -9.94907
▁知名 -9.94935
▁会计 -9.95055
▁死亡 -9.95124
邀 -9.95134
将 -9.95174
▁太阳 -9.95184
▁银 -9.95262
▁节目 -9.95266
糖 -9.95269
▁深圳 -9.95376
▁ISBN -9.95421
断 -9.95446
▁沙 -9.95557
福 -9.95574
▁再次 -9.95598
谈 -9.95606
▁案例 -9.95848
雪 -9.95915
蛋 -9.95971
▁总结 -9.96052
▁修 -9.96088
▁复 -9.96113
▁32 -9.96172
▁韩 -9.96213
▁乱 -9.9633
▁陪 -9.96339
属 -9.96342
不起 -9.96418
深 -9.96431
▁弄 -9.96485
▁投入 -9.96486
▁试 -9.96494
四 -9.96555
辈 -9.96624
▁读者 -9.96719
▁集中 -9.968
▁题 -9.96806
▁论 -9.96844
牛 -9.96878
▁完善 -9.96903
雷 -9.96912
▁规模 -9.96913
▁方案 -9.96984
▁宿舍 -9.9702
▁这一 -9.97041
哥 -9.97049
▁流行 -9.97055
▁女儿 -9.97059
▁旧 -9.97269
▁徐 -9.97271
▁视 -9.97281
▁厘米 -9.97322
▁丢 -9.97351
▁失去 -9.9749
▁既然 -9.97672
▁惊 -9.97754
▁秒 -9.97829
▁300 -9.97864
失 -9.97869
▁天天 -9.9793
▁姑娘 -9.97936
▁常见 -9.97956
▁不得 -9.98002
郎 -9.98023
▁反对 -9.98044
紧 -9.98058
▁升级 -9.98082
‘ -9.98095
▁所在 -9.98096
竟 -9.98116
功 -9.98212
▁竟 -9.98399
▁易 -9.98457
家人 -9.9849
▁现场 -9.9856
▁道德 -9.98593
戏 -9.98595
▁官 -9.98627
▁限制 -9.98644
▁明确 -9.98703
知识 -9.9882
▁久 -9.98828
寺 -9.98834
不上 -9.98946
▁战争 -9.99094
▁相比 -9.99121
▁错误 -9.99154
▁粗 -9.99198
▁适应 -9.99206
▁深入 -9.99376
▁一边 -9.99387
配 -9.99403
t -9.99421
展 -9.99441
兴 -9.99459
▁动力 -9.9947
仪 -9.99487
▁邀请 -9.99544
▁想象 -9.9964
▁及时 -9.99699
▁学科 -9.99756
虫 -9.9981
▁00 -9.9986
围 -9.99861
▁纸 -9.99971
▁29 -10.0007
▁实用 -10.0009
船 -10.0013
• -10.0021
▁赛 -10.0029
▁坏 -10.0031
▁成就 -10.0047
▁充分 -10.0049
▁代理 -10.0055
学位 -10.0062
▁专门 -10.0063
▁传播 -10.0067
码 -10.0074
▁保 -10.0076
▁夏 -10.0078
妙 -10.0101
排 -10.0107
▁安排 -10.011
▁歌曲 -10.011
▁费用 -10.0113
v -10.0115
醒 -10.0125
▁梦想 -10.0134
异 -10.0141
▁足球 -10.0148
▁私 -10.0153
臣 -10.0155
▁最新 -10.0155
▁绝 -10.0156
典 -10.0158
▁研发 -10.0163
▁晚 -10.0169
▁信号 -10.0169
现 -10.0177
▁高度 -10.018
▁青春 -10.018
泉 -10.0186
▁冲 -10.0187
▁尝试 -10.0189
▁难以 -10.0194
▁流 -10.0204
嗯 -10.0206
▁剩 -10.021
▁古代 -10.0211
写 -10.0215
祖 -10.0229
▁人均 -10.0236
H -10.0241
▁防 -10.0244
▁阴 -10.0249
徒 -10.0251
机构 -10.0257
▁翻译 -10.0269
▁认证 -10.028
▁电源 -10.0295
▁软 -10.0295
▁右 -10.0296
杯 -10.0299
▁实习 -10.0301
▁广大 -10.031
百 -10.0313
▁时尚 -10.0314
枪 -10.0314
▁排 -10.0321
▁刚刚 -10.033
▁搭 -10.0333
▁南京 -10.0336
▁年度 -10.0339
▁35 -10.0345
▁f -10.0356
▁现有 -10.0358
▁能量 -10.0363
▁巨大 -10.0364
扬 -10.0371
▁尊重 -10.0378
▁家长 -10.0383
▁战 -10.0386
升 -10.0387
▁粉丝 -10.0388
▁忍 -10.0389
▁处于 -10.0398
▁盐 -10.0408
索 -10.0416
▁充满 -10.0421
▁校园 -10.0421
r -10.0425
▁赚钱 -10.0427
▁宗教 -10.0431
▁公开 -10.0432
市场 -10.044
控 -10.044
低 -10.0447
▁协调 -10.0449
▁汤 -10.0459
▁涉及 -10.0464
▁体 -10.0472
▁专辑 -10.0484
▁中间 -10.0484
▁地球 -10.0499
▁各个 -10.051
▁旅行 -10.0515
▁片 -10.0527
▁观众 -10.0527
伯 -10.0537
尽 -10.0545
末 -10.0554
▁创建 -10.0557
运 -10.0566
▁十二 -10.0566
接 -10.057
▁满 -10.0571
政 -10.0572
销 -10.0572
舞 -10.0575
▁山东 -10.0575
精 -10.0576
▁害 -10.0579
▁宝宝 -10.0587
保 -10.0588
▁杂志 -10.0597
▁一本 -10.0601
▁优质 -10.0604
▁冬 -10.0605
修 -10.061
▁供 -10.0612
▁从小 -10.0613
▁随便 -10.0614
▁系 -10.0617
股 -10.0619
▁有时 -10.0622
▁广 -10.0623
▁一场 -10.0624
狗 -10.0631
晓 -10.0643
▁孤独 -10.0649
▁耐 -10.0656
▁令人 -10.0657
▁细 -10.0657
▁地址 -10.0658
府 -10.0659
▁轻松 -10.066
维 -10.0663
▁美术 -10.0664
▁严格 -10.067
佳 -10.0674
▁整 -10.0676
▁巨 -10.0681
真 -10.0681
能力 -10.0682
▁怕 -10.0696
▁下载 -10.0697
▁普遍 -10.0699
▁抗 -10.0702
资 -10.0705
▁提问 -10.0715
▁改善 -10.072
▁.... -10.0725
▁睡觉 -10.0729
▁升 -10.0745
▁比例 -10.0748
▁内存 -10.0752
服务 -10.0758
▁功 -10.0765
al -10.0766
旨 -10.0768
迷 -10.0773
▁破 -10.0773
▁有着 -10.0774
▁插 -10.0776
▁齐 -10.0777
▁武 -10.0782
因 -10.0783
识 -10.0784
▁困难 -10.079
试 -10.0792
▁职 -10.0818
▁预 -10.0821
▁美丽 -10.0821
▁独特 -10.0822
▁温度 -10.0823
▁独 -10.0826
剑 -10.0826
▁炒 -10.0832
▁外国 -10.0833
苗 -10.0853
▁进一步 -10.0855
▁素质 -10.0858
▁篇 -10.0877
▁奥 -10.0878
▁皆 -10.0879
▁联 -10.0886
铺 -10.0887
▁亚 -10.0887
▁讨厌 -10.0888
▁2015 -10.0896
壁 -10.0897
共 -10.0899
▁# -10.0902
▁打算 -10.0903
送 -10.0907
▁列 -10.0907
▁草 -10.091
▁生长 -10.0925
▁同意 -10.093
▁运用 -10.0931
破 -10.0935
顺 -10.0946
▁魔 -10.0948
▁排名 -10.0958
▁答主 -10.0958
委 -10.0961
▁孙 -10.097
惠 -10.0976
▁观察 -10.0979
▁合同 -10.0984
面前 -10.099
▁输出 -10.0993
查 -10.0999
▁德 -10.1
环 -10.1007
广 -10.1007
▁怪 -10.1014
▁尽量 -10.102
▁细胞 -10.102
▁真心 -10.1035
▁人体 -10.1036
▁骗 -10.1046
▁U -10.1047
锅 -10.1049
▁抽 -10.1052
▁进步 -10.1065
▁组合 -10.1065
▁QQ -10.1067
▁可爱 -10.1068
▁画面 -10.1076
周 -10.1077
▁电视剧 -10.1091
▁圆 -10.1092
▁西方 -10.11
▁打造 -10.1112
千 -10.112
▁45 -10.1125
序 -10.1131
█ -10.1133
▁情感 -10.114
击 -10.114
▁居住 -10.1141
▁同一 -10.1143
▁引 -10.1147
▁等级 -10.1155
▁兴 -10.1164
切 -10.118
▁商务 -10.118
▁广东 -10.118
▁空气 -10.1189
▁河 -10.1193
▁对比 -10.1217
学校 -10.1221
▁群众 -10.1228
▁中华 -10.1233
▁毫米 -10.1244
游 -10.1252
▁相互 -10.1258
溪 -10.1262
▁联合 -10.1268
▁莫 -10.1275
▁浙江 -10.1279
退 -10.1279
税 -10.1286
靠 -10.1288
▁突出 -10.1292
▁药物 -10.1296
日期 -10.1297
▁折 -10.1309
▁施工 -10.1311
鸡 -10.1311
▁1999 -10.1311
▁四川 -10.1313
▁机关 -10.1313
▁奇 -10.1318
▁手段 -10.1318
腾 -10.1319
▁剧 -10.1323
▁感动 -10.1326
▁景 -10.1327
▁迅速 -10.1329
性质 -10.1337
▁天下 -10.1338
▁不足 -10.1346
伦 -10.1348
▁试验 -10.1373
on -10.1374
▁巴 -10.1376
伙伴 -10.1381
▁爱好 -10.1387
● -10.1387
▁配置 -10.1393
▁第八 -10.1395
钟 -10.1399
▁贵 -10.1402
▁不算 -10.1404
▁便宜 -10.1415
▁更是 -10.1417
▁物流 -10.1421
▁兄弟 -10.1425
▁身份 -10.143
▁包装 -10.1437
▁抢 -10.1438
▁PS -10.1439
勒 -10.1444
补 -10.1444
▁前面 -10.1451
▁笔 -10.1455
▁吹 -10.1456
▁常用 -10.1458
in -10.146
▁上课 -10.146
▁异 -10.1466
▁模型 -10.1468
▁神经 -10.147
▁承认 -10.1472
▁粉 -10.1474
诺 -10.1478
▁大众 -10.1478
页 -10.1478
▁双方 -10.1481
▁幼儿 -10.1487
人民政府 -10.1488
▁我妈 -10.1493
h -10.1496
▁聊 -10.1498
▁飞机 -10.1503
作用 -10.1508
▁少年 -10.1509
▁谱 -10.151
▁呵呵 -10.1513
▁身高 -10.1535
尾 -10.1536
▁寻找 -10.1537
▁构成 -10.1547
▁围 -10.1547
藏 -10.1552
▁行动 -10.1553
▁春 -10.1553
▁i -10.1559
怒 -10.156
▁配合 -10.1565
▁味道 -10.1571
▁冠军 -10.1575
容 -10.1575
▁荣誉 -10.1583
▁漫画 -10.1592
▁展示 -10.1592
L -10.1603
▁高速 -10.1613
导 -10.1618
▁群 -10.1623
▁准确 -10.1623
▁权利 -10.1634
富 -10.1637
▁遭 -10.1643
脑 -10.1647
▁装修 -10.1652
▁民主 -10.1656
▁表演 -10.1661
▁坑 -10.1666
智 -10.1666
▁公众 -10.1668
孔 -10.1672
▁皮 -10.1673
▁住房 -10.1678
营 -10.1688
▁革命 -10.1688
▁听说 -10.1696
墙 -10.17
▁31 -10.1701
▁v -10.1705
▁语 -10.1705
齐 -10.1706
康 -10.1709
▁好吃 -10.1712
产业 -10.1715
▁you -10.1716
▁哈 -10.1717
▁厉 -10.1724
▁招 -10.1729
▁礼 -10.1735
z -10.1735
▁层 -10.1737
▁具备 -10.1739
▁绿 -10.1744
▁周边 -10.1745
泽 -10.1748
床 -10.1757
评 -10.177
设备 -10.1775
松 -10.1792
▁理事 -10.1797
连 -10.1797
古 -10.1798
闻 -10.1806
N -10.1812
▁有趣 -10.1814
列 -10.1818
▁有所 -10.1819
▁温 -10.182
准 -10.1827
▁生于 -10.183
▁面试 -10.1831
▁互相 -10.1835
留 -10.1837
▁饭 -10.1838
▁偷 -10.184
▁攻 -10.1844
▁如今 -10.1847
▁军 -10.1849
▁小米 -10.1856
岩 -10.1862
▁深刻 -10.1865
▁曲 -10.187
▁相同 -10.1882
余 -10.1883
▁始终 -10.1888
▁色彩 -10.1893
想起 -10.1894
▁上班 -10.1901
▁资本 -10.1901
▁完整 -10.1911
▁街 -10.1911
▁没想 -10.1911
▁消息 -10.1912
▁1998 -10.1916
▁长大 -10.1929
众 -10.1937
▁很难 -10.1948
密 -10.1951
态 -10.1953
▁说法 -10.1954
庆 -10.1961
▁熟悉 -10.1964
▁练 -10.197
▁减肥 -10.1974
▁火车 -10.1977
港 -10.1977
认 -10.198
▁郑 -10.1981
▁心中 -10.1985
▁截 -10.1997
▁嘴 -10.2001
▁麻烦 -10.2003
▁关心 -10.2006
▁顺 -10.2006
▁婚姻 -10.2016
▁客观 -10.2021
昌 -10.2022
▁模拟 -10.2023
▁爆 -10.204
▁回忆 -10.2044
▁路上 -10.2045
▁厚 -10.2051
▁演唱 -10.2057
▁用来 -10.206
▁鬼 -10.2065
▁平衡 -10.2071
▁吐 -10.2076
▁链接 -10.2076
▁批准 -10.2076
▁08 -10.2082
▁停 -10.2083
▁考研 -10.2085
▁学历 -10.2086
▁特性 -10.209
八 -10.2091
妹 -10.2094
▁郭 -10.2096
币 -10.2102
▁地理 -10.2106
脉 -10.2107
us -10.2107
▁挣 -10.2108
▁瞬间 -10.2113
征 -10.2113
议 -10.2117
▁状况 -10.2121
▁印 -10.2136
▁头发 -10.2139
▁垃圾 -10.2139
▁h -10.2143
刊 -10.2147
瑞 -10.2153
▁回复 -10.2156
▁在线 -10.2161
毒 -10.2172
▁文艺 -10.2174
▁脚 -10.2174
炮 -10.2175
▁大多 -10.2184
人数 -10.2189
沙 -10.2193
质量 -10.2193
▁宝 -10.2197
寸 -10.2203
▁配套 -10.2212
▁脱 -10.2218
▁平方公 -10.2222
▁危险 -10.2225
顶 -10.223
井 -10.2231
若 -10.2233
▁耕地 -10.2234
苦 -10.2247
略 -10.2251
▁去年 -10.2251
守 -10.2255
▁一块 -10.2258
▁拖 -10.226
▁典型 -10.2267
举 -10.227
▁四个 -10.227
▁付出 -10.2274
唱 -10.2275
恋 -10.2276
荣 -10.2276
GB -10.2279
洞 -10.2281
▁绿色 -10.2286
▁意大利 -10.2288
题目 -10.229
▁强烈 -10.2291
▁朝 -10.2295
▁举办 -10.2295
似 -10.23
▁戴 -10.2309
▁流程 -10.2309
负 -10.2323
帅 -10.2323
▁平装 -10.2325
▁前提 -10.2328
▁联盟 -10.2329
丹 -10.233
标 -10.2333
▁重庆 -10.2335
▁呈 -10.2337
▁成果 -10.2337
收 -10.2343
▁单纯 -10.2344
▁究 -10.2346
▁个性 -10.235
▁吵 -10.2358
▁vedio -10.2381
弱 -10.2387
▁发挥 -10.2387
父 -10.2391
娘 -10.2393
▁一套 -10.2396
▁胸 -10.2399
▁冰 -10.2401
▁军事 -10.2405
▁众多 -10.2411
▁十年 -10.2412
▁尽 -10.2414
▁腿 -10.2415
▁派 -10.2415
▁工 -10.2421
▁代码 -10.2424
▁舒服 -10.2426
▁品种 -10.2426
▁生存 -10.2428
系列 -10.2428
▁开放 -10.2439
▁天津 -10.2443
▁盖 -10.2444
一方面 -10.2444
水平 -10.2446
▁高校 -10.2455
▁群体 -10.246
▁资产 -10.2469
▁暗 -10.2473
▁男女 -10.2473
▁伊 -10.2475
▁成都 -10.2479
▁体重 -10.2484
▁存储 -10.2487
经济 -10.2488
▁自信 -10.2488
餐 -10.2489
▁相应 -10.2493
▁作业 -10.2495
▁写作 -10.2496
▁俱乐部 -10.2505
▁的确 -10.2505
▁跨 -10.2506
▁06 -10.2513
载 -10.2514
▁下午 -10.2514
熟 -10.2519
▁上市 -10.252
▁刺激 -10.2525
▁凭 -10.2525
支 -10.2527
不过 -10.2538
▁梁 -10.2539
▁丑 -10.2547
▁09 -10.2549
▁机制 -10.255
▁满意 -10.2555
亲 -10.2562
▁马上 -10.2564
▁煮 -10.2572
▁饮食 -10.2579
体系 -10.2579
▁品质 -10.2581
▁女孩子 -10.2582
▁官方 -10.2582
链 -10.2591
我 -10.2594
▁贡献 -10.2597
威 -10.2597
▁史 -10.2609
▁突破 -10.2618
越 -10.2625
▁装备 -10.2634
▁纠 -10.2635
▁胖 -10.2642
▁游 -10.2643
▁研究所 -10.2646
贤 -10.2648
▁07 -10.2652
▁乘 -10.2658
▁经济学 -10.2668
怪 -10.2669
▁这部 -10.267
迪 -10.2672
▁误 -10.2678
▁散 -10.2685
▁第九 -10.269
岸 -10.2695
▁一步 -10.2698
▁秦 -10.2701
▁挑战 -10.2703
or -10.2704
▁期待 -10.2706
▁早上 -10.2716
▁主席 -10.2716
域 -10.2722
▁声 -10.2726
▁风景 -10.2727
▁服装 -10.273
▁阳光 -10.2735
资源 -10.2736
教师 -10.2736
杰 -10.2744
丁 -10.2749
▁雷 -10.2753
▁蛮 -10.2754
虎 -10.2763
ly -10.2769
▁永 -10.2769
▁资格 -10.2771
缘 -10.2773
主任 -10.2775
吃 -10.2778
▁社 -10.278
六 -10.2786
作品 -10.2788
环境 -10.2791
复 -10.281
温 -10.2812
▁横 -10.2813
▁还会 -10.2817
▁03 -10.2825
▁掉 -10.2828
▁评估 -10.283
▁Z -10.283
▁演出 -10.2831
爸妈 -10.2831
伟 -10.2835
博 -10.2837
癌 -10.2843
陵 -10.2844
▁老公 -10.2844
▁登 -10.2845
吉 -10.285
▁出国 -10.2864
▁规则 -10.2873
▁铜 -10.288
顿 -10.2894
▁玻璃 -10.2895
▁05 -10.2895
半 -10.2897
▁优化 -10.2906
▁维修 -10.2914
懂 -10.2918
▁旁边 -10.292
曼 -10.292
▁老婆 -10.2931
▁明星 -10.2934
瓦 -10.2935
▁外面 -10.2936
▁矛盾 -10.2942
▁1997 -10.2942
▁智慧 -10.2944
累 -10.2944
▁收益 -10.2953
▁两人 -10.2958
▁添加 -10.2968
厚 -10.2969
爷 -10.297
▁1.5 -10.2974
▁影片 -10.2988
▁报道 -10.2992
▁雪 -10.2997
▁缺乏 -10.2998
▁办理 -10.3001
▁尴尬 -10.3006
▁和谐 -10.3006
训 -10.3012
▁心理学 -10.3015
▁采取 -10.3016
▁金属 -10.3016
▁封 -10.3024
▁允许 -10.3029
医 -10.3037
泰 -10.3043
▁策略 -10.3044
▁客 -10.3045
の -10.3047
宇 -10.305
▁反映 -10.3055
▁老人 -10.3062
盛 -10.3069
朗 -10.3073
卖 -10.3075
弹 -10.3085
▁城 -10.3086
▁论坛 -10.3087
批 -10.3089
斗 -10.309
▁布 -10.309
▁课 -10.3099
细胞 -10.3103
▁村民 -10.3103
▁地铁 -10.3106
组织 -10.3111
▁倍 -10.3115
▁印度 -10.3117
听 -10.3121
▁公园 -10.3123
▁差异 -10.3128
▁哈哈 -10.3131
▁少数 -10.315
让 -10.315
极 -10.3153
▁主管 -10.3158
▁增强 -10.3167
▁购物 -10.3177
▁休闲 -10.3177
cm -10.3188
小说 -10.3198
亡 -10.3201
黄 -10.3205
▁房地产 -10.3209
▁好奇 -10.3213
▁策划 -10.322
▁环保 -10.3223
息 -10.3225
问题 -10.3226
▁Q -10.3227
鬼 -10.3229
胎 -10.3237
▁博 -10.3241
▁吸收 -10.3242
觉 -10.3246
▁Windows -10.3246
▁一旦 -10.3246
▁必然 -10.3248
▁跑步 -10.325
▁搜 -10.3257
未 -10.3257
▁湖南 -10.3264
姆 -10.3268
▁来到 -10.3276
▁Y -10.3277
▁财政 -10.3283
娜 -10.3287
▁大师 -10.3288
彩 -10.3292
▁房间 -10.3294
▁效率 -10.3299
迹 -10.33
艳 -10.33
▁公路 -10.3302
▁当代 -10.3303
没 -10.3304
▁本质 -10.3305
▁装饰 -10.3306
▁参 -10.3309
▁流量 -10.3309
▁熟 -10.3312
▁浪费 -10.3315
▁校 -10.3315
恨 -10.3317
O -10.3318
▁昨天 -10.3329
哈 -10.3335
▁出口 -10.3339
▁佛 -10.3341
▁咖啡 -10.3344
▁压 -10.3347
▁1.2 -10.3347
▁给予 -10.3348
▁芯片 -10.3351
散 -10.3352
▁感染 -10.336
烟 -10.3364
▁累 -10.3365
庙 -10.3369
▁鲜 -10.3373
忠 -10.3376
▁黄金 -10.3379
▁导师 -10.3383
▁怀 -10.3384
▁推进 -10.3386
鸟 -10.3386
▁途 -10.3387
▁大大 -10.3387
购 -10.3389
▁否则 -10.34
▁告 -10.34
便 -10.3402
▁社交 -10.3402
▁撕 -10.3404
艺 -10.3404
▁锻炼 -10.3408
▁出场 -10.3413
卫 -10.3414
▁高等 -10.3434
▁无数 -10.3437
▁富 -10.3439
阵 -10.3444
▁推动 -10.3445
杆 -10.3451
▁歌手 -10.3461
▁全市 -10.3464
▁content -10.3465
▁有钱 -10.3467
俩 -10.347
▁鞋 -10.3475
▁救 -10.3477
启 -10.3478
印 -10.3479
▁图书馆 -10.3479
▁这句 -10.348
▁奇怪 -10.3485
▁电池 -10.3486
▁忽 -10.3493
▁丛 -10.3493
▁姐 -10.3494
疼 -10.3496
▁落 -10.3501
▁破坏 -10.3503
人口 -10.3507
梅 -10.3507
▁宇宙 -10.3513
▁传说 -10.3514
▁一片 -10.3517
▁收藏 -10.3518
▁后期 -10.3518
练 -10.3519
▁播放 -10.3524
▁千万 -10.3527
▁for -10.3531
邦 -10.3531
▁弹 -10.3534
▁机器 -10.3536
▁室内 -10.3537
▁步 -10.3542
▁单身 -10.3544
▁一群 -10.355
袭 -10.3553
▁干净 -10.3559
▁测量 -10.356
▁症 -10.356
▁怀疑 -10.3563
▁正是 -10.3564
▁彻底 -10.3566
▁杭州 -10.3571
鲁 -10.3572
▁痛 -10.3575
▁积累 -10.3586
优 -10.3586
▁杜 -10.3587
▁注重 -10.3593
故 -10.3593
▁调节 -10.3596
▁恐怖 -10.3597
▁防止 -10.3605
珠 -10.3606
▁接近 -10.361
谱 -10.361
▁记者 -10.3611
▁顾客 -10.3616
▁铁路 -10.3617
▁亚洲 -10.3618
耳 -10.3618
▁登记 -10.3623
▁黑色 -10.3633
▁四年 -10.3647
务 -10.3654
▁浏览 -10.3654
参数 -10.3658
饰 -10.3658
▁腰 -10.366
10 -10.3662
谅 -10.3663
▁腹 -10.3666
就是 -10.3676
倍 -10.3681
罪 -10.3683
12 -10.3684
壳 -10.3684
▁02 -10.3687
▁多人 -10.3696
▁---- -10.3701
▁武器 -10.3702
▁支 -10.3713
▁块 -10.3718
▁河南 -10.3726
▁动态 -10.3726
钢 -10.3736
▁姜 -10.3738
刚 -10.3739
洛 -10.375
振 -10.375
ar -10.3753
J -10.3754
▁04 -10.3759
▁再也 -10.3765
▁协议 -10.3773
己 -10.3774
▁项 -10.3775
▁俄罗斯 -10.3778
▁仅仅 -10.3779
▁发达 -10.3781
▁此时 -10.3788
▁闹 -10.379
显 -10.379
▁思路 -10.3795
▁1996 -10.3795
▁提前 -10.3796
▁板 -10.3798
籍 -10.3802
▁150 -10.3803
▁泡 -10.3808
▁纪念 -10.3817
▁赢 -10.3818
吧 -10.3821
▁相处 -10.3823
▁江苏 -10.3823
▁艾 -10.3823
▁毫无 -10.3829
科学 -10.383
『 -10.3845
▁急 -10.3846
盐 -10.3852
▁后悔 -10.3853
▁作家 -10.3855
▁秘书 -10.3863
制度 -10.3871
▁频率 -10.3872
▁食 -10.3876
▁恶 -10.3879
▁白色 -10.3881
et -10.3886
▁400 -10.3887
防 -10.3887
▁渠道 -10.3899
材料 -10.39
世纪 -10.39
▁不得不 -10.3904
毕业 -10.3904
es -10.3905
短 -10.3907
▁zhihu -10.3912
▁高三 -10.3914
买 -10.3914
律 -10.3914
▁尽管 -10.3921
▁规律 -10.3922
▁机场 -10.3927
▁审计 -10.3927
▁音 -10.3936
幅 -10.3936
▁玉 -10.3939
▁几次 -10.3941
▁120 -10.3943
规划 -10.3946
▁挑 -10.3953
魂 -10.3953
▁属性 -10.3955
▁理财 -10.3957
▁病人 -10.3964
▁意味 -10.3966
▁从此 -10.3972
键 -10.3975
▁放入 -10.3986
潮 -10.3989
湾 -10.3989
▁CPU -10.3992
』 -10.3994
方案 -10.3999
▁设立 -10.4002
▁敏感 -10.4014
超 -10.402
▁供应 -10.4025
▁宗 -10.4029
▁显得 -10.4029
棒 -10.403
书记 -10.4032
▁物业 -10.4035
赞 -10.4038
岭 -10.404
银行 -10.4041
▁东方 -10.4044
▁格式 -10.4046
▁网友 -10.4055
方法 -10.406
▁蛋白 -10.407
容量 -10.4074
▁诸 -10.4079
▁民间 -10.4083
▁母 -10.4086
城市 -10.4092
▁键盘 -10.4096
皇帝 -10.41
▁塔 -10.4103
▁鼓励 -10.4109
▁轴承 -10.411
▁女子 -10.4112
▁荣获 -10.412
策 -10.4121
Q -10.4122
造 -10.4125
▁消失 -10.4125
▁警察 -10.4126
▁强调 -10.4127
▁股票 -10.4129
提 -10.4132
E -10.4132
▁理性 -10.4141
▁聪明 -10.4144
▁保修 -10.4159
灭 -10.416
▁一辈子 -10.4174
叔 -10.4177
▁举 -10.4178
▁信用 -10.4181
改 -10.4184
ー -10.4184
▁劝 -10.4184
▁高效 -10.4198
▁普 -10.4199
▁肌肉 -10.42
除 -10.42
▁形态 -10.4202
露 -10.4218
j -10.4221
▁松 -10.4221
宜 -10.4223
塘 -10.4223
▁书法 -10.4225
呼 -10.4233
▁背后 -10.4234
幕 -10.4245
▁月份 -10.426
▁十五 -10.4261
▁通知 -10.4265
▁一大 -10.4275
▁令 -10.4281
鉴 -10.4287
专业 -10.4288
▁镜像 -10.4302
访 -10.4304
淡 -10.4307
活动 -10.4309
汁 -10.4313
▁森林 -10.4317
▁航空 -10.4317
▁...... -10.4319
▁两次 -10.4319
▁镜头 -10.4321
▁小区 -10.4334
▁诊断 -10.4337
▁拟 -10.434
▁本地 -10.4341
▁隔 -10.4344
泪 -10.4347
招 -10.4348
模 -10.4351
▁译 -10.4355
两 -10.4357
▁毒 -10.4363
▁措施 -10.437
▁拜 -10.4371
▁预防 -10.4373
付 -10.4375
塞 -10.4376
▁十一 -10.4377
▁通信 -10.4381
▁紫 -10.4382
针 -10.4383
▁生日 -10.439
▁影视 -10.4399
en -10.4401
▁剑 -10.4402
▁忘记 -10.4409
安全 -10.4423
▁下降 -10.4424
▁端 -10.4427
▁动漫 -10.4431
▁较大 -10.4433
▁武汉 -10.4433
▁公元 -10.4434
▁办学 -10.4437
▁占地 -10.4439
▁律师 -10.444
▁缺 -10.4441
▁看来 -10.4445
齿 -10.4454
▁陌生 -10.4463
▁灰 -10.4466
▁举行 -10.4473
▁运输 -10.4475
▁分割 -10.4479
▁明明 -10.4484
坏 -10.449
▁呼吸 -10.4491
▁困 -10.4497
▁设定 -10.4502
R -10.4503
▁抄 -10.4506
▁定价 -10.4509
▁不然 -10.451
▁价值观 -10.451
▁鼻 -10.4511
▁藏 -10.4511
▁收费 -10.4512
绝 -10.4521
▁载 -10.4522
▁委员 -10.4522
▁西安 -10.4525
▁汉族 -10.4538
▁1995 -10.4544
▁一体 -10.4544
▁相似 -10.4546
▁精彩 -10.455
▁探索 -10.4552
▁猪 -10.4552
嘉 -10.4553
▁瘦 -10.4554
▁元素 -10.4576
▁分辨 -10.4583
▁依法 -10.4584
▁标志 -10.4588
顾 -10.459
误 -10.4591
▁没人 -10.4595
▁穷 -10.4598
▁引导 -10.4607
叫 -10.4612
十 -10.4614
细 -10.4618
▁照顾 -10.4631
▁无限 -10.4635
▁浅 -10.4638
▁乌 -10.4639
▁曹 -10.4641
▁保存 -10.4642
敏 -10.4649
▁检验 -10.466
▁循环 -10.4662
▁梅 -10.467
貌 -10.4675
― -10.4681
▁部队 -10.4685
▁编程 -10.4685
▁question -10.4686
▁社会主义 -10.4687
▁世 -10.4688
▁江 -10.4691
▁干部 -10.4693
文化 -10.4694
▁陆 -10.4701
订 -10.4709
st -10.4709
▁维 -10.4718
九 -10.4722
▁辖 -10.4725
▁2013 -10.4729
it -10.473
▁长得 -10.4731
▁产 -10.4732
▁时刻 -10.4733
粒 -10.4734
琴 -10.4736
▁赤 -10.4738
▁蒙 -10.4738
▁总体 -10.4738
▁天然 -10.4742
▁波 -10.4748
▁半年 -10.4751
▁贷款 -10.4752
▁认可 -10.4756
▁.. -10.4776
▁搭配 -10.4777
▁星 -10.4779
▁少女 -10.478
▁烤 -10.4782
济 -10.4782
坐 -10.4783
普 -10.4785
▁小组 -10.4785
袋 -10.4785
▁西南 -10.48
▁甚 -10.4801
▁体会 -10.4806
▁输 -10.4806
慢 -10.4808
▁外观 -10.4813
航 -10.4815
▁广场 -10.4817
▁核 -10.4817
档 -10.4818
隆 -10.4822
材 -10.4829
豆 -10.4832
▁硬盘 -10.4833
▁欣赏 -10.4835
▁工程师 -10.4836
归 -10.4837
播 -10.4843
沟 -10.4844
犯 -10.4845
裂 -10.4849
▁犯罪 -10.4852
宾 -10.4852
▁闲 -10.4855
▁划 -10.4862
姓 -10.4865
▁主人 -10.4868
▁兼职 -10.4869
▁集体 -10.4872
▁必 -10.4873
16 -10.4879
▁休息 -10.4881
▁指挥 -10.4885
信息 -10.4893
▁胃 -10.49
▁彼此 -10.4904
骤 -10.4906
裤 -10.4911
薄 -10.4913
▁基因 -10.4917
▁弱 -10.4917
▁离婚 -10.492
登 -10.4923
▁课题 -10.4934
▁五年 -10.4934
研 -10.4934
肯 -10.4937
腿 -10.4937
▁废 -10.4942
成果 -10.4943
▁运 -10.4944
枝 -10.4946
▁示范 -10.4946
▁貌似 -10.4948
▁地下 -10.495
尘 -10.4953
▁一半 -10.4954
▁好处 -10.4957
▁伟大 -10.4957
▁摄像 -10.4962
创 -10.4964
永 -10.4964
▁爷爷 -10.4966
▁嘉 -10.4972
丰 -10.4973
▁创意 -10.4975
▁一致 -10.4975
▁早期 -10.498
▁问问 -10.498
▁甜 -10.498
必 -10.4985
亩 -10.4999
▁公务员 -10.5003
Z -10.5004
▁公平 -10.5009
绕 -10.5014
▁出发 -10.5016
▁假设 -10.5018
▁编 -10.5027
▁经理 -10.5031
▁贸易 -10.5033
ia -10.5043
旗 -10.5044
换 -10.5045
▁标 -10.506
▁福 -10.5061
th -10.5062
瓶 -10.5064
▁一路 -10.5065
▁降 -10.5077
所以 -10.5083
芳 -10.5086
▁国产 -10.5089
▁扔 -10.5094
▁生气 -10.5098
▁温暖 -10.5098
▁消 -10.5099
▁团体 -10.5099
社会 -10.5099
▁辅助 -10.5101
▁电力 -10.5101
▁海拔 -10.5103
秋 -10.5108
炉 -10.5109
▁盛 -10.5115
▁提醒 -10.5115
▁气质 -10.5115
▁西北 -10.5117
▁夏天 -10.5121
欲 -10.5125
▁人文 -10.5129
▁情节 -10.5134
▁分配 -10.5147
▁重复 -10.5149
祥 -10.5153
▁节奏 -10.5156
▁只会 -10.5157
庭 -10.5157
▁渣 -10.5158
▁一座 -10.5161
▁滚 -10.5164
▁脑 -10.5167
▁硬件 -10.5171
▁敢 -10.5179
▁原料 -10.5179
▁难度 -10.518
▁罢 -10.518
屋 -10.5181
▁服务器 -10.5182
▁仿 -10.5182
静 -10.5187
▁并非 -10.5189
▁舍 -10.5191
▁书籍 -10.5193
le -10.5198
坑 -10.5198
▁著 -10.5199
▁刀 -10.52
▁树 -10.5201
软件 -10.5202
▁缺点 -10.521
▁搬 -10.5212
▁负 -10.5222
▁略 -10.5224
刑 -10.5227
标准 -10.5242
▁心灵 -10.5243
▁无聊 -10.5245
▁省级 -10.525
▁书记 -10.5252
响 -10.5255
▁断 -10.5255
总 -10.5257
▁一把 -10.5257
▁..... -10.5262
▁热情 -10.5271
与 -10.5274
▁有用 -10.528
▁实例 -10.5285
实话 -10.5287
▁湖 -10.5289
▁抑郁 -10.5289
韵 -10.529
▁出门 -10.5296
▁吓 -10.5297
锋 -10.5297
▁北方 -10.5298
▁上学 -10.53
农 -10.5301
▁云南 -10.5306
氧 -10.5306
▁所属 -10.5307
▁夹 -10.5307
▁紧张 -10.5311
▁观念 -10.5315
慕 -10.5316
▁不让 -10.5317
▁肾 -10.5318
幸 -10.5322
▁意外 -10.5328
扎 -10.5329
▁养殖 -10.5329
▁腾讯 -10.5332
坛 -10.5333
▁智商 -10.5336
▁魏 -10.534
▁环 -10.5352
▁事物 -10.5354
▁剪 -10.5354
▁飞行 -10.5361
▁扩大 -10.5363
▁摆 -10.5367
恼 -10.5371
医院 -10.5373
▁嫌 -10.5374
湿 -10.5379
猫 -10.538
▁遇 -10.5381
男 -10.5383
盖 -10.5389
敌 -10.539
▁祝 -10.5394
▁指标 -10.5408
▁招聘 -10.5409
折 -10.5409
▁再说 -10.543
▁种类 -10.5436
冷 -10.5437
他人 -10.5441
背 -10.5443
▁凡 -10.5443
▁转移 -10.5446
▁信任 -10.5452
▁现状 -10.5459
▁第十 -10.5459
▁1994 -10.5462
▁随时 -10.5463
▁各地 -10.5464
仙 -10.5465
卫生 -10.5469
假 -10.5472
蒂 -10.5482
▁维持 -10.5486
翼 -10.5489
▁明天 -10.5489
菲 -10.549
▁扩展 -10.5493
▁美容 -10.5496
▁复习 -10.5499
▁3000 -10.5503
▁删 -10.5505
▁军队 -10.5508
▁互动 -10.5508
▁成分 -10.5511
▁信仰 -10.5514
▁歌词 -10.5515
机关 -10.5516
七 -10.5517
▁难过 -10.5519
▁家族 -10.5522
▁收集 -10.553
▁近年 -10.553
▁居民 -10.5531
▁造型 -10.5536
署 -10.5544
id -10.5544
▁依据 -10.5551
▁烟 -10.5553
▁级别 -10.5553
聚 -10.5554
▁沿 -10.5556
舍 -10.5556
▁性别 -10.5566
在于 -10.5569
▁西班牙 -10.5569
▁中华人民共和国 -10.5572
▁800 -10.5573
▁街道 -10.5573
▁走向 -10.5577
▁集成 -10.558
跑 -10.5584
▁年纪 -10.5593
▁最低 -10.5595
▁iPhone -10.5595
▁中有 -10.5595
▁混合 -10.5595
▁利润 -10.5609
责任 -10.5617
▁主体 -10.5619
▁空调 -10.5628
冠 -10.5636
▁清晰 -10.5638
▁宣 -10.564
▁寄 -10.5648
▁单元 -10.5651
▁1993 -10.5653
▁致力 -10.5659
▁暂时 -10.5659
辉 -10.566
▁团 -10.5661
▁命运 -10.5662
▁逛 -10.5669
▁刺 -10.567
▁南方 -10.5672
▁吸 -10.5675
纪 -10.5676
11 -10.5682
腔 -10.5686
▁暴 -10.5686
▁吉 -10.5691
̄ -10.5695
el -10.5705
▁保留 -10.5716
患 -10.5719
▁高端 -10.5721
▁烦 -10.5722
▁同志 -10.5727
鼠 -10.5731
▁app -10.5733
▁印刷 -10.5733
午 -10.5734
▁超市 -10.5738
▁乃 -10.5738
▁预测 -10.5741
▁提示 -10.5745
▁耳机 -10.5748
协会 -10.575
▁舒适 -10.5751
▁各自 -10.5753
▁自主 -10.5754
▁is -10.5758
▁覆盖 -10.5761
▁临 -10.5762
▁红色 -10.5764
▁36 -10.5766
磨 -10.5771
▁鼠标 -10.5775
▁上网 -10.5778
▁编写 -10.5779
T -10.5779
ta -10.5781
▁岗位 -10.5783
夏 -10.5783
▁整理 -10.579
菌 -10.5791
▁佛教 -10.58
▁秘密 -10.5801
竹 -10.5801
▁确认 -10.5802
频 -10.5803
▁蔬菜 -10.5807
妇 -10.5809
is -10.581
▁哥哥 -10.5812
▁启动 -10.5818
高兴 -10.5825
▁w -10.5828
▁广西 -10.5829
▁拼 -10.5829
持 -10.5829
洋 -10.5833
▁司机 -10.5836
▁扯 -10.5837
▁某种 -10.584
・ -10.5841
▁大约 -10.5849
▁对外 -10.585
护 -10.5852
▁恶心 -10.5853
讲 -10.5861
▁车辆 -10.5863
.8 -10.5866
́ -10.5866
▁在家 -10.5868
▁表情 -10.5874
` -10.5875
▁访问 -10.5875
鼓 -10.5876
▁潮 -10.5885
▁顺利 -10.5885
▁ps -10.5896
谓 -10.5898
▁图像 -10.5906
▁无线 -10.5909
▁造 -10.5926
▁依旧 -10.5926
▁妻子 -10.593
喜 -10.5935
▁获取 -10.5937
▁股份 -10.595
▁函数 -10.5953
▁魅力 -10.5954
▁数据库 -10.5955
▁各项 -10.5955
▁脂肪 -10.5955
▁发出 -10.5956
柜 -10.596
▁差别 -10.5962
▁温柔 -10.5965
▁多数 -10.5965
萨 -10.5968
▁污染 -10.5971
▁首次 -10.5978
▁内置 -10.5984
▁篮球 -10.5988
予 -10.599
▁二十 -10.5991
奥 -10.5994
▁碰到 -10.6002
▁2.5 -10.6006
羊 -10.601
▁记载 -10.601
轴 -10.6014
▁悲 -10.6015
▁源 -10.6015
寒 -10.6017
▁狂 -10.6019
▁限 -10.602
▁小心 -10.6023
▁d -10.6025
▁热爱 -10.6027
▁预算 -10.6027
▁晋 -10.6028
▁星期 -10.6028
京 -10.6028
▁仔细 -10.603
▁人工 -10.6032
▁亲戚 -10.6035
MB -10.6038
▁仙 -10.6039
▁展开 -10.6044
素质 -10.6047
太多 -10.6048
▁1992 -10.6048
▁戏 -10.6048
▁创 -10.605
▁合法 -10.6051
▁浓 -10.6056
▁队伍 -10.6061
▁地震 -10.6067
碎 -10.6068
▁2.2 -10.6068
▁概率 -10.6075
▁碰 -10.6077
▁蔡 -10.6077
▁解放 -10.6081
▁皇 -10.609
▁秋 -10.6104
▁农 -10.6106
▁司法 -10.6107
▁进口 -10.6112
▁一流 -10.612
教授 -10.612
▁没事 -10.612
▁没法 -10.6123
▁对手 -10.6124
虚 -10.6125
▁骑 -10.6134
▁亮 -10.6138
▁查看 -10.6138
▁教练 -10.6138
▁趋势 -10.6139
▁保健 -10.6142
▁决策 -10.6149
▁人士 -10.615
▁海洋 -10.6151
▁不必 -10.6156
▁USB -10.6158
▁密码 -10.6158
▁甲 -10.6165
▁认同 -10.6168
▁浮 -10.6173
▁欺 -10.6174
行业 -10.6176
▁个体 -10.6177
▁证据 -10.618
▁档案 -10.6188
▁法规 -10.6194
▁喊 -10.6196
▁授予 -10.6196
▁l -10.62
肠 -10.6205
▁躺 -10.6206
▁董事 -10.6207
▁丁 -10.6214
▁可靠 -10.6217
▁擦 -10.6218
▁复合 -10.6224
▁那时 -10.6227
▁认知 -10.623
▁一对 -10.624
▁厂商 -10.625
▁接下 -10.6251
▁辅导 -10.6252
▁分子 -10.6253
▁通讯 -10.6256
▁依靠 -10.6258
▁擅长 -10.6263
▁肝 -10.6263
▁虚拟 -10.6271
阁 -10.6272
洲 -10.6272
▁规格 -10.6273
意识 -10.6283
▁新疆 -10.6285
险 -10.6286
▁妹妹 -10.6292
▁竹 -10.6293
▁小姐 -10.6294
▁优惠 -10.6298
▁鸡蛋 -10.6299
▁主角 -10.6302
凤 -10.6306
▁改造 -10.6308
俊 -10.631
人士 -10.631
▁欲 -10.6315
▁转换 -10.6316
▁中共 -10.6319
▁删除 -10.6319
婚 -10.632
民币 -10.6322
▁作出 -10.6324
健 -10.6326
▁疯狂 -10.6326
▁神秘 -10.6327
▁确保 -10.6327
▁贝 -10.6331
▁雨 -10.6336
▁建成 -10.6338
▁2014 -10.6338
▁灵魂 -10.6339
▁男子 -10.6359
▁赞 -10.6364
▁结论 -10.6365
▁视觉 -10.6367
▁寺 -10.6367
救 -10.6368
▁击 -10.6368
▁胜利 -10.6369
扰 -10.6371
▁适当 -10.6381
梁 -10.6384
▁流动 -10.6384
堡 -10.6387
▁损失 -10.6389
▁查询 -10.6392
▁人性 -10.6399
▁生理 -10.64
▁海外 -10.6401
▁嫁 -10.6404
▁一级 -10.6405
▁同性恋 -10.6408
凉 -10.6415
▁乙 -10.6421
▁即将 -10.6422
▁一项 -10.6424
▁o -10.6431
▁上下 -10.6434
▁港 -10.6435
▁精力 -10.644
▁范 -10.6443
▁开车 -10.6447
▁尚 -10.6451
▁外地 -10.6455
▁北京市 -10.6476
▁奖励 -10.6487
14 -10.6494
▁肩 -10.6494
▁灵 -10.6494
▁一颗 -10.6499
▁停止 -10.65
▁果 -10.6502
▁当前 -10.6503
仅仅 -10.6504
殿 -10.6509
▁倾向 -10.651
莱 -10.6512
▁货币 -10.6516
将来 -10.6518
▁故障 -10.6518
.6 -10.6519
▁维生素 -10.6527
操 -10.6527
你 -10.6528
▁黑暗 -10.6532
▁适量 -10.6533
▁平面 -10.6536
谢 -10.6539
▁一篇 -10.6542
嘴 -10.6544
▁工厂 -10.6547
▁对面 -10.6548
▁75 -10.655
an -10.6563
稿 -10.6563
▁深度 -10.6567
▁有机 -10.6567
获 -10.6573
▁等待 -10.6574
▁防治 -10.6575
即 -10.6576
矿 -10.6577
▁发明 -10.6578
▁层次 -10.6586
▁著作 -10.6586
u -10.6589
▁驻 -10.6593
巧 -10.6593
▁600 -10.6594
▁欧 -10.6597
▁无奈 -10.6602
▁奇葩 -10.6604
▁须 -10.6604
刷 -10.6608
▁审核 -10.6608
▁望 -10.6609
▁证券 -10.661
▁同步 -10.661
▁一批 -10.6611
▁异常 -10.6612
▁太多 -10.6614
▁默默 -10.6615
▁引进 -10.6616
▁地面 -10.6618
美元 -10.6623
▁上述 -10.6624
亭 -10.6627
▁美女 -10.6628
▁墙 -10.6632
尖 -10.6633
▁自学 -10.6637
▁不久 -10.6637
薪 -10.664
▁反复 -10.6641
▁免疫 -10.665
锁 -10.6654
▁接着 -10.6654
▁主流 -10.6657
核 -10.6658
臂 -10.6661
封 -10.6666
▁姓名 -10.6666
▁大力 -10.6671
雄 -10.6678
▁学者 -10.6679
▁治 -10.6682
▁试试 -10.6683
▁自卑 -10.6694
▁便利 -10.6696
▁入门 -10.67
童 -10.6704
示 -10.6708
▁构 -10.6709
▁麦 -10.6711
▁苏联 -10.6715
▁车站 -10.6716
▁舞台 -10.6716
扣 -10.6717
▁In -10.6725
▁美食 -10.6726
▁把握 -10.6737
▁见面 -10.6738
48 -10.6739
▁广东省 -10.6744
▁熊 -10.6748
▁病毒 -10.6751
▁总统 -10.6752
▁魔法 -10.6757
▁cm -10.6759
▁沈 -10.6768
▁上升 -10.6769
▁原本 -10.6773
▁优越 -10.6775
啥 -10.6784
▁权力 -10.6785
学习 -10.6789
▁佩 -10.6789
文献 -10.6795
测 -10.6798
▁音频 -10.68
▁河北 -10.6801
▁不怎么 -10.6812
射 -10.6814
▁撒 -10.6821
▁咬 -10.6821
▁支撑 -10.6829
▁编制 -10.6832
粹 -10.6832
趣 -10.6839
▁38 -10.6841
▁书画 -10.6842
纹 -10.6846
▁柳 -10.6846
▁共有 -10.6847
柱 -10.6847
▁和平 -10.6848
▁网页 -10.685
▁虚 -10.6852
▁电子商务 -10.6852
▁喂 -10.6856
▁发动机 -10.6858
窗 -10.6861
▁1990 -10.6861
▁护理 -10.6864
▁水果 -10.687
▁忙 -10.6874
▁展 -10.6875
▁西部 -10.6875
▁苏州 -10.6875
▁授权 -10.6876
▁整合 -10.6876
▁打击 -10.6876
▁朝鲜 -10.6879
一下 -10.6883
繁 -10.6883
协 -10.6887
▁船 -10.6897
劲 -10.69
▁奔 -10.6901
▁顺便 -10.6903
▁命名 -10.6904
▁糖 -10.6904
▁没错 -10.691
▁呆 -10.6911
勇 -10.6915
损 -10.6915
▁随意 -10.6916
睡 -10.6924
哭 -10.6932
▁餐 -10.6933
▁钢 -10.6936
▁寝 -10.6938
▁尿 -10.6942
▁没用 -10.6943
▁寒 -10.6943
苏 -10.6944
电视台 -10.6944
冲 -10.6952
▁1.1 -10.6954
▁就行 -10.6955
▁妖 -10.696
攻 -10.6961
▁歧视 -10.6971
▁这时 -10.6978
部分 -10.6978
▁找个 -10.698
▁爆发 -10.6987
▁极大 -10.6987
▁免 -10.6993
▁到达 -10.6997
运动 -10.7002
▁大脑 -10.7005
▁移民 -10.7005
▁收录 -10.701
▁慢 -10.701
▁税 -10.7018
▁除非 -10.7024
▁校长 -10.7025
▁权威 -10.7031
▁de -10.7033
▁财富 -10.7035
▁装置 -10.7036
轻 -10.7037
▁自杀 -10.704
▁舞蹈 -10.7041
▁社团 -10.7042
▁手动 -10.7045
▁超越 -10.7046
▁材质 -10.7047
ad -10.7049
俗 -10.705
▁笔记本 -10.7056
▁抵 -10.7063
▁指南 -10.7064
▁路线 -10.7069
纲 -10.707
▁疼 -10.7072
胶 -10.7072
▁模块 -10.708
▁石油 -10.7082
▁安徽 -10.7092
廷 -10.7095
润 -10.7095
▁莫名 -10.7096
▁me -10.7099
不下 -10.7102
▁专利 -10.7104
▁珍 -10.7105
▁爆炸 -10.7106
▁办事 -10.7108
▁收获 -10.7109
▁地上 -10.7109
▁执 -10.7111
▁神奇 -10.7113
▁田 -10.7114
▁不停 -10.7115
▁根 -10.7119
▁修复 -10.712
▁--- -10.7122
▁对话 -10.7123
▁磨 -10.7124
▁冯 -10.7128
▁玉米 -10.7132
脂 -10.7133
▁一两 -10.7134
▁一面 -10.7135
▁1988 -10.714
▁身材 -10.7141
不好 -10.7143
泄 -10.7143
▁统治 -10.7145
▁采购 -10.7147
利于 -10.7149
▁院校 -10.7152
自己 -10.7156
▁界面 -10.7158
▁就读 -10.716
▁批评 -10.7161
▁奶 -10.7161
烈 -10.7165
▁恐惧 -10.7169
资料 -10.7169
▁专用 -10.717
▁证书 -10.7173
▁异地 -10.7175
▁继承 -10.7182
同学 -10.7184
▁鉴定 -10.7186
▁实体 -10.7193
责 -10.7206
▁康 -10.7207
▁融 -10.721
宋 -10.7211
▁交往 -10.7214
▁逐步 -10.7216
▁愤 -10.7216
▁准 -10.7218
▁欧美 -10.7224
▁宁 -10.7224
▁境内 -10.7228
▁直播 -10.7233
拍 -10.7233
▁乡村 -10.7234
▁公布 -10.7238
程序 -10.7239
论文 -10.7239
▁违法 -10.7242
姨 -10.7243
▁33 -10.7247
▁2.1 -10.7248
▁一点点 -10.7253
模式 -10.7258
▁加油 -10.726
▁穿着 -10.7262
▁前后 -10.7265
置 -10.7266
不懂 -10.7267
▁宫 -10.7267
▁日语 -10.7268
寨 -10.7271
▁公主 -10.7274
仓 -10.7275
▁协助 -10.7278
▁袁 -10.7282
审 -10.7285
兹 -10.7288
▁对应 -10.7297
▁东南 -10.7302
▁言论 -10.7307
洁 -10.7309
降 -10.7316
▁一堆 -10.7324
▁半个 -10.7329
▁领先 -10.733
尺 -10.7336
战争 -10.7336
洪 -10.7338
▁新鲜 -10.7338
▁格 -10.7338
▁清洁 -10.7341
▁密 -10.7341
Y -10.7349
▁合并 -10.735
▁爬 -10.7353
▁脑子 -10.7356
▁诚信 -10.7356
承 -10.7356
▁罪 -10.736
▁终 -10.7364
▁博物馆 -10.7374
财 -10.7377
▁旗 -10.7377
▁展现 -10.7382
兽 -10.7386
帮 -10.739
▁1.3 -10.7391
▁湖北 -10.7392
伐 -10.7393
开发 -10.7398
▁顾问 -10.7399
汗 -10.7399
▁院长 -10.74
▁物品 -10.7401
▁电路 -10.7401
▁塞 -10.7403
▁掌 -10.7411
玛 -10.7413
▁愿 -10.7413
迷茫 -10.7415
▁强化 -10.7419
▁助 -10.7422
津 -10.7422
▁改进 -10.7426
址 -10.7429
伊 -10.7433
▁丰 -10.7434
▁可见 -10.7435
逸 -10.7437
颜 -10.7443
▁加拿大 -10.7448
▁疏 -10.7449
爽 -10.7451
▁福建 -10.7451
▁英文名 -10.7454
▁总会 -10.7459
▁合成 -10.7461
▁分数 -10.7463
い -10.7464
▁原始 -10.7469
▁商家 -10.7471
意义 -10.7471
▁杯 -10.748
▁拆 -10.748
▁摔 -10.7495
▁37 -10.7497
鸣 -10.7508
kg -10.7511
17 -10.7512
▁严 -10.7514
圣 -10.7518
▁IT -10.7522
▁鸟 -10.7523
▁驾驶 -10.7523
零 -10.7531
舰 -10.7538
盒 -10.7543
▁女神 -10.7545
▁框架 -10.7545
▁Re -10.7551
▁一线 -10.7555
▁大二 -10.756
处理 -10.7565
▁舒 -10.7566
ion -10.7568
耀 -10.7574
蒙 -10.7576
半年 -10.7579
▁讲述 -10.7583
▁钢琴 -10.7585
烂 -10.7586
检 -10.7587
▁原创 -10.7589
阴 -10.7594
▁赢得 -10.7594
▁卢 -10.7596
▁公益 -10.7601
▁天气 -10.7601
▁值 -10.7606
▁摸 -10.7608
▁1991 -10.7608
ation -10.7611
希 -10.7612
▁瞎 -10.7614
乘 -10.7615
▁一周 -10.7616
▁一支 -10.7618
▁5000 -10.7619
▁化工 -10.7621
拳 -10.7622
▁margin -10.7623
▁奶奶 -10.7625
来讲 -10.7626
▁三国 -10.7627
▁分离 -10.7629
▁逃 -10.7631
技 -10.7633
麻 -10.7638
▁正好 -10.7639
▁寻 -10.764
▁感兴趣 -10.7643
▁十八 -10.7652
▁国务院 -10.7654
▁辛苦 -10.7656
▁谈恋 -10.7658
▁类别 -10.7661
肿 -10.7661
▁山东省 -10.7662
▁遗 -10.7666
▁透明 -10.7668
▁64 -10.7668
▁肺 -10.7669
▁过于 -10.7669
▁假如 -10.7672
13 -10.7673
▁交换 -10.7678
▁塑料 -10.768
▁贾 -10.7682
▁55 -10.7686
▁人人 -10.7687
▁粮食 -10.7688
氛 -10.769
发展 -10.7692
▁游泳 -10.7696
▁彩 -10.7705
▁般 -10.7706
▁像素 -10.7707
▁欲望 -10.7711
冰 -10.7716
坡 -10.7719
缺 -10.7722
桌 -10.7725
▁牌子 -10.7727
▁豆瓣 -10.7733
▁致 -10.7736
▁取消 -10.7742
▁伤 -10.7742
贝 -10.7744
大战 -10.7745
▁基层 -10.7747
▁常务 -10.7747
▁探讨 -10.7749
慧 -10.775
▁监控 -10.7758
▁暴力 -10.776
▁耳 -10.7767
▁算法 -10.7771
24 -10.7771
阀 -10.7772
▁守 -10.7773
▁拍照 -10.7783
益 -10.7787
▁饰 -10.7789
▁专科 -10.779
▁槽 -10.7797
▁男孩 -10.7798
▁放心 -10.7799
▁绘画 -10.78
筒 -10.7801
▁采 -10.7805
▁夺 -10.7807
▁中药 -10.7807
柄 -10.7813
育 -10.7821
▁这点 -10.7825
▁猜 -10.7827
▁中午 -10.7828
▁山西 -10.783
▁定期 -10.7835
▁砸 -10.7837
▁一所 -10.7845
▁食用 -10.7849
▁药品 -10.785
浪 -10.7851
岗 -10.7855
▁依赖 -10.7856
▁审美 -10.7861
▁触 -10.7864
为准 -10.7869
▁护肤 -10.7869
▁加速 -10.7872
▁呈现 -10.7876
▁洪 -10.7877
▁人会 -10.7877
▁花钱 -10.7879
▁偶像 -10.7888
un -10.7891
魔 -10.7891
▁应对 -10.7892
工业 -10.7895
▁十六 -10.7897
▁三星 -10.7897
▁遂 -10.7898
▁鲁 -10.7899
▁血液 -10.79
寿 -10.7901
丑 -10.7903
▁时光 -10.7904
▁带有 -10.7904
羽 -10.7905
结合 -10.7906
▁驱动 -10.7906
莲 -10.7906
▁洞 -10.791
▁十三 -10.7915
▁融合 -10.7915
▁无关 -10.7916
▁精品 -10.7916
▁第二次 -10.7916
▁四大 -10.7917
▁填 -10.7918
▁家乡 -10.7919
▁危害 -10.792
▁1984 -10.7922
猪 -10.7923
▁杂 -10.7923
条件 -10.7924
▁炸 -10.7925
▁自行 -10.7925
▁家具 -10.7926
ro -10.7931
▁线路 -10.7931
▁国民 -10.7933
▁老年 -10.7933
各 -10.7936
▁上帝 -10.7938
▁民国 -10.794
霜 -10.7944
▁羊 -10.7946
¥ -10.7946
迁 -10.7948
▁野 -10.7949
▁抛 -10.7952
▁案件 -10.7953
▁吊 -10.7956
时代 -10.7957
▁特定 -10.7958
烦 -10.7965
▁禁止 -10.7967
茂 -10.797
▁34 -10.7981
ン -10.7983
▁义务 -10.7983
▁语文 -10.7983
▁虎 -10.7986
▁恐 -10.7987
▁妇女 -10.7987
▁宣布 -10.7996
▁两者 -10.7998
il -10.7998
▁住宅 -10.7999
man -10.8001
64 -10.8001
▁95 -10.8002
▁进攻 -10.8004
回事 -10.8005
▁房屋 -10.8008
▁48 -10.8009
▁争 -10.8009
价比 -10.8017
浦 -10.8017
▁互 -10.8019
▁识别 -10.8019
▁众 -10.8023
伴 -10.8024
▁均匀 -10.8025
▁始 -10.8025
▁面向 -10.803
斑 -10.8031
▁从业 -10.8034
双 -10.8035
▁於 -10.8036
痘 -10.8037
▁岛 -10.8037
分子 -10.8038
▁青岛 -10.8039
▁加盟 -10.8043
▁暑假 -10.8049
▁冲突 -10.8053
▁党委 -10.8053
▁高层 -10.8055
圆 -10.8055
涂 -10.8057
投 -10.8071
▁财产 -10.8071
国家 -10.8077
▁转变 -10.8078
▁巧 -10.8078
▁合格 -10.808
▁赴 -10.8081
▁全身 -10.8087
▁乔 -10.8089
释 -10.8094
▁表明 -10.8098
▁税收 -10.8108
▁普及 -10.8109
形码 -10.8111
语言 -10.8112
▁普通人 -10.8115
▁传奇 -10.8123
▁委托 -10.8125
▁监测 -10.8126
▁单独 -10.8127
框 -10.813
免 -10.8134
▁汉语 -10.8134
▁随机 -10.8136
▁江西 -10.8139
▁郡 -10.8142
▁将军 -10.8142
▁签 -10.8143
▁消除 -10.8143
囊 -10.8148
▁周末 -10.8153
▁短信 -10.8153
▁页面 -10.8154
am -10.8158
▁匿名 -10.8159
劳 -10.8162
健康 -10.8164
眼里 -10.8169
▁标签 -10.8171
▁聚 -10.8173
凡 -10.8175
民族 -10.8178
▁猎 -10.818
▁无人 -10.8182
21 -10.8183
研究 -10.8184
▁残 -10.8184
引 -10.8188
▁认定 -10.8195
▁古典 -10.82
▁彭 -10.8201
▁脸上 -10.8215
老师 -10.8216
▁转化 -10.8217
▁360 -10.8221
软 -10.8223
▁村委 -10.8226
暴 -10.8227
▁即便 -10.8227
500 -10.8236
27 -10.8237
▁0.5 -10.8245
拜 -10.8249
▁餐饮 -10.825
▁杀人 -10.8253
仲 -10.8258
▁骨 -10.8259
工作者 -10.8263
哲 -10.8267
今 -10.8268
▁监管 -10.827
▁承受 -10.8271
▁安慰 -10.8273
▁回头 -10.8273
▁边缘 -10.8274
▁股 -10.8277
▁1989 -10.8277
▁精英 -10.8279
▁潜 -10.8279
▁孔 -10.828
▁烂 -10.8292
▁穆斯林 -10.8292
32 -10.8296
▁江苏省 -10.8303
▁甘 -10.8304
▁雄 -10.8304
▁上海市 -10.8305
▁铺 -10.8311
▁第十一 -10.8311
诚 -10.8311
ang -10.8312
▁尤 -10.8313
▁39 -10.8314
▁厕所 -10.8317
▁学员 -10.8319
▁be -10.832
透 -10.832
▁能源 -10.8328
▁2.3 -10.8342
▁一首 -10.8345
▁避 -10.8347
ic -10.835
▁发射 -10.835
恶 -10.8354
▁1985 -10.8356
▁涂 -10.8358
▁圈子 -10.8358
▁陕西 -10.8365
▁语音 -10.8369
均 -10.8374
▁花园 -10.8374
戒 -10.8377
▁获奖 -10.838
▁天赋 -10.838
▁楼上 -10.8384
▁钟 -10.8386
▁3.2 -10.8386
辞 -10.8388
▁扣 -10.8389
▁党员 -10.8389
▁布局 -10.839
▁遗憾 -10.8391
政策 -10.8392
▁赋 -10.8396
奶 -10.8397
▁奋斗 -10.8397
▁大三 -10.8397
▁秀 -10.8401
考试 -10.8403
▁99 -10.8407
▁缓 -10.8409
▁引发 -10.8415
▁营业 -10.8415
所谓 -10.8417
灾 -10.8422
▁考核 -10.8425
▁命令 -10.8426
▁透 -10.8427
不同 -10.8433
代表 -10.8435
▁末 -10.8437
楚 -10.844
▁俩 -10.8442
秘 -10.8447
委员 -10.845
▁消化 -10.8456
▁消耗 -10.8459
▁床 -10.8459
▁1.4 -10.8466
▁科学技术 -10.8467
▁复制 -10.847
▁雅 -10.847
理事 -10.8472
▁渐 -10.8472
▁景观 -10.8478
⋯ -10.848
警 -10.8482
敬 -10.8484
▁灵活 -10.8486
▁新型 -10.8488
▁宠物 -10.8493
▁打印 -10.8494
不定 -10.8494
▁创办 -10.8494
▁煎 -10.8496
▁通用 -10.8497
壮 -10.85
▁泰 -10.8501
▁内涵 -10.8508
邻 -10.8514
贴 -10.852
▁租 -10.8522
尚 -10.8526
▁床上 -10.8532
廊 -10.8535
▁旅 -10.8539
▁贯彻 -10.8542
▁图形 -10.8547
▁禁 -10.8548
▁总面 -10.8548
告 -10.8549
▁地质 -10.8549
▽ -10.855
▁犯 -10.8553
▁异性 -10.8555
衰 -10.8562
▁演技 -10.8565
▁征 -10.857
谋 -10.857
霞 -10.8573
▁平常 -10.8577
▁出于 -10.8578
要说 -10.858
▁幻想 -10.8581
▁客服 -10.8584
▁作战 -10.8587
音乐 -10.8587
▁前期 -10.8595
▁斗 -10.8596
▁静 -10.8596
▁最多 -10.8597
莉 -10.86
茨 -10.8601
▁桥 -10.8608
▁事故 -10.8609
▁电压 -10.861
▁1987 -10.8613
▁出身 -10.8614
▁工人 -10.8616
文学 -10.8619
▁十几 -10.8621
▁确 -10.8621
▁法学 -10.8623
▁二级 -10.8624
▁同年 -10.8625
▁正面 -10.863
皇 -10.8631
资产 -10.8634
ine -10.8634
ω -10.8635
▁勾 -10.8637
▁较为 -10.8638
▁肌肤 -10.8639
.9 -10.864
▁mm -10.8642
▁外形 -10.8645
▁仪器 -10.8647
▁务工 -10.8649
▁诗人 -10.866
反 -10.8661
▁几十 -10.8662
▁种子 -10.8676
▁1986 -10.8677
▁潘 -10.8677
▁怀孕 -10.8678
▁胜 -10.868
▁扶 -10.8686
▁专题 -10.8688
▁故意 -10.8696
▁危机 -10.8699
▁尾 -10.87
▁牵 -10.87
莫 -10.8712
▁粘 -10.8719
▁180 -10.872
账 -10.8728
.12 -10.8732
▁微笑 -10.8741
▁售后 -10.8743
肥 -10.8744
急 -10.8747
▁枪 -10.8748
目录 -10.875
▁强制 -10.875
▁平等 -10.8752
▁频 -10.8754
▁审查 -10.8755
▁君 -10.8756
几 -10.8761
株 -10.8764
▁召开 -10.877
▁弟 -10.8771
请 -10.8771
ur -10.8771
灰 -10.8774
▁到处 -10.8775
▁威胁 -10.8781
▁生成 -10.8783
▁大会 -10.8783
禁 -10.8784
▁考察 -10.8789
▁排除 -10.8789
▁激光 -10.8792
▁大致 -10.8792
▁睡眠 -10.8794
▁敬 -10.8794
分析 -10.8797
▁3.5 -10.8799
▁暖 -10.8801
▁因而 -10.8802
毁 -10.8804
▁受伤 -10.8807
▁索 -10.8808
ter -10.8811
▁选手 -10.8814
▁一代 -10.8816
燕 -10.8824
膀 -10.8826
▁座 -10.8832
▁飘 -10.8834
▁充电 -10.8838
▁纵 -10.8839
77 -10.8842
▁舞 -10.8858
医学 -10.886
▁冲击 -10.8864
▁加快 -10.8864
▁立体 -10.8865
▁文科 -10.8865
桶 -10.8866
▁打工 -10.8867
▁失望 -10.8868
.7 -10.8878
▁肿瘤 -10.8884
绿 -10.8886
行为 -10.8888
▁二次 -10.8889
▁承诺 -10.889
▁南北 -10.889
▁一层 -10.8897
▁发育 -10.8898
▁口味 -10.8901
▁瓦 -10.8903
49 -10.8905
▁新加坡 -10.8911
▁42 -10.8911
▁一眼 -10.8915
ate -10.8928
争 -10.893
▁授 -10.893
▁麻 -10.8931
▁隶属 -10.8933
▁孤 -10.894
▁纹 -10.8942
汇 -10.8944
▁好玩 -10.8947
▁种种 -10.895
▁萌 -10.8953
▁哒 -10.8955
▁激动 -10.8955
移 -10.8959
92 -10.8959
▁苍 -10.8962
▁狼 -10.8964
36 -10.8968
▁随后 -10.8972
萌 -10.8975
▁抑制 -10.8975
▁110 -10.8977
▁医药 -10.8989
▁学期 -10.8994
际 -10.8994
▁踢 -10.8995
▁青少年 -10.9
▁全是 -10.9003
.10 -10.9003
▁邓 -10.9005
▁成人 -10.9007
▁立即 -10.9009
▁二是 -10.9014
节目 -10.9018
▁电动 -10.9021
▁唯 -10.9023
▁主观 -10.9028
– -10.9029
▁金钱 -10.9033
▁html -10.9035
▁部长 -10.9036
▁两天 -10.9038
▁不良 -10.9041
▁障碍 -10.9047
▁吐槽 -10.9054
婆 -10.9057
龄 -10.9058
▁咋 -10.9059
▁已有 -10.906
整 -10.9061
▁万人 -10.9061
▁庞 -10.9063
脱 -10.9065
▁85 -10.9071
▁词汇 -10.9073
▁探 -10.9074
๑ -10.9075
2.0 -10.9078
▁墓 -10.9079
▁极其 -10.908
▁节能 -10.9081
妻 -10.9088
▁隐 -10.9088
孙 -10.9091
▁更大 -10.9092
▁测 -10.9094
感觉 -10.9099
翻 -10.9102
▁强度 -10.9102
▁Android -10.9103
▁简体中文 -10.9106
祸 -10.9107
▁唱歌 -10.9111
欧 -10.9113
穿 -10.9122
▁融资 -10.9123
╯ -10.9124
▁录音 -10.9125
▁3.1 -10.9128
▁更为 -10.9131
▁吉他 -10.9131
泡 -10.9134
▁土壤 -10.9134
▁深深 -10.9139
依 -10.9139
▁逗 -10.914
▁踩 -10.9141
反应 -10.9141
at -10.9143
▁回国 -10.9145
▁霍 -10.9146
▁寂寞 -10.9148
菇 -10.915
▁作曲 -10.9152
as -10.9154
涛 -10.9154
▁编剧 -10.9164
▁当事 -10.9168
▁牺牲 -10.9174
▁人格 -10.9176
▁丈 -10.9176
im -10.9177
▁PC -10.9185
ml -10.9186
▁盈利 -10.9186
▁功率 -10.9191
▁公交 -10.9195
▁外出 -10.92
▁Intel -10.9201
旅 -10.9205
歉 -10.9206
粮 -10.921
滑 -10.9212
▁新人 -10.9214
尊 -10.9216
▁职工 -10.9216
▁正规 -10.9218
▁验证 -10.922
▁进程 -10.9222
▁2.4 -10.9226
▁相反 -10.9227
保险 -10.923
▁东京 -10.9231
浅 -10.9233
▁斜 -10.9239
▁钻 -10.9243
▁课堂 -10.9247
▁qq -10.9249
▁img -10.9258
公里 -10.9263
砖 -10.9267
▁忽略 -10.9269
▁Pro -10.9275
电影 -10.9279
▁吕 -10.9279
▁疯 -10.9281
▁高手 -10.9284
▁极端 -10.9286
▁51 -10.9286
▁凉 -10.9288
▁伤心 -10.9288
▁墨 -10.929
▁一时 -10.929
▁撞 -10.9291
os -10.9303
▁疑 -10.9307
▁质疑 -10.9308
狂 -10.9309
▁形容 -10.9311
据 -10.9312
▁成年 -10.9315
独 -10.9317
▁等于 -10.9317
侠 -10.9317
译 -10.9323
符 -10.9325
肌 -10.9327
▁晒 -10.9327
助 -10.933
▁指点 -10.9333
冬 -10.9334
▁滑 -10.9336
▁信心 -10.9338
毕 -10.9342
ir -10.9344
▁一是 -10.9351
▁不但 -10.9351
师范大学 -10.9354
▁市委 -10.9355
▁竞赛 -10.9356
▁损 -10.9357
▁一番 -10.9358
ie -10.9359
位置 -10.9359
沉 -10.9362
途 -10.9367
▁躲 -10.9369
▁肌 -10.937
per -10.9372
▁浪漫 -10.9376
.1 -10.9376
孝 -10.9377
宅 -10.9379
▁关闭 -10.9392
▁诞生 -10.9392
价值 -10.9392
▁长江 -10.9393
▁迷 -10.9395
郁 -10.9396
▁效应 -10.9397
甜 -10.9399
▁人事 -10.94
▁大一 -10.9404
不见 -10.9406
▁接收 -10.9408
膏 -10.941
▁御 -10.9415
▁五个 -10.9416
▁做饭 -10.9417
▁传输 -10.9422
▁广播 -10.9437
q -10.9443
▁销 -10.9449
卿 -10.9451
▁拓展 -10.9453
▁不高 -10.9455
─ -10.9456
--------------- -10.946
▁职责 -10.9465
.2 -10.9467
▁回报 -10.9471
⊙ -10.9472
规 -10.9477
逝 -10.9482
腹 -10.9483
▁道歉 -10.9485
伏 -10.9487
▁微商 -10.949
▁干燥 -10.949
诊 -10.9496
浆 -10.9499
▁浙江省 -10.9499
▁体制 -10.9501
鹿 -10.9501
▁211 -10.9506
▁释 -10.9508
▁抱怨 -10.951
▁真相 -10.9517
▁适宜 -10.9519
▁显著 -10.9521
▁旁 -10.9522
▁甩 -10.9526
▁违反 -10.9528
▁优雅 -10.9533
▁放下 -10.9536
窄 -10.9536
▁研制 -10.9538
▁网易 -10.9539
蓝 -10.9543
▁安静 -10.9544
▁小小 -10.955
▁保密 -10.9551
忧 -10.9558
▁连锁 -10.956
son -10.957
▁碎 -10.9574
▁绿化 -10.9585
▁数码 -10.9588
▁挤 -10.9588
▁低于 -10.9592
法院 -10.9592
▁985 -10.9595
▁临时 -10.9596
饼 -10.9602
▁采访 -10.9604
▁体内 -10.9605
聊 -10.9607
ce -10.9608
▁APP -10.9611
20 -10.9616
▁65 -10.9621
爹 -10.9622
文明 -10.9622
思想 -10.9624
▁师范 -10.9624
▁贫困 -10.9631
▁子女 -10.9633
逢 -10.9633
▁公民 -10.9636
▁撤 -10.9639
▁不吃 -10.964
▁继 -10.964
▁各大 -10.9642
盆 -10.9643
▁公安 -10.9645
▁萧 -10.9649
▁名人 -10.9659
▁催 -10.9671
▁演讲 -10.9674
企 -10.9674
▁术 -10.9676
候 -10.9677
▁幸运 -10.9679
▁证 -10.9684
▁进化 -10.9685
赫 -10.9685
▁审批 -10.969
▁中文名 -10.969
▁习题 -10.9691
▁夫妻 -10.9692
▁手里 -10.9692
束 -10.9694
腰 -10.9705
不该 -10.9706
▁汪 -10.9715
▁撩 -10.9717
▁用途 -10.9717
网络 -10.9718
▁凭借 -10.9719
▁气候 -10.972
▁题材 -10.9726
▁58 -10.9729
▁炮 -10.973
72 -10.9732
50 -10.9734
浩 -10.9735
▁微软 -10.9737
▁这边 -10.974
▁私人 -10.974
▁护 -10.9742
mo -10.9742
▁抹 -10.9743
▁蛋 -10.9745
洗 -10.9748
勤 -10.9754
▁移 -10.9756
▁新手 -10.9761
▁思 -10.9761
芬 -10.9765
▁回归 -10.9766
▁洗净 -10.9766
te -10.9771
▁氛 -10.9771
▁用心 -10.9772
▁若干 -10.9773
▁容量 -10.9774
▁薄 -10.9774
蛇 -10.9776
▁释放 -10.9777
▁承 -10.9787
▁跪 -10.9789
▁执法 -10.979
▁眼泪 -10.9791
▁每日 -10.9796
拔 -10.9797
▁董 -10.9797
▁停车 -10.98
age -10.9809
▁运作 -10.9813
▁一人 -10.9813
▁托 -10.9814
▁伪 -10.9814
佩 -10.9815
▁水利 -10.9816
▁爱国 -10.9819
▁鄙 -10.9823
▁争取 -10.9824
▁传递 -10.9825
▁距 -10.9831
▁文献 -10.9832
▁厦门 -10.9834
只 -10.9839
▁第十二 -10.9844
▁多媒体 -10.9846
▁婚礼 -10.9847
▁幽 -10.9848
▁疼痛 -10.9849
▁帖 -10.9855
▁一方 -10.9857
▁环节 -10.9859
▁河南省 -10.986
墨 -10.9861
▁顾 -10.9861
ma -10.9862
▁教育部 -10.9863
▁cn -10.9868
▁职能 -10.987
▁砍 -10.9873
习惯 -10.9875
漠 -10.9875
泥 -10.9877
▁电流 -10.988
▁纤维 -10.9884
▁阿里 -10.9885
▁可怜 -10.9889
▁毫 -10.9891
▁颇 -10.9891
▁不止 -10.9892
▁解析 -10.9892
▁十四 -10.9895
▁闪 -10.9898
▁k -10.9899
▁泥 -10.99
▁地域 -10.9901
▁好评 -10.9906
▁一门 -10.9911
▁骑士 -10.9912
▁不在 -10.9913
帽 -10.9914
▁蒋 -10.9921
▁划分 -10.9929
▁林业 -10.9929
休 -10.9929
提高 -10.9933
瘤 -10.9934
▁犹 -10.9936
功能 -10.9939
▁十大 -10.9939
蛋白 -10.9941
▁场所 -10.9942
▁自动化 -10.9944
45 -10.9945
▁过度 -10.9947
▁MP -10.995
▁举报 -10.995
▁戏剧 -10.9951
杂 -10.9951
▁招生 -10.9952
▁淡 -10.9957
▁容 -10.9957
事业 -10.996
▁1982 -10.9961
▁货物 -10.9963
▁沉 -10.9963
▁返 -10.9966
▁传感 -10.9967
苑 -10.9971
偏 -10.9974
▁逆 -10.9978
▁穿越 -10.9983
惑 -10.9984
▁刊 -10.999
▁虾 -10.9993
▁冒 -10.9994
▁三大 -11.0001
▁说服 -11.0004
轨 -11.0005
▁落后 -11.0009
▁中国共产党 -11.001
馨 -11.0013
▁放大 -11.0015
▁荣 -11.0015
▁凤凰 -11.0015
▁on -11.0016
▁辆 -11.0022
匀 -11.0024
re -11.0025
▁凯 -11.0026
▁号码 -11.0026
喝 -11.0029
▁期望 -11.0032
柏 -11.0033
建设 -11.0033
▁报警 -11.0034
▁录取 -11.0034
内容 -11.0038
▁好久 -11.0041
胖 -11.0041
▁加班 -11.0042
15 -11.0046
▁缺陷 -11.0051
um -11.0052
生涯 -11.0056
▁库 -11.0059
▁导 -11.0061
▁过敏 -11.0062
▁遗传 -11.0064
▁罗马 -11.0065
▁一百 -11.0065
阶 -11.0066
革命 -11.0067
平台 -11.0067
▁优先 -11.0069
▁毁 -11.0073
▁民众 -11.0077
▁争议 -11.008
交通 -11.0083
一代 -11.0083
晨 -11.0085
琳 -11.0088
坊 -11.0094
▁从未 -11.0094
▁建造 -11.0095
▁园林 -11.0096
▁绝望 -11.0097
▁欠 -11.01
▁总有 -11.0101
▁楚 -11.0103
▁女主 -11.0103
▁1983 -11.0105
欢迎 -11.0106
ac -11.0106
▁天使 -11.0109
▁锁 -11.0109
▁澳大利亚 -11.0119
▁一声 -11.0122
▁戒 -11.0126
震 -11.0127
▁with -11.013
叉 -11.0135
▁答应 -11.0145
.4 -11.0149
I -11.015
▁局部 -11.0152
▁代替 -11.0154
▁几句 -11.0158
范围 -11.0159
▁恩 -11.0161
▁处女 -11.0162
▁全程 -11.0167
MHz -11.0168
▁以此 -11.0169
迫 -11.017
▁买房 -11.0171
▁内地 -11.0174
动物 -11.0178
▁三十 -11.0178
▁期刊 -11.0179
▁II -11.0182
▁奉 -11.0185
ne -11.0186
▁挖 -11.0193
▁自行车 -11.0193
▁技 -11.0194
▁羡 -11.0195
▁乡镇 -11.0198
▁饰演 -11.0198
▁游客 -11.0212
▁98 -11.0227
▁祭 -11.0231
铃 -11.0231
▁孟 -11.0233
▁科普 -11.0233
▁44 -11.0234
mg -11.0236
▁福利 -11.0236
▁由此 -11.024
▁酷 -11.0244
耗 -11.0245
▁损害 -11.0246
▁正义 -11.0248
▁非洲 -11.0249
▁机器人 -11.0251
▁250 -11.0252
▁r -11.0253
看来 -11.0254
阶段 -11.0255
旁 -11.0255
玩笑 -11.0257
▁报名 -11.0258
▁原文 -11.0258
▁每周 -11.0259
▁不怕 -11.026
▁形状 -11.026
▁咯 -11.0263
▁天地 -11.0265
▁帕 -11.0266
▁主机 -11.0266
决 -11.0271
▁打扮 -11.0271
▁水泥 -11.0272
▁笔记 -11.0281
翁 -11.0282
▁百科 -11.0283
▁指数 -11.0288
▁激烈 -11.0291
方位 -11.0292
▁No -11.0296
▁各级 -11.0304
▁电气 -11.0307
▁大人 -11.0308
++ -11.031
▁学到 -11.0311
▁出轨 -11.0313
▁亲自 -11.0314
▁慢性 -11.032
▁辐射 -11.033
签 -11.0331
96 -11.0334
▁京 -11.0335
ra -11.0341
▁江南 -11.0342
ol -11.0346
▁幼 -11.0348
范 -11.035
▁养生 -11.035
▁隐藏 -11.0353
翔 -11.0353
控制 -11.0354
▁西藏 -11.0355
▁滴 -11.0356
▁战士 -11.0356
▁一口 -11.036
终 -11.0366
渣 -11.0367
戈 -11.0369
▁偶然 -11.037
百姓 -11.037
▁简历 -11.0372
▁价 -11.0372
▁职务 -11.0373
品质 -11.0374
固 -11.0381
▁抓住 -11.0381
▁周期 -11.0384
.3 -11.0386
▁中介 -11.0387
▁全省 -11.0389
▁闺蜜 -11.0389
▁拿出 -11.0391
习 -11.0397
开放 -11.0399
宏 -11.0404
43 -11.0404
▁渴望 -11.0405
▁长沙 -11.0407
▁扫 -11.0408
▁萨 -11.0411
莎 -11.0413
▁男票 -11.0419
▁顺序 -11.042
▁北大 -11.042
▁赌 -11.0422
▁不论 -11.043
▁4.2 -11.0431
演员 -11.0434
▁更换 -11.0436
▁VIP -11.0436
逊 -11.044
唇 -11.0444
▁坚定 -11.0446
79 -11.0453
▁阐 -11.0455
▁文物 -11.0456
U -11.0457
兆 -11.0458
▁偷偷 -11.0463
▁活性 -11.0466
鹏 -11.047
▁人间 -11.0471
▁蒸 -11.0472
珍 -11.0479
▁天生 -11.048
▁加热 -11.0481
效力 -11.0485
渡 -11.0486
▁本会 -11.049
▁灭 -11.0491
桃 -11.0492
昂 -11.0495
▁海军 -11.0497
▁兔子 -11.0501
▁手法 -11.0501
▁幽默 -11.0501
缓 -11.0503
▁试图 -11.0503
▁湿 -11.0505
▁前景 -11.0507
▁3.3 -11.0509
▁治理 -11.0514
下列 -11.0514
▁新浪 -11.0515
▁直径 -11.0518
▁窗口 -11.0519
成本 -11.0519
▁中等 -11.0519
▁稳 -11.0521
▁眼神 -11.0523
▁陶 -11.0525
▁夸 -11.0528
▁活跃 -11.053
▁大气 -11.0531
▁吨 -11.0534
▁帝国 -11.0534
▁早已 -11.0535
▁联想 -11.0535
▁收购 -11.0536
▁1500 -11.0537
▁批判 -11.0539
▁按摩 -11.054
▁畜牧 -11.0543
▁好几 -11.0544
▁防御 -11.0544
增 -11.0546
消 -11.0546
▁奥运会 -11.055
▁存 -11.055
▁绝不 -11.055
▁人气 -11.0554
榜 -11.0554
▁远远 -11.0556
ent -11.0566
旺 -11.0569
凯 -11.057
筋 -11.0572
▁氧化 -11.0573
▁电梯 -11.0574
▁真诚 -11.0577
34 -11.0582
▁放松 -11.0582
▁房价 -11.0587
▁一道 -11.0594
▁清洗 -11.0595
监督管理 -11.0596
▁塑造 -11.0596
不出 -11.0596
▁凶 -11.0605
▁请求 -11.0606
经验 -11.0608
pa -11.061
▁崔 -11.0614
▁传媒 -11.0619
疑 -11.062
艺术家 -11.0622
▁共享 -11.0628
▁大道 -11.0633
▁纽约 -11.0633
▁兴奋 -11.0637
▁账户 -11.0639
▁眼前 -11.0641
▁整形 -11.0642
靖 -11.0644
▁替代 -11.0647
▁辞职 -11.0648
▁变形 -11.0653
▁恶意 -11.0654
戴 -11.0655
▁象征 -11.0655
▁眼中 -11.0656
▁选举 -11.0656
▁负面 -11.0659
▁返回 -11.0662
▁柔 -11.0664
▁禅 -11.0665
and -11.067
▁46 -11.0671
▁猛 -11.0679
辈子 -11.0679
夹 -11.0679
▁含义 -11.068
▁滤 -11.0684
▁富有 -11.0684
统 -11.0685
延 -11.0692
▁re -11.0693
▁荷兰 -11.0693
▁四川省 -11.0695
▁遭遇 -11.0696
▁醉 -11.0698
故事 -11.0699
▁大于 -11.0701
▁诗歌 -11.0702
▁将会 -11.0703
▁观看 -11.0703
▁促 -11.0704
▁1980 -11.0709
▁舆论 -11.0715
丛 -11.0716
▁社会科学 -11.0717
▁轨道 -11.0717
▁团结 -11.0728
▁承包 -11.0728
▁43 -11.0728
66 -11.073
羞 -11.0732
悦 -11.0733
操作 -11.0733
▁陷入 -11.0737
44 -11.0737
▁键 -11.0747
▁警 -11.0747
▁全体 -11.0747
盟 -11.0747
乳 -11.0748
▁悬 -11.0752
▁区分 -11.0752
▁激励 -11.0752
▁业绩 -11.0752
▁栽培 -11.0754
▁扫描 -11.0756
遗址 -11.0757
▁激情 -11.0768
▁词语 -11.0768
▁介 -11.0769
▁充 -11.077
▁乳 -11.0771
▁耳朵 -11.0771
▁寻求 -11.0773
捕 -11.0777
▁赞同 -11.0778
引用 -11.0778
▁辣 -11.078
▁强迫 -11.0781
▁俗 -11.0785
▁乐趣 -11.0787
▁严谨 -11.0788
▁160 -11.0789
▁生意 -11.0791
▁天空 -11.0794
▁斗争 -11.0803
▁渔 -11.0804
眉 -11.0806
▁蓝色 -11.0806
▁大专 -11.0809
▁压缩 -11.0811
垫 -11.0812
瘦 -11.0814
焦 -11.0816
0.00 -11.0817
▁答题 -11.0818
▁血管 -11.0819
▁砖 -11.0821
裕 -11.0823
▁模仿 -11.0823
▁智 -11.0825
▁外部 -11.083
▁LED -11.0833
▁账号 -11.0833
朴 -11.0835
▁业余 -11.0837
▁今后 -11.0838
ge -11.0838
▁相机 -11.0839
33 -11.0843
▁拔 -11.0859
▁it -11.0861
▁吾 -11.0862
▁壮 -11.0863
▁句子 -11.0865
▁新生 -11.0867
▁56 -11.0869
▁崩溃 -11.0874
电视 -11.0878
▁手指 -11.0882
▁St -11.0885
▁未知 -11.0886
▁迁 -11.0895
碑 -11.0896
▁安卓 -11.0896
▁不满 -11.0898
▁碧 -11.09
▁搞笑 -11.0903
第 -11.0904
▁宜 -11.0906
▁开关 -11.0908
▁呗 -11.091
▁一一 -11.0911
91 -11.0914
▁航 -11.0915
不知 -11.0917
▁入选 -11.0917
▁唱片 -11.0917
晋 -11.0921
▁洛 -11.0922
▁频道 -11.0927
▁葱 -11.0928
困 -11.0929
划 -11.0933
继 -11.0935
▁错过 -11.0939
巾 -11.0953
▁后果 -11.0953
▁口腔 -11.0954
▁52 -11.0954
▁理工 -11.0959
情况 -11.096
▁葡萄 -11.096
▁41 -11.0962
▁榜 -11.0965
38 -11.0966
▁童年 -11.097
▁改编 -11.0973
▁扎 -11.0973
▁000 -11.0974
妃 -11.0977
▁玩具 -11.098
▁开拓 -11.0982
阅 -11.0983
▁进来 -11.0983
▁愉 -11.0989
▁茶叶 -11.099
▁统 -11.0994
▁帅 -11.0995
栏 -11.0995
▁洋 -11.0996
始 -11.0996
▁路面 -11.1
狼 -11.1003
▁47 -11.1003
▁突 -11.1006
▁讲解 -11.1006
气温 -11.1007
罩 -11.1008
会议 -11.101
如此 -11.1011
辖 -11.1013
穴 -11.1015
▁人大 -11.1017
▁语法 -11.1018
▁初步 -11.1018
席 -11.1018
一枚 -11.102
豪 -11.1025
19 -11.1026
▁城镇 -11.1027
▁鼓 -11.1031
掌 -11.1032
▁防守 -11.1034
23 -11.1037
▁方言 -11.1037
▁迈 -11.1039
▁专项 -11.1043
银 -11.1047
酷 -11.1049
▁两三 -11.1052
▁缺少 -11.1055
▁模糊 -11.1057
执行 -11.1072
lo -11.1075
品牌 -11.1081
▁税务 -11.1081
▁实战 -11.1085
广告 -11.1087
目标 -11.1088
▁玄 -11.1089
▁弥 -11.1089
▁厨房 -11.1091
舟 -11.1092
一刻 -11.1094
◆ -11.1096
▁十七 -11.1097
▁IP -11.1099
▁旋转 -11.1099
敢 -11.1103
藤 -11.1105
▁Me -11.1106
▁今日 -11.1106
▁拉丁 -11.1108
姿 -11.1108
▁巴黎 -11.1109
规定 -11.1111
▁巴西 -11.1111
辩 -11.1113
▁电磁 -11.1119
▁长相 -11.1121
▁中小 -11.1121
▁露 -11.1124
手机 -11.1125
▁无需 -11.1127
▁乃至 -11.1127
咒 -11.1128
▁酱 -11.1129
▁that -11.1138
做事 -11.1139
潭 -11.114
▁果然 -11.1143
▁负担 -11.1147
▁恨 -11.1147
▁主办 -11.1148
须 -11.1151
▁尖 -11.1153
▁他会 -11.1156
▁肥 -11.1158
▁小型 -11.1159
晶 -11.1172
▁次数 -11.1175
▁饿 -11.1178
▁一万 -11.1179
▁燃烧 -11.118
▁一脸 -11.118
▁可用 -11.1182
▁情景 -11.1185
▁自私 -11.1188
推 -11.1189
▁姿 -11.119
公园 -11.1192
▁申 -11.1195
犬 -11.1205
青年 -11.1206
▁挡 -11.1215
▁卵 -11.1217
仗 -11.1217
裔 -11.1222
▁麦克 -11.1224
岳 -11.1226
▁焦虑 -11.1227
▁上级 -11.1227
▁起码 -11.1229
酱 -11.1232
▁不愿 -11.1236
▁最小 -11.1237
▁婚 -11.1239
椒 -11.124
▁辞 -11.1241
▁牛奶 -11.1243
▁胶 -11.1246
▁阵 -11.1246
100 -11.1247
瓜 -11.1247
18 -11.1249
▁懒 -11.1256
愁 -11.1261
丸 -11.1261
▁no -11.1267
25 -11.1269
▁赔偿 -11.1274
▁悲剧 -11.1276
轩 -11.1283
▁多项 -11.1286
壶 -11.1286
▁脱离 -11.1289
淫 -11.1291
▁同行 -11.1293
▁三次 -11.13
▁理科 -11.1301
用品 -11.1302
▁精心 -11.1303
▁应急 -11.1303
工具 -11.1304
滴 -11.1305
惊 -11.1306
隐 -11.1307
清楚 -11.1307
毅 -11.1309
窝 -11.1311
▁牛肉 -11.1313
燃 -11.1314
▁勿 -11.1317
▁1.6 -11.1317
▁泰国 -11.1326
▁精准 -11.1326
▁不幸 -11.133
▁勇敢 -11.1333
▁作文 -11.1334
不得 -11.1337
▁快递 -11.1342
▁颜值 -11.1343
▁研究院 -11.1345
▁热水 -11.1351
▁辛 -11.1351
梯 -11.1351
▁焦 -11.1355
▁描写 -11.1356
纯 -11.1359
▁巡 -11.1359
▁股东 -11.136
▁埃 -11.1361
▁沈阳 -11.1361
▁97875 -11.1361
▁通道 -11.1361
▁暴露 -11.1363
▁痘 -11.1363
▁退出 -11.1368
▁都市 -11.1372
▁助理 -11.1374
▁替 -11.1377
▁常规 -11.1378
▁多样 -11.1383
▁局限 -11.1385
な -11.1386
▁第十三 -11.1388
▁加大 -11.1389
▁快捷 -11.139
▁保养 -11.1391
愈 -11.1392
▁城乡 -11.1394
▁细菌 -11.1397
▁威 -11.1397
▁投票 -11.1399
▁4.1 -11.1404
聘 -11.1409
▁房产 -11.1409
艺术 -11.1413
效率 -11.1414
▁导航 -11.1415
刺 -11.1416
痒 -11.1417
to -11.1419
▁老子 -11.1422
▁勇士 -11.1425
▁弯 -11.1426
▁捡 -11.1426
旬 -11.1427
▁提取 -11.1429
▁魔兽 -11.143
肝 -11.1431
烧 -11.1433
▁一带 -11.1436
▁牙齿 -11.1436
▁与其 -11.1436
▁达成 -11.1439
▁郑州 -11.1439
▁矣 -11.1439
▁咸 -11.144
▁声明 -11.1441
▁岂 -11.1444
▁华为 -11.1444
▁持 -11.1448
芽 -11.1451
▁上映 -11.1451
出去 -11.1452
▁党组 -11.1456
涨 -11.1456
▁5.2 -11.1457
▁疑问 -11.1458
▁举个 -11.1459
▁构造 -11.146
▁大规模 -11.1462
▁过年 -11.1465
把 -11.1466
▁卓越 -11.1468
▁勇气 -11.1469
▁白天 -11.1472
▁NBA -11.1472
▁补偿 -11.1477
便是 -11.1486
▁申报 -11.1488
▁扮演 -11.1489
▁曝光 -11.149
辑 -11.1496
▁莱 -11.1497
▁纳税 -11.1504
▁额 -11.1516
▁棉 -11.1518
循环 -11.1519
融 -11.152
▁现金 -11.1521
▁惨 -11.1523
▁人力资源 -11.1525
坪 -11.1528
▁中小学 -11.1531
▁情侣 -11.1531
桂 -11.1533
▁消防 -11.1535
抗 -11.1542
择 -11.1545
▁高于 -11.1546
86 -11.1548
妮 -11.1549
自然 -11.156
▁听听 -11.1562
闹 -11.1563
▁对抗 -11.1566
人物 -11.1577
酒店 -11.1583
▁学士 -11.1583
▁五十 -11.1588
.11 -11.1588
93 -11.1588
▁饮 -11.1591
▁哥们 -11.1593
▁创立 -11.1593
辰 -11.1593
糟 -11.1595
哀 -11.1599
人才 -11.16
▁拌 -11.1601
se -11.1603
颈 -11.1604
瘾 -11.1607
▁净 -11.1617
▁气体 -11.1617
▁下班 -11.1618
不说 -11.1622
▁商标 -11.1624
穗 -11.1627
▁防护 -11.1629
▁Google -11.1629
▁落实 -11.1629
滩 -11.1631
▁混乱 -11.1631
▁蛇 -11.1631
▁电信 -11.1632
▁首发 -11.1636
鲜 -11.1643
▁佳 -11.1644
▁剧本 -11.1647
▁艺人 -11.1653
▁单词 -11.1656
▁国王 -11.1658
阔 -11.166
▁一阵 -11.1662
▁诺 -11.1663
右 -11.1667
▁管辖 -11.1668
▁j -11.1671
ai -11.1674
▁倾 -11.1676
▁130 -11.1677
▁强行 -11.1677
▁截止 -11.168
▁起源 -11.168
▁纷纷 -11.1683
▁矮 -11.1691
▁标题 -11.1691
▁脏 -11.1694
▁短期 -11.1694
锦 -11.1695
▁优良 -11.1695
塑 -11.1695
▁养成 -11.1696
▁黄色 -11.1696
▁发病 -11.17
▁远程 -11.17
盾 -11.1701
▁友好 -11.1701
ment -11.1706
▁独自 -11.171
▁资讯 -11.171
▁外交 -11.171
▁万平方 -11.1712
▁其余 -11.1714
硬 -11.1714
减 -11.1718
▁几百 -11.1719
▁针 -11.1722
▁4000 -11.1723
ch -11.1724
贞 -11.1728
▁汉字 -11.1728
▁度过 -11.1729
▁so -11.1735
▁洗衣 -11.1737
▁出色 -11.1738
怡 -11.1739
▁自觉 -11.1739
措施 -11.1743
▁身份证 -11.1745
沃 -11.1746
▁公交车 -11.1749
亿 -11.1753
▁宝贝 -11.1758
▁别说 -11.1759
▁天才 -11.1759
染 -11.1768
创新 -11.1769
▁传承 -11.1771
为此 -11.1773
▁豆腐 -11.1773
▁在我 -11.1778
▁国民党 -11.1779
▁动手 -11.1781
57 -11.1783
▁服饰 -11.1784
▁起床 -11.1786
▁实时 -11.1787
▁侯 -11.1787
▁卫星 -11.1788
▁曲线 -11.1788
▁心脏 -11.1792
院校 -11.1793
▁闻 -11.1796
▁豪华 -11.1799
输 -11.18
200 -11.18
▁延伸 -11.1806
▁进球 -11.1807
▁初恋 -11.1814
▁陶瓷 -11.182
▁英格兰 -11.1824
颗 -11.1825
▁秘 -11.1826
赢 -11.1827
▁谷 -11.1831
▁ISO -11.1836
骂 -11.1841
▁版权 -11.1841
▁文档 -11.1842
规则 -11.1842
▁童 -11.1844
▁觉 -11.1844
▁彩色 -11.1846
▁唉 -11.1852
▁by -11.1853
▁pro -11.1853
▁国有 -11.1858
▁豆 -11.186
探 -11.186
精致 -11.1865
▁履行 -11.1866
▁不爱 -11.1868
▁骄傲 -11.1868
▁展览 -11.1869
▁一看 -11.1869
▁种族 -11.187
▁提交 -11.187
▁两位 -11.1872
帖 -11.1876
IC -11.1878
丘 -11.1879
▁后续 -11.1882
▁冒险 -11.1884
▁创始 -11.1886
▁博客 -11.1887
▁诸多 -11.1892
▁my -11.1893
自治区 -11.1894
▁详 -11.1896
▁中部 -11.1897
荡 -11.1901
▁眼光 -11.1904
▁刮 -11.1906
▁邮件 -11.1909
▁架构 -11.191
ip -11.1912
▁常年 -11.1912
.5 -11.1918
▁主力 -11.1921
炒 -11.1922
▁岁月 -11.1926
▁演奏 -11.1928
▁洗澡 -11.1929
▁得出 -11.1929
▁二年 -11.193
▁1978 -11.193
▁大连 -11.1932
▁桑 -11.1934
胆 -11.1935
▁from -11.1937
▁克服 -11.194
▁眼镜 -11.194
▁肚子 -11.1941
▁哈尔滨 -11.1941
ry -11.1946
65 -11.1948
▁赏 -11.195
▁4.3 -11.1951
▁惹 -11.1954
▁绕 -11.1956
▁轴 -11.196
▁插槽 -11.1961
em -11.1963
接口 -11.1966
铭 -11.1966
are -11.1967
▁热门 -11.1969
▁暂 -11.197
▁公寓 -11.1971
▁严肃 -11.1973
▁无力 -11.1974
▁打破 -11.1978
▁毛泽东 -11.198
ive -11.1992
隔 -11.1996
▁口碑 -11.1997
▁悲伤 -11.2
▁精华 -11.2001
▁刻意 -11.2001
▁深圳市 -11.2002
▁智力 -11.2003
▁婴儿 -11.2003
▁春秋 -11.2008
▁煤 -11.201
醉 -11.201
▁饮料 -11.2013
▁生育 -11.2013
▁好听 -11.2015
▁混凝土 -11.2016
颂 -11.2021
▁姓 -11.2022
▁发起 -11.2022
捷 -11.2022
▁流传 -11.2027
▁山水 -11.2028
陶 -11.203
▁宾馆 -11.203
▁资质 -11.203
职工 -11.2032
▁师资 -11.2033
26 -11.2034
est -11.2035
▁蒙古 -11.2036
▁空中 -11.2038
续 -11.2038
▁一身 -11.2039
▁神话 -11.2041
蝶 -11.2043
文件 -11.2043
▁不易 -11.2043
た -11.2043
效益 -11.2044
▁牙 -11.2048
▁伴 -11.2048
▁大叔 -11.205
严 -11.2051
▁票 -11.2051
▁湖北省 -11.2055
▁希 -11.2057
▁尊 -11.2061
▁未必 -11.2062
▁从不 -11.2063
▁财 -11.2063
▁校区 -11.2066
拼 -11.2068
▁营 -11.2071
▁政 -11.2072
交流 -11.2076
▁杀死 -11.2077
▁战术 -11.2079
▁幼稚 -11.2079
▁退休 -11.208
▁序 -11.208
▁实训 -11.2084
▁湖南省 -11.2085
▁前端 -11.2086
▁豪 -11.2089
▁掏 -11.2091
芝 -11.2092
▁股市 -11.2095
▁一辆 -11.2095
▁百年 -11.2099
▁雾 -11.2099
▁一会 -11.2103
▁手中 -11.2104
宴 -11.2105
▁生动 -11.2105
76 -11.2105
▁发动 -11.2107
▁主导 -11.2108
▁祝福 -11.2109
摩 -11.2111
渊 -11.2112
▁拨 -11.212
▁埋 -11.2122
▁汇 -11.2125
▁外语 -11.2127
▁凑 -11.2128
▁清新 -11.2132
▁叔 -11.2137
▁泛 -11.2138
▁势力 -11.214
雕 -11.2143
不可 -11.2144
椅 -11.2144
▁工商 -11.2145
▁意志 -11.2147
▁通电 -11.2149
▁约定 -11.2149
侯 -11.2149
▁流通 -11.2151
全书 -11.2151
▁圆形 -11.2151
▁脾气 -11.2152
▁别墅 -11.2154
映 -11.2157
▁测定 -11.2158
▁六年 -11.216
▁越南 -11.2161
54 -11.2162
▁高等教育 -11.2163
宿 -11.2164
▁典 -11.2165
▁瑞 -11.2168
▁反驳 -11.2175
姑娘 -11.2176
▁辩论 -11.218
▁服用 -11.2181
▁投影 -11.2182
▁季节 -11.2183
▁酸 -11.2186
碗 -11.2189
▁轻轻 -11.219
▁极限 -11.2191
▁玫瑰 -11.2191
▁看上 -11.2194
▁平安 -11.2194
▁斩 -11.2196
▁北京大学 -11.2197
▁dota -11.2197
溢 -11.22
▁高速公路 -11.2207
▁宏观 -11.2208
化学 -11.2209
▁DVD -11.221
过来 -11.2211
▁磁 -11.2211
▁图案 -11.2212
桑 -11.2217
▁取代 -11.2218
▁立法 -11.2222
唐 -11.2225
▁大街 -11.2225
▁南部 -11.2228
▁淘汰 -11.2233
▁贵州 -11.2233
诉 -11.2234
▁三角 -11.2238
▁市政府 -11.2239
▁很久 -11.224
▁崇 -11.2244
合同 -11.2246
▁县级 -11.225
首歌 -11.2255
▁吉林 -11.2257
▁经费 -11.226
道德 -11.226
▁首都 -11.2261
▁按钮 -11.2262
▁模特 -11.2262
▁战场 -11.2263
析 -11.2263
蜜 -11.2266
▁ng -11.2267
残 -11.2268
霸 -11.2268
io -11.2269
▁缘 -11.2273
奔 -11.2278
舌 -11.228
▁帝 -11.228
83 -11.2282
▁拯救 -11.2283
▁辽宁 -11.2284
▁蹲 -11.2285
▁感冒 -11.2291
▁有利 -11.2292
胸 -11.2294
▁友谊 -11.2295
▁大厦 -11.2296
成绩 -11.2296
▁阻止 -11.2298
▁初级 -11.2299
▁石头 -11.2299
▁北部 -11.23
our -11.2301
▁另一 -11.2301
▁海南 -11.2303
棚 -11.2304
▁相亲 -11.2304
▁好感 -11.2305
▁干扰 -11.2308
▁射 -11.2309
▁height -11.231
▁扑 -11.231
恒 -11.231
之处 -11.2311
▁湘 -11.2315
▁猴 -11.2318
IN -11.2319
ers -11.2322
匙 -11.2323
▁发音 -11.2326
▁冲动 -11.233
▁摘 -11.2332
▁一副 -11.2332
▁曾任 -11.2339
中学 -11.234
▁清华 -11.2346
▁前进 -11.2347
密度 -11.235
▁亲人 -11.2351
81 -11.2353
▁伦敦 -11.2358
╰ -11.2358
穷 -11.2361
ba -11.2361
▁转型 -11.2362
▁商城 -11.2364
▁加以 -11.2365
▁度假 -11.2366
较 -11.2368
▁否定 -11.2372
▁整天 -11.2372
▁铝 -11.2373
▁水分 -11.2374
▁5.1 -11.2376
84 -11.2379
▁深受 -11.2382
▁有何 -11.2384
one -11.2385
▁CD -11.2386
▁尽可能 -11.2387
▁恋 -11.2388
邪 -11.239
75 -11.2391
▁近代 -11.2398
▁垂直 -11.2399
欣 -11.24
▁绝大多数 -11.2401
主角 -11.2403
▁岩 -11.2404
▁Li -11.2406
▁1.8 -11.2407
净 -11.2409
▁坦克 -11.241
▁闭 -11.2411
漏 -11.2414
▁笑话 -11.2417
▁物种 -11.2417
▁注射 -11.2423
▁反馈 -11.2426
▁牌 -11.243
▁零售 -11.2431
狱 -11.2433
不清 -11.2435
73 -11.2442
▁残酷 -11.2446
▁过分 -11.2447
券 -11.2448
容易 -11.2449
鼎 -11.2449
▁49 -11.245
学历 -11.2452
▁讲究 -11.2453
▁高大 -11.2453
▁清水 -11.2457
▁切换 -11.2457
▁关节 -11.246
凌 -11.246
▁小白 -11.2472
▁对照 -11.2474
培 -11.2476
▁大四 -11.2477
▁触摸 -11.2477
西医 -11.2479
▁弹性 -11.2479
尸 -11.2481
▁胆 -11.2488
▁iOS -11.249
▁sh -11.2493
▁1981 -11.2494
▁单曲 -11.2496
▁蛋糕 -11.2497
事故 -11.2498
▁营造 -11.2504
役 -11.2504
▁啪 -11.2505
▁沉默 -11.251
▁单一 -11.251
荒 -11.2512
▁福建省 -11.2513
▁迎 -11.2514
▁流畅 -11.2518
▁极为 -11.2519
▁男主 -11.2519
坝 -11.2519
▁发言 -11.252
▁颗粒 -11.2522
▁薛 -11.2525
▁算了 -11.2526
▁通俗 -11.2526
▁士兵 -11.2527
▁凌晨 -11.253
痕 -11.2531
▁再见 -11.2532
▁高一 -11.2533
▁祖国 -11.2534
▁有过 -11.2534
▁精确 -11.2536
▁隔壁 -11.2536
▁撰写 -11.2544
▁期货 -11.2546
顷 -11.2548
▁无意 -11.255
▁栏目 -11.2552
▁你好 -11.2553
▁终端 -11.2554
▁视为 -11.2559
▁宁波 -11.2561
▁醒来 -11.2562
▁分泌 -11.2562
▁承办 -11.2564
▁符号 -11.2565
他 -11.2568
▁关联 -11.2572
▁韦 -11.2574
奈 -11.2576
▁轮 -11.2579
co -11.2579
供 -11.2583
▁教研 -11.2583
▁辉煌 -11.2586
盈 -11.2589
腻 -11.2593
ot -11.2594
▁奢侈 -11.2594
▁La -11.2594
参 -11.2595
▁割 -11.2595
酶 -11.2595
▁220 -11.2596
介 -11.2597
▁240 -11.2601
辣 -11.2602
▁竞 -11.2603
37 -11.2604
▁sp -11.2606
▁播出 -11.2606
身为 -11.2607
▁陪伴 -11.2607
▁首歌 -11.2607
弄 -11.2609
▁面包 -11.2609
min -11.261
し -11.2611
▁善于 -11.2612
杰伦 -11.2614
雾 -11.2614
▁东莞 -11.2616
▁档 -11.2618
▁外界 -11.2619
▁卫 -11.2619
▁平凡 -11.2621
▁尚未 -11.2623
42 -11.2624
▁知识产权 -11.2625
46 -11.2626
▁新建 -11.2627
▁父 -11.2627
旧 -11.2628
▁二人 -11.2629
▁baidu -11.263
▁黄河 -11.263
▁1979 -11.2635
▁澳门 -11.2641
辛 -11.2641
▁官网 -11.2641
▁发送 -11.2652
暖 -11.2652
▁爱上 -11.2654
▁一千 -11.266
充 -11.2666
97 -11.2668
▁起点 -11.2675
▁消灭 -11.2676
诀 -11.2678
▁De -11.2679
▁鸭 -11.268
兄 -11.2681
▁庄 -11.2682
▁一根 -11.2684
▁立场 -11.2685
▁家居 -11.2685
▁当今 -11.2687
麦 -11.2688
▁一顿 -11.2692
墓 -11.2695
▁多多 -11.2696
▁过滤 -11.2696
▁乐器 -11.2699
63 -11.2707
泵 -11.2707
▁饭店 -11.2708
状况 -11.2714
▁散文 -11.2717
▁担保 -11.2717
▁投诉 -11.272
▁阿尔 -11.272
▁趁 -11.272
▁河北省 -11.272
无疑 -11.2721
▁明年 -11.2723
▁分解 -11.2724
▁灌溉 -11.2726
当作 -11.2728
其妙 -11.273
番 -11.2735
▁第十四 -11.2737
▁贺 -11.2738
▁早餐 -11.2738
▁红包 -11.2743
ass -11.2743
▁急性 -11.2743
记录 -11.2744
▁嘲笑 -11.2745
▁颁发 -11.2746
▁不宜 -11.2747
▁遵守 -11.2749
▁带动 -11.275
国人 -11.275
▁酱油 -11.2754
▁艰难 -11.2755
▁文本 -11.2757
▁矿 -11.276
▁拼音 -11.276
▁绘制 -11.2761
▁礼仪 -11.2761
▁喜剧 -11.2761
▁老大 -11.2761
▁这类 -11.2761
▁恒 -11.2763
▁讨 -11.2765
▁狭 -11.2767
▁受害 -11.2768
鸿 -11.2771
▁转会 -11.2776
▁爸 -11.2777
什 -11.278
▁取决 -11.2782
宽 -11.2784
▁编号 -11.2784
廉 -11.2785
▁江湖 -11.2787
〈 -11.2793
工资 -11.2794
▁画画 -11.2794
▁挖掘 -11.2794
end -11.2796
▁非法 -11.2797
▁https -11.2799
▁li -11.28
▁启 -11.2802
▁夸张 -11.2803
勉强 -11.2804
▁科目 -11.2805
▁碳 -11.2805
▁59 -11.2806
▁招商 -11.2809
▁高二 -11.281
▁本土 -11.2813
▁培育 -11.2814
理论 -11.2817
▁画家 -11.2818
▁庆 -11.282
▁体积 -11.282
▁GT -11.2823
▁昨 -11.2823
▁堆 -11.2829
▁万一 -11.2831
▁转发 -11.2833
▁糟糕 -11.2836
▁便捷 -11.2839
▁医师 -11.2844
前任 -11.285
▁报纸 -11.2851
▁京东 -11.2851
▁二代 -11.2854
▁予以 -11.2857
影响 -11.2858
玲 -11.286
▁公式 -11.2861
▁跌 -11.2864
▁方针 -11.2867
钠 -11.2867
é -11.287
不去 -11.2871
速度 -11.2873
啊啊 -11.2873
▁糖尿病 -11.2873
▁计划生育 -11.2874
衷 -11.2874
▁口语 -11.2874
▁备 -11.2875
▁坐标 -11.2875
▁不见 -11.2877
▁密封 -11.2883
▁客户端 -11.2884
▁冷静 -11.2884
▁这才 -11.2884
应用 -11.2885
▁各有 -11.2885
▁光明 -11.2886
▁满满 -11.2886
旦 -11.2887
▁走过 -11.2891
专 -11.2892
保护 -11.2894
▁没见 -11.2896
▁大哥 -11.2897
拨 -11.2903
▁基督教 -11.2903
▁不去 -11.2905
详 -11.2912
▁济南 -11.2913
咨询 -11.2915
╭ -11.2915
39 -11.2921
▁中山 -11.2925
▁接待 -11.2926
▁附件 -11.2928
▁不变 -11.2928
▁过得 -11.2933
na -11.2933
▁县委 -11.2934
▁日记 -11.2935
▁睡着 -11.2937
▁高温 -11.2937
▁质保 -11.294
▁资深 -11.2942
▁财经 -11.2942
集团 -11.2947
▁产量 -11.2948
▁开学 -11.2948
▁暗恋 -11.2952
▁ル -11.2956
▁师傅 -11.2959
▁任意 -11.2961
▁树立 -11.2961
匹 -11.2962
▁愈 -11.2963
左右 -11.2963
售 -11.2965
▁繁殖 -11.2966
▁欢乐 -11.2967
抱 -11.2967
ri -11.2968
▁川 -11.2971
▁星座 -11.2972
▁姚 -11.2973
仔 -11.2974
▁医学院 -11.2975
ell -11.2975
▁纺织 -11.2975
$ -11.2976
op -11.2979
▁中学生 -11.2981
▁诚 -11.2981
▁玩意 -11.2988
▁万吨 -11.2988
▁摆脱 -11.2995
▁三天 -11.2995
▁功夫 -11.3001
泳 -11.3001
▁净化 -11.3005
▁国家队 -11.3005
罐 -11.3008
允 -11.3008
▁高压 -11.3008
ut -11.3012
▁安徽省 -11.3016
熊 -11.3017
▁享有 -11.3019
▁金额 -11.3019
▁外表 -11.3021
▁厂家 -11.3022
▁附属 -11.3023
▁差点 -11.3027
▁仪式 -11.3028
▁1949 -11.3029
▁半天 -11.3033
▁疗 -11.3034
▁Le -11.3035
▁出租车 -11.3036
▁53 -11.3039
▁一下子 -11.3041
▁XX -11.3042
▁绑 -11.3045
讯 -11.3046
▁签订 -11.3048
▁情人 -11.3057
▁穷人 -11.3059
形象 -11.306
▁绩效 -11.306
▁泪 -11.3062
▁门槛 -11.3062
缝 -11.3063
▁贪 -11.3067
▁液体 -11.3067
▁暗示 -11.3068
▁几千 -11.3068
▁半夜 -11.3069
▁举例 -11.307
▁肖 -11.3072
爆 -11.3073
▁有助 -11.3073
▁聚集 -11.3075
▁小事 -11.3076
GA -11.3079
▁预期 -11.3079
蕾 -11.308
仕 -11.308
▁App -11.308
标准化 -11.3082
▁劳 -11.3085
▁下车 -11.3088
▁三四 -11.3088
╮ -11.3089
▁刑事 -11.3092
▁康复 -11.3092
缸 -11.3092
▁五百 -11.3093
支持 -11.3093
li -11.3095
▁国籍 -11.3098
妆 -11.3099
▁惊喜 -11.3099
▁88 -11.3101
▁控 -11.3101
▁To -11.3103
爵 -11.3104
▁资助 -11.3109
人民 -11.3112
赖 -11.3117
▁指令 -11.3122
▁形势 -11.3126
距离 -11.3126
▁贵族 -11.3128
▁u -11.3133
▁事后 -11.3134
▁反思 -11.3135
酬 -11.3135
▁处罚 -11.3136
▁呵 -11.3138
▁持有 -11.3138
▁锅 -11.3141
▁积分 -11.3141
▁困惑 -11.3145
基础 -11.3148
熙 -11.3153
▁有种 -11.3158
▁遥 -11.316
撞 -11.3162
▁清理 -11.3165
82 -11.3169
ul -11.3169
ss -11.3171
▁交互 -11.3171
▁当选 -11.3176
跃 -11.3177
▁法制 -11.3181
▁座位 -11.3183
▁用法 -11.3183
▁歪 -11.3186
停 -11.3188
▁使命 -11.319
般 -11.319
▁事务 -11.3193
▁镜 -11.3196
▁志愿 -11.3202
▁辽 -11.3202
▁理智 -11.3204
▁人力 -11.3204
▁延长 -11.3205
渠 -11.3205
▁柯 -11.3209
▁不像 -11.3211
▁54 -11.3212
预 -11.3213
22 -11.3217
狮 -11.3218
鹰 -11.3219
▁常识 -11.3223
▁帧 -11.3223
ス -11.3224
▁签约 -11.3224
债 -11.3226
▁电机 -11.3226
计划 -11.3228
▁代谢 -11.3231
宪 -11.3235
▁筹 -11.3237
▁写出 -11.3238
▁引入 -11.3238
植 -11.3239
悟 -11.3241
MA -11.3241
疗 -11.3245
▁多元 -11.3247
▁立马 -11.3247
▁奖学 -11.325
▁洛阳 -11.3251
走路 -11.3252
脏 -11.3253
ton -11.3253
▁付 -11.3254
▁还原 -11.3254
▁挑选 -11.3254
▁近期 -11.3256
▁钢铁 -11.3259
衡 -11.3265
▁学长 -11.3266
电脑 -11.327
▁诗词 -11.3271
▁书中 -11.3273
▁延续 -11.3276
▁指示 -11.3278
▁重建 -11.3282
SE -11.3283
41 -11.3286
▁布置 -11.3286
▁在职 -11.3289
▁固 -11.329
▁摩托车 -11.3292
▁水准 -11.3296
▁骨干 -11.3296
找 -11.33
▁128 -11.33
▁瓶 -11.3303
▁终身 -11.3304
▁竞技 -11.3305
巷 -11.3305
▁旗舰 -11.3307
▁浇 -11.3308
▁管道 -11.3313
▁科幻 -11.3313
精神 -11.3314
高新技术 -11.3314
▁特效 -11.3318
▁粤语 -11.3318
▁开头 -11.3321
副 -11.3322
▁谈谈 -11.3323
▁震撼 -11.3327
▁披 -11.3327
▁灌 -11.3328
▁希腊 -11.3329
▁遮 -11.333
▁师生 -11.3331
▁采集 -11.3333
▁包容 -11.334
▁百姓 -11.334
朵 -11.3347
▁生死 -11.3348
▁入学 -11.335
▁缴 -11.335
▁模具 -11.335
感谢 -11.3351
▁坚强 -11.3352
斯特 -11.3353
▁调研 -11.3354
奉 -11.3356
禄 -11.3356
名称 -11.3357
↓ -11.3357
▁减轻 -11.336
▁con -11.336
▁赐 -11.3361
◎ -11.3361
委会 -11.3362
▁Co -11.3362
▁出自 -11.3363
▁跟踪 -11.3364
叹 -11.3365
▁碗 -11.3367
▁有名 -11.3372
▁澳洲 -11.3373
▁捧 -11.3374
▁978750 -11.3376
某 -11.3381
葬 -11.3382
▁靠近 -11.3387
漆 -11.3388
IS -11.3388
当做 -11.339
▁场地 -11.3394
▁舰 -11.3394
绪 -11.3398
▁教会 -11.3401
耶 -11.3401
▁3.4 -11.3402
▁邮 -11.3406
▁桌子 -11.3406
▁丹 -11.341
察 -11.341
辆 -11.3413
▁瑞士 -11.3415
触 -11.3419
▁续 -11.342
▁撑 -11.342
▁长久 -11.3425
▁平静 -11.3428
▁骗子 -11.3428
▁我用 -11.3429
▁驾 -11.3431
傲 -11.3439
教学 -11.3441
▁长安 -11.3442
▁不明 -11.3443
▁96 -11.3446
夕 -11.3447
郊 -11.3449
▁二维 -11.3452
ard -11.3454
▁伴随 -11.3454
▁字体 -11.3454
▁开发区 -11.3456
▁0.1 -11.3456
▁猪肉 -11.3459
▁密切 -11.3463
栋 -11.3463
▁孔子 -11.3465
▁做成 -11.3469
▁170 -11.3473
▁′ -11.3473
▁巫 -11.3478
▁列车 -11.3479
▁影像 -11.348
不动 -11.3482
跳 -11.3486
▁节省 -11.3486
科技进步 -11.3491
陷 -11.3492
▁许可证 -11.3493
瑟 -11.3493
る -11.3495
徽 -11.3496
▁懵 -11.3498
▁分支 -11.3503
▁内在 -11.3504
▁上门 -11.3505
▁委屈 -11.3508
▁双手 -11.3513
▁两句 -11.3513
缩 -11.3515
喻 -11.3515
▁气象 -11.3519
州市 -11.3519
悉 -11.3519
▁照明 -11.3522
姜 -11.3535
▁约束 -11.3538
一天 -11.3543
▁酒吧 -11.3547
分之一 -11.3548
▁弘 -11.3549
困难 -11.3549
▁土豆 -11.355
▁内蒙古 -11.3551
蓄 -11.3551
▁默认 -11.3552
坤 -11.356
▁一会儿 -11.356
▁1200 -11.3563
▁盯 -11.3566
▁试卷 -11.3567
▁一夜 -11.3574
之内 -11.3576
ty -11.3577
▁试题 -11.3582
▁激发 -11.3583
▁实质 -11.3583
构 -11.3585
▁置 -11.3586
▁清华大学 -11.3588
▁可行 -11.3588
万元 -11.3589
▁匹配 -11.359
▁不锈钢 -11.3592
日起 -11.3593
▁娶 -11.3594
验 -11.3595
ˋ -11.3595
▁寿命 -11.3595
be -11.3595
▁名牌 -11.3597
▁学霸 -11.3597
▁憋 -11.3599
▁山区 -11.36
棉 -11.3602
勺 -11.3603
▁国土 -11.3604
裁 -11.3604
▁华人 -11.3607
▁占领 -11.3608
▁登录 -11.3608
彦 -11.3612
▁计量 -11.3613
▁仇 -11.3614
▁孕妇 -11.3616
炭 -11.3616
▁颜 -11.3616
▁一瞬间 -11.3617
▁脑袋 -11.3617
甘 -11.3619
琪 -11.3619
涵 -11.3624
▁老家 -11.3625
▁射击 -11.3625
▁制成 -11.3631
▁道具 -11.3633
▁大门 -11.3634
年级 -11.3635
拿 -11.3636
锤 -11.3637
▁砂 -11.3638
弗 -11.3641
纱 -11.3643
▁细腻 -11.3645
畏 -11.3646
CC -11.3646
▁个别 -11.3648
▁手工 -11.3651
▁讲师 -11.3655
74 -11.3657
▁魅族 -11.3657
▁扭 -11.3657
▁户外 -11.3661
▁激 -11.3662
▁封面 -11.3663
と -11.3665
▁登陆 -11.3666
▁光学 -11.3668
▁140 -11.3669
▁封闭 -11.3674
▁扩张 -11.3674
................ -11.3675
▁坚决 -11.3677
▁活力 -11.3677
▁翅 -11.3681
▁体力 -11.3682
僧 -11.3682
▁物理学 -11.3686
▁4.5 -11.3687
▁家境 -11.369
▁拼命 -11.369
71 -11.3694
雀 -11.3697
▁装逼 -11.3699
▁裙 -11.37
▁黑人 -11.37
▁毫不 -11.3703
娃 -11.3703
30 -11.3707
▁节约 -11.3709
▁休 -11.3713
队伍 -11.3715
▁敲 -11.3722
▁97 -11.3722
奏 -11.3734
▁称之 -11.3734
彻 -11.3735
▁後 -11.3736
▁批 -11.3736
▁雅思 -11.374
▁1958 -11.3742
か -11.3745
▁兽 -11.3745
▁水库 -11.3745
▁辣椒 -11.3749
▁便于 -11.3749
▁保守 -11.375
绍 -11.3752
▁中原 -11.3753
▁6.2 -11.3754
砂 -11.3755
▁說 -11.3755
▁研究员 -11.3755
▁规 -11.3756
▁路由 -11.3756
▁第十五 -11.3757
乙 -11.3759
に -11.376
▁染 -11.3761
▁行李 -11.3766
▁穆 -11.377
○ -11.377
▁啤酒 -11.377
▁占据 -11.3771
▁百万 -11.3776
▁美元 -11.378
▁这份 -11.378
▁命 -11.3782
▁家属 -11.3784
▁二手 -11.3785
▁农田 -11.3785
类型 -11.3793
35 -11.3804
▁乐观 -11.3804
▁DDR -11.3805
▁棒 -11.3805
酯 -11.3806
▁春天 -11.3808
▁副会 -11.3811
▁不爽 -11.3814
▁生涯 -11.3814
▁潜在 -11.3816
锡 -11.3818
▁嘲讽 -11.382
▁阴阳 -11.382
▁世界杯 -11.382
▁詹姆斯 -11.382
绘 -11.3821
▁抽烟 -11.3822
昭 -11.3825
▁精灵 -11.383
▁走上 -11.3831
骑 -11.3831
▁共产党 -11.3833
ity -11.3836
▁奇迹 -11.3839
亏 -11.3841
▁法院 -11.3843
▁审 -11.3843
▁信誉 -11.3849
▁火箭 -11.3853
▁法定 -11.3855
▁沾 -11.3856
▁ノ -11.3857
▁几十年 -11.3858
▁很长 -11.386
▁熟练 -11.3861
▁平板 -11.3863
sh -11.3865
〉 -11.3867
▁鼻子 -11.3868
▁涨 -11.3875
说出 -11.3876
▁变态 -11.3877
▁命题 -11.3879
激 -11.3879
▁传送 -11.388
▁计 -11.3881
vi -11.3882
▁癌症 -11.3886
钉 -11.3888
鸭 -11.3888
ab -11.3889
▁任职 -11.389
▁专栏 -11.389
决赛 -11.3891
▁摇 -11.3892
▁诉讼 -11.3892
忌 -11.3894
▁琴 -11.3898
AS -11.3899
▁You -11.3901
▁花生 -11.3902
▁蜜 -11.3912
▁排列 -11.3912
▁几率 -11.3913
崖 -11.3916
▁犹豫 -11.3918
采 -11.3922
▁涵盖 -11.3926
槽 -11.3927
利亚 -11.393
▁三维 -11.3931
▁懒得 -11.3933
▁子宫 -11.3935
ant -11.3937
牢 -11.3939
▁提名 -11.394
▁用药 -11.3945
▁码 -11.3946
▁开口 -11.3946
空间 -11.3947
29 -11.3947
▁总经理 -11.3951
吻 -11.3951
冈 -11.3955
▁low -11.3958
▁纪录片 -11.3959
家中 -11.3959
ong -11.3959
ee -11.396
▁进而 -11.3961
▁样本 -11.3962
舱 -11.3966
▁宅 -11.3966
法规 -11.3969
▁评审 -11.397
吏 -11.397
て -11.3972
▁衡量 -11.3974
▁LOL -11.3974
▁复印 -11.3975
▁烫 -11.3981
▁爹 -11.3982
ice -11.3984
▁监察 -11.3985
培训 -11.3985
▁紧密 -11.3986
▁过多 -11.3986
▁散热 -11.3988
变更 -11.3991
▁充足 -11.3993
葱 -11.3995
▁规章 -11.3997
▁冬季 -11.3999
抛弃 -11.4001
▁名为 -11.4002
▁操 -11.4005
▁水稻 -11.4006
▁68 -11.4008
▁写字 -11.401
▁一笔 -11.401
▁同类 -11.401
资金 -11.4011
▁步行 -11.4011
▁摇滚 -11.4013
▁指责 -11.4017
▁紧急 -11.4018
▁强奸 -11.4022
▁抗日 -11.4022
▁局长 -11.4026
贼 -11.4028
▁祖 -11.4028
植物 -11.4029
▁以往 -11.4031
▁充实 -11.4034
▁ex -11.4038
▁气息 -11.4041
▁端口 -11.4041
← -11.4044
▁遵循 -11.4044
▁环境保护 -11.4046
▁同情 -11.4046
87 -11.4047
▁买卖 -11.4048
▁イ -11.4049
豹 -11.4052
▁Mac -11.4056
▁left -11.4056
▁圣诞 -11.4057
▁蒜 -11.4058
▁崛 -11.4058
▁大都 -11.406
▁携带 -11.4065
饮 -11.4066
▁机身 -11.4072
贫 -11.4073
ES -11.4073
附 -11.4074
▁上传 -11.4079
▁备注 -11.408
95 -11.4082
▁菜单 -11.4085
贷 -11.4086
▁流浪 -11.4093
▁简洁 -11.4096
▁转载 -11.4101
▁凤 -11.4102
菊 -11.4104
▁杰出 -11.4105
▁两条 -11.4106
▁Java -11.4106
▁邻居 -11.4108
800 -11.411
▁约翰 -11.4114
▁王子 -11.4115
ti -11.4124
▁中共党 -11.4125
▁考验 -11.4126
髓 -11.4127
▁钙 -11.4128
▁哥 -11.4129
▁第一百 -11.413
▁太过 -11.4133
弦 -11.4136
▁现为 -11.4137
▁沟 -11.4139
▁号线 -11.4143
▁良心 -11.4144
ㄧ -11.4144
ER -11.4146
▁地步 -11.4147
▁二战 -11.4151
▁战役 -11.4151
▁情报 -11.4152
▁臭 -11.4155
▁大事 -11.4157
47 -11.4158
▁搅拌 -11.4159
▁温州 -11.4161
大部分 -11.4162
▁三级 -11.4168
da -11.4173
▁到位 -11.4177
▁穿衣 -11.418
▁俄 -11.418
▁密度 -11.4181
▁少量 -11.4182
▁验收 -11.4182
钦 -11.4187
不及 -11.4188
▁硬化 -11.4189
▁堵 -11.4192
▁封建 -11.4195
▁九章 -11.4201
▁孝 -11.4201
避 -11.4204
有限责任 -11.4206
▁非要 -11.4209
▁盗版 -11.421
ear -11.4214
▁高科技 -11.4223
▁精选 -11.4224
▁附加 -11.4225
▁72 -11.4229
嫁 -11.423
▁先锋 -11.4231
▁演变 -11.4232
劫 -11.4235
▁民营 -11.4238
▁市民 -11.4239
▁好事 -11.4239
CH -11.424
培养 -11.4244
▁屁股 -11.4244
▁昆明 -11.4245
▁手续 -11.4245
或 -11.4248
▁压抑 -11.4249
▁必备 -11.4249
▁恋人 -11.425
▁延 -11.4251
▁笨 -11.4252
▁can -11.4254
▁讲话 -11.4256
km -11.4257
▁漏 -11.4258
外科 -11.4259
乏 -11.4262
代表大会 -11.4262
醇 -11.4262
塌 -11.4263
▁出血 -11.4267
▁坐落 -11.4269
▁熬夜 -11.4271
▁崇拜 -11.4272
▁出席 -11.4272
▁前途 -11.4276
▁公正 -11.4276
▁胎 -11.4276
▁捞 -11.4277
▁礼貌 -11.4277
匠 -11.4278
▁answer -11.4279
▁治愈 -11.428
不开 -11.4282
▁鲁迅 -11.4283
尝 -11.4284
▁选项 -11.4286
▁央视 -11.4288
▁桌面 -11.4289
ru -11.4289
ide -11.429
▁众人 -11.4292
▁婚纱 -11.4295
▁草原 -11.4297
报告 -11.4297
夷 -11.4302
婉 -11.4306
誉为 -11.4307
▁失眠 -11.4309
▁交叉 -11.431
98 -11.4311
▁78 -11.4312
▁高档 -11.4314
▁响 -11.4315
▁电视机 -11.4315
▁潜力 -11.4315
▁王朝 -11.4318
▁永久 -11.4319
▁全家 -11.432
▁5.3 -11.432
▁省内 -11.4321
▁浓度 -11.4324
▁批发 -11.4324
嫩 -11.4325
▁向前 -11.4325
不算 -11.4327
柔 -11.4329
援 -11.4329
▁供电 -11.433
串 -11.4332
▁男孩子 -11.4334
▁目光 -11.4337
▁粥 -11.4338
制品 -11.4341
old -11.4342
94 -11.4342
▁后台 -11.4344
▁借助 -11.4345
▁not -11.4349
需 -11.435
日军 -11.4355
▁拒 -11.4356
▁Ma -11.4356
▁赔 -11.4357
▁风云 -11.4357
郡 -11.4358
▁background -11.4359
▁井 -11.4359
ian -11.4361
▁揭 -11.4362
▁战胜 -11.4363
▁太子 -11.4366
〕 -11.4367
▁列表 -11.4369
▁高达 -11.4369
▁中级 -11.437
▁耍 -11.4374
▁舌 -11.4374
▁期限 -11.4379
▁巅峰 -11.4379
神经 -11.438
▁路径 -11.4381
▁从小到 -11.4384
袖 -11.4385
涩 -11.4385
▁卧槽 -11.4385
▁黑白 -11.4387
▁Ch -11.4392
▁分散 -11.4393
赏 -11.4394
▁紧 -11.4399
▁温泉 -11.44
〔 -11.4401
▁kg -11.4402
▁特有 -11.4407
偶 -11.4408
▁股权 -11.4411
▁军人 -11.4411
67 -11.4413
▁光驱 -11.4417
▁溶液 -11.4417
▁一股 -11.4419
胺 -11.442
▁主干 -11.4421
▁日益 -11.4421
▁强势 -11.4428
▁回收 -11.4432
▁集合 -11.4432
▁co -11.4433
桐 -11.4435
▁欺骗 -11.4439
数据 -11.4441
ill -11.4442
▁电器 -11.4447
61 -11.4454
▁高潮 -11.4457
▁哇 -11.4457
怨 -11.4459
▁难得 -11.4461
瑶 -11.4469
▁更名 -11.4469
▁租赁 -11.4469
▁高中生 -11.4471
▁卧 -11.4472
▁冷却 -11.4472
▁年初 -11.4473
丧 -11.4477
默 -11.4479
过程 -11.4481
▁发票 -11.4484
▁反抗 -11.4485
▁部落 -11.4489
▁爽 -11.4489
挂 -11.449
▁亲密 -11.4491
▁垂 -11.4492
ve -11.4493
趋势 -11.4493
替 -11.4494
▁提起 -11.4495
▁内外 -11.4496
▁生物学 -11.4502
▁守护 -11.4504
▁审判 -11.4505
▁向往 -11.4509
̀ -11.4515
53 -11.4515
▁产品质量 -11.4515
奶奶 -11.4516
▁消毒 -11.4516
▁突发 -11.4516
▁入口 -11.4519
▁起步 -11.4525
▁导向 -11.4527
引擎 -11.4527
▁do -11.4528
▁参照 -11.4529
AT -11.453
弊 -11.453
十岁 -11.4531
女孩 -11.4533
▁经济林 -11.4535
哟 -11.4536
▁停留 -11.4537
▁团购 -11.4539
ix -11.4543
▁对称 -11.4544
喵 -11.4544
▁转向 -11.4544
硕 -11.455
ist -11.4551
▁太大 -11.4555
▁处处 -11.4556
监 -11.4558
▁护士 -11.4563
革 -11.4567
▁询问 -11.4568
▁修正 -11.4576
肚 -11.4577
资格 -11.4578
▁研讨 -11.4578
▁地产 -11.4578
▁发电 -11.458
▁友情 -11.458
BA -11.458
AN -11.4589
咪 -11.4589
阿 -11.4589
▁珍贵 -11.4591
▁古人 -11.4591
▁发酵 -11.4595
▁工科 -11.46
利用 -11.46
300 -11.4601
克斯 -11.4602
▁鲜明 -11.4605
畔 -11.4607
裙 -11.4607
▁第十六 -11.461
肩 -11.4612
▁大奖 -11.4612
奋 -11.4617
▁秉承 -11.4617
▁高低 -11.4619
▁失恋 -11.4625
▁订单 -11.4626
▁爱心 -11.4627
▁TM -11.4628
▁zh -11.463
▁霸 -11.463
▁书写 -11.4631
学科 -11.4631
耕 -11.4632
▁住宿 -11.4632
▁三分 -11.4633
▁大声 -11.4635
陆 -11.4635
▁驱 -11.4637
▁top -11.4637
▁参观 -11.4638
▁建国 -11.4639
▁生活费 -11.4639
姬 -11.4642
net -11.4646
▁传染 -11.4649
▁户口 -11.4656
▁Linux -11.4657
▁鹿 -11.4658
▁聚会 -11.4658
▁楼下 -11.4658
▁钓鱼 -11.4658
▁综艺 -11.4662
▁天堂 -11.4663
懒 -11.4667
▁做人 -11.4671
▁立足 -11.4674
▁涉 -11.4674
▁一圈 -11.4677
la -11.4678
▁q -11.4678
▁嫌疑 -11.4679
污 -11.468
电路 -11.4685
▁联合国 -11.4685
姑 -11.4685
ni -11.4686
侵犯 -11.469
干部 -11.469
▁春节 -11.4693
一体化 -11.4693
▁-------- -11.4694
▁前言 -11.4694
柳 -11.4696
不通 -11.4696
▁HR -11.47
▁占有 -11.4707
邑 -11.4708
▁浪 -11.471
▁一小 -11.4711
▁真理 -11.4714
▁奉献 -11.4714
▁污 -11.4715
▁钻石 -11.4716
翠 -11.4719
坞 -11.4721
▁无知 -11.4722
▁农产品 -11.4726
ha -11.4732
▁讽刺 -11.4733
▁两大 -11.474
课程 -11.4742
▁修建 -11.4743
▁观赏 -11.4746
▁正当 -11.4746
荷 -11.475
瓷 -11.4752
▁产物 -11.4752
▁蜀 -11.4757
▁防水 -11.4758
▁屁 -11.4758
啪 -11.4759
▁惟 -11.4761
▁明代 -11.4762
▁配音 -11.4765
绳 -11.4767
▁见解 -11.4767
▁人心 -11.4769
▁重庆市 -11.4772
▁五月 -11.4772
宣 -11.4777
▁任教 -11.4777
▁入侵 -11.4779
▁太平 -11.4783
鹤 -11.4784
仇 -11.4787
价格 -11.4787
▁演绎 -11.4787
钩 -11.4791
▁庙 -11.4792
▁自来水 -11.4793
▁名义 -11.4795
▁数值 -11.4795
执 -11.4796
忍 -11.48
瓣 -11.4801
▁心思 -11.4801
▁周年 -11.4807
混 -11.4809
厨 -11.4812
▁弟子 -11.4814
▁数十 -11.4814
碱 -11.4814
驻 -11.4815
▁足以 -11.4816
系数 -11.4818
▁地狱 -11.4818
▁被动 -11.4819
▁场面 -11.482
▁老妈 -11.4821
bit -11.4821
▁身心 -11.4822
▁伏 -11.4823
▁盲目 -11.4825
▁黏 -11.4825
斌 -11.4827
弯 -11.4827
▁备用 -11.4827
▁傅 -11.4828
摸 -11.4828
▁陕西省 -11.4829
▁挨 -11.483
力学 -11.4834
▁农作 -11.4834
ig -11.4836
▁水中 -11.4838
▁备案 -11.4839
▁62 -11.484
▁说好 -11.4842
殊 -11.4843
▁马克思主义 -11.4845
▁经纪 -11.4848
▁国企 -11.4851
PS -11.4851
▁长城 -11.4852
堤 -11.4856
棋 -11.4856
▁幻 -11.4856
▁水晶 -11.486
▁解除 -11.4862
▁桥梁 -11.4863
▁代言 -11.4865
▁王国 -11.4866
▁果断 -11.4868
▁养老 -11.4869
oo -11.4871
▁谷歌 -11.4873
▁好人 -11.4874
▁听力 -11.4874
▁云南省 -11.4874
▁巧克力 -11.4877
▁一手 -11.4878
▁稀 -11.4879
▁棋 -11.488
▁your -11.4881
▁最强 -11.4881
慰 -11.4883
汽 -11.4883
▁6.1 -11.4883
奴 -11.4885
▁再生 -11.4886
避免 -11.4887
区域 -11.4888
▁扇 -11.4895
▁戳 -11.4895
▁海上 -11.4896
▁召唤 -11.4897
▁new -11.4897
▁8. -11.4899
▁现已 -11.4902
▁先说 -11.4903
浮 -11.4903
▁清除 -11.4903
杉 -11.4904
▁爱人 -11.4905
▁索尼 -11.4907
▁见过 -11.4907
▁收拾 -11.4908
▁七年 -11.4911
巢 -11.4914
go -11.4914
▁饼 -11.4921
. -11.4921
▁700 -11.4924
de -11.4924
▁住院 -11.4924
▁结尾 -11.4925
▁事务所 -11.4926
畜 -11.4926
▁width -11.4928
▁姐妹 -11.4928
▁逃避 -11.4933
▁感激 -11.4933
佑 -11.4938
▁此类 -11.4939
▁死刑 -11.4939
▁政权 -11.494
▁账 -11.4945
▁lol -11.495
pe -11.495
▁2.0 -11.4956
尉 -11.4957
诶 -11.4958
▁胡萝卜 -11.4959
▁win -11.4962
▁焊接 -11.4962
赋 -11.4965
▁甘肃 -11.4965
▁下属 -11.497
贺 -11.4971
▁旋律 -11.4971
▁慈善 -11.4972
▁发型 -11.4975
▁推理 -11.4975
▁热点 -11.4976
▁美学 -11.4978
才能 -11.4982
▁直属 -11.4982
▁外婆 -11.4984
RO -11.4985
▁几点 -11.4985
▁椭圆 -11.4985
▁攀 -11.4986
虹 -11.4986
▁评选 -11.4994
深处 -11.4996
Hz -11.5
▁温和 -11.5001
之地 -11.5002
0.0 -11.5002
▁大胆 -11.5005
☆ -11.5006
▁输送 -11.5007
▁溶 -11.5011
▁余人 -11.5011
ure -11.5012
租 -11.5013
▁水面 -11.5017
谜 -11.5018
▁57 -11.5018
棍 -11.5019
sa -11.502
▁配件 -11.5022
40 -11.5032
▁燕 -11.5034
▁兰州 -11.5037
▁陷 -11.5038
▁4.4 -11.5039
▁客房 -11.5041
原因 -11.5042
▁丧失 -11.5042
▁十多 -11.5044
ラ -11.5044
危机 -11.5048
▁猜测 -11.5049
▁裸 -11.5054
▁自称 -11.5054
▁导弹 -11.5055
▁益 -11.5056
▁工会 -11.5063
▁同桌 -11.5064
学名 -11.5067
▁透露 -11.5067
▁勤奋 -11.5068
▁客运 -11.507
辅 -11.5072
▁HD -11.5072
▁开创 -11.5076
▁降水 -11.5076
▁亏 -11.5079
▁纳米 -11.5081
▁盗 -11.5082
▁ta -11.5083
▁报考 -11.5088
誉 -11.5088
▁只想 -11.5088
▁买个 -11.509
▁家用 -11.5091
堪 -11.5094
▁圆桌 -11.5094
ow -11.5095
AD -11.5097
CM -11.5098
▁Con -11.5098
▁动机 -11.5099
▁转基因 -11.51
▁疗效 -11.5103
婴 -11.5104
▁决 -11.5105
▁底层 -11.5105
▁---------- -11.5105
▁高清 -11.5106
▁2017 -11.5107
▁底线 -11.5108
▁特种 -11.511
▁Al -11.5112
▁谭 -11.5112
玄 -11.5112
▁相声 -11.5113
▁Se -11.5115
▁业内 -11.5117
▁高峰 -11.5118
▁正版 -11.5121
斯坦 -11.5122
不时 -11.5124
▁金牌 -11.5125
趟 -11.5125
▁武术 -11.5126
-------------- -11.5127
▁铸 -11.5134
▁作物 -11.5136
坚 -11.5137
▁叶片 -11.5138
69 -11.5139
蜂 -11.5141
障 -11.5143
▁贫穷 -11.515
▁商人 -11.5154
▁波动 -11.5154
▁部件 -11.5154
函 -11.5154
▁顿 -11.5155
▁渗透 -11.5155
▁矿物 -11.5156
享 -11.5157
▁近几年 -11.5157
惨 -11.5157
▁as -11.5157
金属 -11.5158
▁均衡 -11.5162
▁向上 -11.5162
▁八卦 -11.5163
▁假期 -11.5163
う -11.5163
▁公关 -11.5163
亦 -11.5164
工艺 -11.5166
刃 -11.5166
▁终止 -11.5166
▁一事 -11.5167
▁女士 -11.5169
晚上 -11.5171
▁脆弱 -11.5171
▁武装 -11.5174
▁零件 -11.5176
± -11.5178
▁哄 -11.5178
佐 -11.5179
▁献 -11.5182
▁抬 -11.5183
ME -11.5184
▁浸 -11.5186
▁晕 -11.5186
业务 -11.5187
▁汗 -11.5188
灶 -11.5188
卖家 -11.519
呆 -11.5193
▁这条 -11.5193
way -11.5194
▁病情 -11.5195
▁在校 -11.5195
建筑 -11.5197
▁赫 -11.5199
▁见识 -11.52
爱好 -11.52
▁淘 -11.5201
▁尼玛 -11.5203
▁love -11.5203
day -11.5205
▁高血压 -11.5205
判 -11.5206
▁多大 -11.5207
▁分管 -11.5207
▁求助 -11.5207
▁新兴 -11.5208
氮 -11.5212
▁奥运 -11.5216
▁响应 -11.5219
0000 -11.5219
▁杰 -11.5223
▁主场 -11.5226
.09 -11.5226
▁储 -11.5228
338 -11.5228
▁忌 -11.5229
▁南昌 -11.5229
笼 -11.523
用户 -11.5231
消息 -11.5232
ˊ -11.5237
兔 -11.5237
▁沿海 -11.5237
no -11.5239
AP -11.5244
▁拥抱 -11.5244
勉 -11.5247
▁美味 -11.5247
▁效益 -11.5248
▁外科 -11.525
▁灯光 -11.5251
▁Web -11.5257
▁76 -11.526
▁储存 -11.5263
发挥 -11.5263
▁UI -11.5264
▁国防 -11.5265
▁公告 -11.5266
▁大牛 -11.5268
▁话语 -11.527
▁畅销 -11.5271
▁真人 -11.5273
▁何必 -11.5275
▁素材 -11.5276
▁争论 -11.5278
ud -11.528
▁宽带 -11.5281
七八 -11.5284
▁船舶 -11.5285
醋 -11.5286
▁作词 -11.5286
堆 -11.5288
▁or -11.529
蕴 -11.5291
▁复读 -11.5292
▁毁灭 -11.5295
进士 -11.5296
▁摩擦 -11.5297
障碍 -11.5298
▁AC -11.5304
媒 -11.5305
电站 -11.5307
▁容积 -11.5309
▁广州市 -11.5312
▁务 -11.5312
贩 -11.5315
▁决赛 -11.5317
太太 -11.5319
▁极度 -11.5322
▁瑞典 -11.5324
▁一期 -11.5327
▁阶 -11.5327
▁编码 -11.5329
▁泡沫 -11.5331
▁直线 -11.5332
乃 -11.5334
▁媳妇 -11.5335
▁惩罚 -11.5336
▁警方 -11.5337
▁ul -11.534
▁几何 -11.5341
▁He -11.5342
▁排队 -11.5345
▁中考 -11.5348
▁产权 -11.5348
成功 -11.5353
还 -11.5354
▁光线 -11.5356
▁付款 -11.5358
bi -11.5358
▁出生地 -11.5358
▁看不起 -11.536
▁86 -11.5365
▁隔离 -11.537
芯 -11.537
▁text -11.5371
家伙 -11.5373
▁外汇 -11.5373
▁顶尖 -11.5374
农业 -11.5375
▁累计 -11.5376
▁决心 -11.5382
▁狮子 -11.5385
▁兼容 -11.539
▁业主 -11.5392
▁网游 -11.5394
ary -11.5394
学生 -11.5397
▁订 -11.5398
迅 -11.5399
▁7.2 -11.5401
▁裤子 -11.5402
▁繁荣 -11.5403
总局 -11.5404
▁夫人 -11.5404
占 -11.5406
▁沃 -11.5407
▁机遇 -11.5407
ji -11.5408
▁童话 -11.5416
▁第十七 -11.5417
▁额外 -11.5418
科技 -11.5421
▁宪法 -11.5425
▁筛选 -11.5425
▁葛 -11.5429
▁水质 -11.5432
▁震惊 -11.5435
性能 -11.5436
▁报刊 -11.5437
▁转让 -11.5437
ms -11.544
▁分成 -11.5442
机械 -11.5443
▁兼任 -11.5443
贯 -11.5446
su -11.5447
▁应有 -11.5448
横 -11.5448
▁拿来 -11.5451
职责 -11.5452
贡献 -11.5453
▁z -11.5459
担 -11.5459
▁ト -11.546
▁摩 -11.5461
▁xx -11.5468
▁前段 -11.5468
忘 -11.5469
▁监 -11.5475
▁排斥 -11.5475
▁煤炭 -11.5476
▁倡导 -11.5477
▁全民 -11.5477
▁粮 -11.5478
.07 -11.5479
▁厌恶 -11.5479
▁学生会 -11.5481
▁详情 -11.5481
▁月亮 -11.5481
▁户型 -11.5481
▁太高 -11.5482
▁感慨 -11.5486
▁劣 -11.5493
▁海关 -11.5495
▁心想 -11.5495
讶 -11.5497
▁勤 -11.5498
因素 -11.5498
帕 -11.5501
▁进展 -11.5502
▁暧昧 -11.5505
▁早点 -11.5506
疾 -11.5506
▁诱惑 -11.5509
31 -11.5512
▁十几年 -11.5518
▁闪光 -11.5521
▁纠正 -11.5521
▁0.2 -11.5523
狐 -11.5524
扇 -11.5527
▁精美 -11.5529
▁司令 -11.553
▁裂 -11.5531
吟 -11.5536
▁好坏 -11.5537
鑫 -11.5537
▁老实 -11.5538
▁出售 -11.5539
风格 -11.554
▁娇 -11.5541
tr -11.5544
▁首席 -11.5544
▁63 -11.5544
▁求职 -11.5549
杠 -11.5551
▁领袖 -11.5553
▁人身 -11.5554
▁幸 -11.5559
▁抖 -11.5559
▁排水 -11.5566
▁大队 -11.5567
▁纽 -11.5571
▁不适 -11.5574
▁主板 -11.5575
▁阶级 -11.5579
▁副局 -11.558
▁节日 -11.5582
51 -11.5582
机制 -11.5586
▁应届 -11.5587
梨 -11.5588
▁点赞 -11.5589
▁93 -11.5589
▁世上 -11.5589
▁领导人 -11.559
▁膜 -11.5591
▁10000 -11.5591
事件 -11.5593
▁忍受 -11.5595
▁1956 -11.5596
扎实 -11.5596
闲 -11.5596
▁死去 -11.5597
▁新增 -11.5598
▁复旦 -11.5601
生意 -11.5601
▁直观 -11.5603
▁抽象 -11.5604
▁侍 -11.5605
▁面部 -11.5606
▁人品 -11.5611
ps -11.5614
▁----- -11.5617
▁合肥 -11.5621
▁And -11.5627
家庭 -11.5628
▁菌 -11.563
▁這 -11.5631
▁换个 -11.5632
▁邮政 -11.5633
▁丧 -11.5633
▁冰箱 -11.5636
窍 -11.5643
SS -11.5644
▁唐代 -11.5647
▁细致 -11.5649
两天 -11.5649
▁原子 -11.565
▁桃 -11.565
▁菩萨 -11.565
▁皇家 -11.5655
▁74 -11.5656
▁发放 -11.5657
▁Be -11.5658
▁感恩 -11.566
▁概括 -11.5661
咏 -11.5663
▁320 -11.5667
▁不难 -11.5668
▁脑洞 -11.5671
ner -11.5671
▁被迫 -11.5674
吸 -11.5683
▁带头 -11.5683
▁炖 -11.5685
肤 -11.5686
疾病 -11.5689
60 -11.569
▁从中 -11.5693
▁两点 -11.5696
▁姿态 -11.5696
▁浙 -11.5697
▁6000 -11.5697
▁改名 -11.5697
雁 -11.57
笛 -11.5703
▁加载 -11.5703
▁康熙 -11.5704
av -11.5705
口气 -11.5706
陀 -11.5709
雍 -11.571
▁月经 -11.5712
插 -11.5713
紫 -11.5714
▁资本主义 -11.5719
悠 -11.5722
▁尽快 -11.5725
▁叙述 -11.5725
▁有益 -11.5726
▁观测 -11.5727
▁83 -11.5728
▁静脉 -11.5729
eng -11.5729
89 -11.573
▁兔 -11.5731
▁恶魔 -11.5731
▁沙发 -11.5732
▁语气 -11.5734
▁保卫 -11.5735
简 -11.5736
▁到来 -11.5739
仆 -11.5741
▁鉴别 -11.5742
大堆 -11.5742
▁后代 -11.5743
▁嫉妒 -11.5744
▁小儿 -11.5746
▁人去 -11.5749
▁启发 -11.5749
ん -11.575
违 -11.5751
侃 -11.5756
▁氨基 -11.5757
个月 -11.5758
▁Unit -11.5759
▁For -11.5761
▁橙 -11.5763
▁深厚 -11.5766
▁论证 -11.5766
▁平行 -11.5767
▁第十八 -11.5771
胡 -11.5772
▁通风 -11.5773
▁一双 -11.5774
格式 -11.5774
▁曹操 -11.5774
▁每一 -11.5777
雄厚 -11.578
喷 -11.5781
▁航天 -11.5787
▁崩 -11.5787
▁东南亚 -11.5792
汽车 -11.5797
▁5.5 -11.5797
▁屏蔽 -11.5798
▁电影院 -11.5801
∠ -11.5803
▁全力 -11.5804
▁打败 -11.5805
▁科比 -11.5805
▁淀粉 -11.5807
▁核桃 -11.581
▁江西省 -11.5812
陋 -11.5815
萍 -11.5816
▁放假 -11.582
▁小于 -11.5823
▁一中 -11.5826
ಥ -11.5836
▁热量 -11.5837
▁垫 -11.5837
ance -11.5839
▁信念 -11.5841
尿 -11.5842
▁揉 -11.5843
重要 -11.5848
屿 -11.5849
▁有害 -11.585
▁四十 -11.5851
▁作弊 -11.5851
▁115 -11.5855
缴 -11.5855
▁秩序 -11.5855
▁优美 -11.5861
胃 -11.5864
▁存款 -11.5864
▁仰 -11.5865
▁角落 -11.5865
me -11.5867
ay -11.5871
匪 -11.5871
▁某某 -11.5879
▁熬 -11.588
▁二十年 -11.588
▁保安 -11.5882
▁油画 -11.5882
▁矫情 -11.5883
▁熔 -11.5884
▁右手 -11.5884
▁乾隆 -11.5885
▁伯 -11.5887
▁リ -11.5891
個 -11.5893
5.0 -11.5899
▁剧场 -11.59
▁棕 -11.5903
▁柴 -11.5906
盼 -11.5907
纤维 -11.5908
di -11.5908
脊 -11.5909
效果 -11.5909
▁6.3 -11.5912
PA -11.5912
▁协作 -11.5915
▁想过 -11.5915
桩 -11.5918
▁内衣 -11.5919
▁票房 -11.5921
▁Win -11.5922
▁码头 -11.5924
▁建材 -11.5924
漂亮 -11.5925
▁长发 -11.5926
▁山西省 -11.5927
▁预警 -11.5927
▁开业 -11.5927
▁疲劳 -11.5928
▁0.3 -11.5928
▁邮箱 -11.5929
▁灰色 -11.593
▁学业 -11.593
▁可达 -11.593
------------- -11.5932
营销 -11.5933
▁补贴 -11.5933
▁征收 -11.5934
慎 -11.5936
眼皮 -11.5941
▁后卫 -11.5941
▁第二百 -11.5943
▁僧 -11.5944
▁基督 -11.5944
闷 -11.5944
▁阿拉伯 -11.5945
▁经销 -11.5951
敦 -11.5951
耍 -11.5953
媚 -11.5953
lu -11.5953
滨 -11.5953
▁通话 -11.5954
▁一丝 -11.5955
▁钥匙 -11.5958
▁舍不得 -11.596
▁多功能 -11.596
▁成型 -11.5967
煤 -11.5972
▁太平洋 -11.5974
▁ID -11.5974
▁诸如 -11.5975
▁变动 -11.5976
驾 -11.5976
▁脖子 -11.5976
丰富 -11.5977
▁名词 -11.5977
▁捏 -11.5978
▁模板 -11.5978
搏 -11.5981
80 -11.5982
▁十万 -11.5983
▁路边 -11.5984
▁毕 -11.5985
▁洗脑 -11.5986
DS -11.5986
ao -11.5986
▁执着 -11.5987
▁1976 -11.5989
▁美观 -11.599
▁运算 -11.5991
悔 -11.5992
▁忧 -11.5993
▁饮水 -11.5993
矩 -11.5993
▁2.6 -11.5994
技能 -11.5994
▁付费 -11.5995
▁果实 -11.5995
▁打死 -11.5996
不来 -11.6004
▁僵尸 -11.6005
▁垄断 -11.6005
▁尸体 -11.6005
pi -11.601
疹 -11.6014
泊 -11.6014
程度 -11.6017
▁法语 -11.6018
▁大地 -11.6022
丈 -11.6023
▁阴谋 -11.6026
疮 -11.6027
▁打架 -11.6027
▁人能 -11.6028
▁成交 -11.6033
▁引领 -11.6033
▁热血 -11.6034
▁防火 -11.6044
离职 -11.6047
▁弗 -11.6048
▁按键 -11.6049
▁支出 -11.6049
校学生 -11.6049
▁干脆 -11.605
▁反射 -11.6051
▁用人 -11.6052
▁隐私 -11.6052
▁精密 -11.6053
糕 -11.6053
▁回顾 -11.6054
▁利率 -11.6054
▁论述 -11.6056
▁利息 -11.6058
▁线性 -11.6059
は -11.6061
▁透过 -11.6062
阶级 -11.6062
枫 -11.6064
▁发热 -11.6064
oc -11.6067
▁伞 -11.607
▁不堪 -11.6071
▁抗战 -11.6071
▁插入 -11.6074
▁孕 -11.6076
笋 -11.6077
▁乘客 -11.6078
▁千年 -11.6081
▁邪 -11.6083
坦 -11.6085
▁头脑 -11.6086
▁既有 -11.6086
▁文革 -11.6087
地址 -11.6088
▁锦 -11.6089
奎 -11.6089
▁忠诚 -11.6091
600 -11.6092
▁博览 -11.6092
▁光盘 -11.6093
成分 -11.6093
56 -11.6094
▁89 -11.6094
▁珍珠 -11.6095
▁土耳其 -11.6095
▁百分 -11.6098
▁小伙 -11.6099
.08 -11.61
▁手表 -11.61
▁七十 -11.6103
▁辽宁省 -11.6103
▁活泼 -11.6106
▁相爱 -11.6106
媒体 -11.6106
逼 -11.6109
▁六个 -11.6109
bo -11.6116
▁拳 -11.612
bb -11.6121
▁总线 -11.6123
言之 -11.6125
EC -11.6127
▁真空 -11.6132
▁此事 -11.6134
▁一行 -11.6135
▁编译 -11.6135
▁租房 -11.6135
▁出名 -11.6135
▁皇后 -11.6137
▁碑 -11.6139
▁螺 -11.6142
學 -11.6142
▁谈判 -11.6143
▁市政 -11.6145
OS -11.6147
▁握 -11.6149
▁深夜 -11.6152
der -11.6152
▁结算 -11.6153
▁超出 -11.6155
蜡 -11.6158
▁太空 -11.6158
▁沪 -11.6159
▁星星 -11.616
▁永恒 -11.6162
▁大夫 -11.6164
▁978780 -11.6164
▁言语 -11.6164
▁深层 -11.6164
进取 -11.6166
▁自然科学 -11.6167
▁66 -11.6167
▁问答 -11.6177
▁0.8 -11.6178
▁过渡 -11.618
旋 -11.618
企业家 -11.6182
▁量子 -11.6183
▁八年 -11.6188
▁同龄 -11.6188
US -11.6188
▁南海 -11.619
▁风雨 -11.619
犯罪 -11.6191
▁61 -11.6191
▁中国科学院 -11.6194
▁扒 -11.6198
▁出差 -11.6199
箭 -11.62
▁虐 -11.62
▁太原 -11.6201
▁黎 -11.6204
▁台北 -11.6205
吐 -11.6207
▁变量 -11.6208
▁奠 -11.6209
▁做个 -11.621
------------ -11.6212
▁评分 -11.6213
▁味精 -11.6214
▁热带 -11.6216
▁小麦 -11.622
碧 -11.6223
▁签证 -11.6223
▁宝马 -11.6224
▁大腿 -11.6225
▁未能 -11.6225
▁粒 -11.6226
▁饲料 -11.6226
▁自愿 -11.6226
▁联合会 -11.6227
温度 -11.6228
▁宾 -11.6231
▁福州 -11.6233
▁长篇 -11.6235
▁iPad -11.6237
▁域名 -11.6238
過 -11.6239
罚 -11.624
牧 -11.6243
闭 -11.6243
▁宝贵 -11.6245
▁82 -11.6251
暗 -11.6254
▁三人 -11.6254
▁总裁 -11.6254
▁扬州 -11.6255
▁武侠 -11.6256
▁征服 -11.6257
▁权益 -11.6258
▁人为 -11.6258
▁1950 -11.626
▁忽悠 -11.6261
炼 -11.6263
▁出处 -11.6264
▁变换 -11.6265
▁港口 -11.6265
▁思念 -11.6272
≤ -11.6273
▁广阔 -11.6274
à -11.6274
▁悠 -11.6276
▁打个 -11.6277
▁劲 -11.628
▁介质 -11.6281
▁新颖 -11.6282
▁真题 -11.6282
▁发掘 -11.6283
▁枚 -11.6283
▁精度 -11.6288
▁名片 -11.629
▁好不 -11.6292
▁番茄 -11.6293
▁地板 -11.6293
▁褐 -11.6294
▁耗 -11.6295
LE -11.6297
▁储备 -11.6299
委员会 -11.6299
▁特长 -11.6301
志愿者 -11.6302
▁执业 -11.6302
▁调料 -11.6302
▁婆婆 -11.6302
经理 -11.6304
▁丙 -11.6304
▁臀 -11.6305
▁梗 -11.6306
概念 -11.6308
截 -11.6311
▁试用 -11.6314
▁Ro -11.6314
▁好用 -11.6318
▁漏洞 -11.6319
▁we -11.632
▁聊聊 -11.6321
▁蜂蜜 -11.6321
▁电台 -11.6321
▁仓 -11.6322
▁7.1 -11.6324
▁CS -11.6325
▁马路 -11.6325
儒 -11.6325
▁提倡 -11.6326
▁不信 -11.633
▁饲养 -11.6332
▁北美 -11.6333
▁渲染 -11.6333
▁数控 -11.6335
▁105 -11.6338
▁毗 -11.6339
虾 -11.6339
▁线条 -11.635
▁间接 -11.6356
▁一日 -11.6358
▁蝴蝶 -11.6358
▁内向 -11.6358
前景 -11.6359
涅 -11.6359
▁野生 -11.636
改革 -11.6364
▁受益 -11.6365
▁隧道 -11.6366
手术 -11.637
▁妇科 -11.637
▁字母 -11.637
00 -11.6371
▁得分 -11.6373
▁一书 -11.6373
▁凌 -11.6375
▁亲身经 -11.6376
▁佐 -11.6377
▁男神 -11.6378
▁影子 -11.638
▁衍生 -11.6381
旭 -11.6382
▁河流 -11.6387
▁行驶 -11.6387
▁颈 -11.6387
亨 -11.6388
▁无可 -11.6392
▁醒 -11.6394
▁收缩 -11.6396
AM -11.6396
▁麻醉 -11.6397
▁67 -11.6397
▁风水 -11.6397
▁风光 -11.6397
▁两家 -11.64
▁外贸 -11.6402
▁纠纷 -11.6412
▁用力 -11.6413
▁定时 -11.6415
▁捉 -11.6415
锐 -11.6417
▁债务 -11.6419
▁94 -11.6423
▁谨慎 -11.6423
虐 -11.6424
▁振动 -11.6425
▁卫视 -11.6425
▁锤子 -11.6427
▁剂量 -11.643
▁悟 -11.6431
▁府 -11.6432
▁产地 -11.6433
要求 -11.6434
头部 -11.6434
碟 -11.6435
FC -11.6435
▁古城 -11.6436
▁年级 -11.6442
▁司马 -11.6444
▁清醒 -11.6445
art -11.6446
申 -11.6452
▁NO -11.6452
疆 -11.6452
▁黑龙江 -11.6454
估 -11.6456
MI -11.6457
▁即时 -11.6458
mon -11.6458
阻 -11.646
▁天文 -11.6461
织 -11.6463
▁撤销 -11.6464
▁精通 -11.6464
▁分工 -11.6465
▁章节 -11.6467
▁很想 -11.647
▁纲 -11.6473
▁进修 -11.6474
▁荒 -11.6475
▁头上 -11.6478
▁六十 -11.6479
▁新华 -11.6481
▁前锋 -11.6483
龟 -11.6484
陆续 -11.6487
▁历代 -11.6488
▁参赛 -11.6489
▁炫耀 -11.649
▁同性 -11.6492
▁进出口 -11.6493
▁制片 -11.6494
▁雕塑 -11.6496
一看 -11.6499
PE -11.65
▁1970 -11.6501
▁要死 -11.6504
募 -11.6505
▁龙头 -11.6506
▁Mo -11.6507
优势 -11.6509
▁常委 -11.6509
屈 -11.6511
▁不大 -11.6511
实实 -11.6513
▁受理 -11.6515
▁姓氏 -11.6515
▁短暂 -11.6516
茅 -11.6517
mi -11.6519
gu -11.6519
霾 -11.6522
▁漫 -11.6522
骗 -11.6523
▁累积 -11.6525
献 -11.6528
▁偏向 -11.6529
ue -11.653
▁音效 -11.6533
▁蹭 -11.6534
籽 -11.6535
▁课本 -11.6536
▁北上 -11.6536
▁唇 -11.6537
▁诱 -11.6538
▁漫长 -11.654
突 -11.6543
▁奖金 -11.6543
酿 -11.6545
▁感悟 -11.6548
▁性感 -11.6552
▁9787 -11.6554
加工 -11.6556
炸 -11.6556
心酸 -11.6557
▁1600 -11.6557
▁犬 -11.6557
á -11.656
▁引擎 -11.656
各地 -11.6561
▁天涯 -11.6565
▁哀 -11.6565
▁VS -11.6566
▁党支 -11.6566
CP -11.657
▁实务 -11.6573
▁范畴 -11.6574
▁球场 -11.6574
59 -11.658
薯 -11.6583
崇 -11.6583
▁法人 -11.6583
り -11.6584
▁四季 -11.6585
▁天津市 -11.6586
▁食材 -11.6586
▁摊 -11.6587
▁艺 -11.6588
▁清真 -11.6589
hi -11.6589
▁制造业 -11.659
挡 -11.6595
▁探测 -11.6596
哉 -11.6598
▁天上 -11.66
▁更高 -11.6602
授 -11.6603
▁默 -11.6608
▁无效 -11.6609
▁are -11.6609
淑 -11.6612
▁嫩 -11.6614
▁缓慢 -11.6614
▁分会 -11.6614
っ -11.6615
▁合伙 -11.6617
▁生平 -11.662
年内 -11.6622
▁编曲 -11.6622
▁一分 -11.6624
▁字符 -11.6625
调控 -11.6625
▁怨 -11.6626
▁不成 -11.6628
废 -11.6629
稳 -11.6629
▁蠢 -11.6629
栓 -11.6632
▁损坏 -11.6632
▁液压 -11.6633
▁假装 -11.6634
▁OS -11.6634
▁平民 -11.6635
▁玛 -11.6635
▁剩余 -11.6637
▁综 -11.6638
联保 -11.6641
▁自拍 -11.6641
▁87 -11.6642
▁客厅 -11.6644
栈 -11.6644
ds -11.6646
▁制冷 -11.6647
湛 -11.6647
▁占用 -11.6651
sc -11.6652
卒 -11.6653
高中 -11.6656
晕 -11.6656
▁狗狗 -11.6657
竞争 -11.6658
叠 -11.666
RS -11.6661
▁老虎 -11.6663
肢 -11.6666
▁香水 -11.6668
▁面粉 -11.6668
▁马来西亚 -11.6669
▁China -11.667
▁机电 -11.667
▁边界 -11.667
▁罚款 -11.6671
▁第十九 -11.6675
au -11.6675
帆 -11.6676
▁地形 -11.6679
创作 -11.6681
▁体检 -11.6683
胀 -11.6686
▁屎 -11.6687
元素 -11.6687
逆 -11.6688
▁尚书 -11.669
▁全年 -11.669
▁迟 -11.6691
恭 -11.6691
▁才华 -11.6692
逃 -11.6693
▁窝 -11.6693
▁监狱 -11.6693
▁灵感 -11.6694
冠军 -11.6696
摊 -11.6698
▁1960 -11.6699
▁艰苦 -11.67
▁up -11.67
澄 -11.6701
拓 -11.6702
▁殷 -11.6707
杏 -11.671
▁样式 -11.6713
吾 -11.6715
國 -11.6717
ap -11.672
400 -11.672
夺 -11.6722
▁旅客 -11.6722
▁恶性 -11.6723
▁------ -11.6727
201 -11.6728
▁用地 -11.673
▁左手 -11.6732
仅 -11.6732
▁刑法 -11.6732
▁伦理 -11.6735
▁GHz -11.6736
▁碾 -11.6737
▁术语 -11.674
▁移植 -11.6742
半岛 -11.6742
作家 -11.6742
▁注入 -11.6744
▁削 -11.6745
▁推行 -11.6747
膝 -11.6748
▁安心 -11.6749
鉴于 -11.6754
皆 -11.6757
▁初始 -11.6761
ま -11.6764
▁两边 -11.677
讨 -11.6776
生产 -11.6776
▁吻 -11.6776
▁脑残 -11.6781
▁杀手 -11.6781
ob -11.6786
▁阴道 -11.6788
▁裁判 -11.679
▁许可 -11.679
▁串 -11.679
▁开门 -11.679
▁讲座 -11.6796
2000 -11.6796
900 -11.6796
▁螺旋 -11.6797
▁书面 -11.6798
能源 -11.68
▁烹饪 -11.6801
▁时空 -11.6802
▁美人 -11.6802
▁教科 -11.6802
▁女王 -11.6803
▁生殖 -11.6811
▁侵略 -11.6812
抚 -11.6815
漫 -11.6818
▁判 -11.6819
▁ch -11.682
凭 -11.682
▁匿 -11.682
▁橡胶 -11.6821
▁趣味 -11.6822
▁邪恶 -11.6825
用车 -11.6828
▁健 -11.6829
▁foot -11.683
▁天猫 -11.6831
▁终究 -11.6832
招呼 -11.6833
▁茎 -11.6835
▁瑜伽 -11.6836
▁帽子 -11.6837
芒 -11.6838
绒 -11.6839
▁涌 -11.6842
斥 -11.6842
▁行程 -11.6843
▁分裂 -11.6843
▁风情 -11.6845
▁1945 -11.6846
▁社会学 -11.6847
ea -11.6847
▁相差 -11.6849
▁宏 -11.6853
TO -11.6853
▁900 -11.6857
▁何时 -11.6857
▁不利 -11.6858
CA -11.6862
▁海报 -11.6862
▁巩固 -11.6863
大会 -11.6864
滋 -11.687
过去 -11.6871
▁民俗 -11.6873
▁几位 -11.6877
▁控股 -11.6879
谦 -11.688
▁出手 -11.6888
▁名家 -11.689
▁ア -11.689
▁跨越 -11.6895
▁DNA -11.6895
▁69 -11.6901
▁1.7 -11.6902
绩 -11.6904
▁91 -11.6905
▁本能 -11.6906
▁代表作 -11.6906
6.0 -11.6909
▁染色 -11.6909
▁总监 -11.6911
▁放置 -11.6911
od -11.6913
黄色 -11.6915
▁时机 -11.6917
▁报复 -11.6918
▁踏 -11.6919
▁基友 -11.692
遗产 -11.692
▁器官 -11.6921
▁瓜 -11.6921
來 -11.6923
able -11.6923
▁窗 -11.6923
▁书名 -11.693
▁卒 -11.6931
▁侮辱 -11.6934
功率 -11.6935
▁控制器 -11.6936
▁繁华 -11.6939
▁仲裁 -11.6944
▁议 -11.6947
勋 -11.6948
▁两岸 -11.6952
▁犹太 -11.6955
薇 -11.6955
氢 -11.6958
▁中期 -11.6958
▁来得 -11.6959
▁落地 -11.696
▁半小 -11.696
▁灾难 -11.6961
▁头像 -11.6962
RE -11.6962
▁相连 -11.6965
▁商贸 -11.6965
▁街头 -11.6965
飞机 -11.6966
▁gay -11.6966
▁这首 -11.6967
▁show -11.6969
▁竖 -11.6972
▁受众 -11.6972
▁陆军 -11.6973
▁惊人 -11.6973
EN -11.6975
▁失业 -11.6978
▁人际 -11.6979
▁外星 -11.6981
物质 -11.6981
▁感叹 -11.6981
▁政协 -11.6984
眠 -11.6985
▁TV -11.6986
如何 -11.6989
▁回想 -11.6992
▁节点 -11.6994
▁怒 -11.6997
▁十九 -11.6997
▁救援 -11.6998
▁顶端 -11.6999
▁饮用 -11.7004
▁歌唱 -11.7004
▁首选 -11.7005
ag -11.7007
▁中毒 -11.7008
篮 -11.7009
ick -11.7009
秦 -11.701
▁描绘 -11.701
白了 -11.7011
▁Mar -11.7012
▁尹 -11.7013
▁PCI -11.7014
▁丸 -11.7015
▁是非 -11.7015
▁君子 -11.7018
▁屯 -11.702
Cr -11.7025
旱 -11.7025
▁物资 -11.7025
▁押 -11.7031
▁星球 -11.7033
▁5.4 -11.7033
肃 -11.7034
▁双重 -11.7037
▁估 -11.7038
▁辅 -11.7039
▁这段 -11.7044
感兴趣 -11.7044
▁醋 -11.7047
识别 -11.7047
▁习 -11.7047
▁伤口 -11.7049
▁四五 -11.705
▁OK -11.7052
▁伴侣 -11.7056
▁一杯 -11.7056
▁1975 -11.7059
▁选购 -11.7061
取向 -11.7061
.01 -11.7068
▁难免 -11.7069
▁警告 -11.7073
▁关怀 -11.7075
▁无穷 -11.7075
ron -11.708
妖 -11.7081
▁初学 -11.7084
up -11.7084
危 -11.7085
▁呼 -11.7086
AR -11.7087
▁华夏 -11.7087
▁行走 -11.709
▁枯 -11.709
▁露出 -11.7091
▁早晨 -11.7094
▁葡萄酒 -11.7094
▁MV -11.7097
▁昆虫 -11.7098
▁看得 -11.7099
颖 -11.7099
职称 -11.71
▁沉淀 -11.7102
▁赠 -11.7103
▁罕 -11.7103
▁香味 -11.7103
▁this -11.7104
▁自尊 -11.7105
▁AV -11.7106
▁十月 -11.7107
▁绑架 -11.7107
难看 -11.7109
ou -11.711
霉 -11.7115
ら -11.7116
▁扶持 -11.7116
得起 -11.7117
▁一发 -11.7118
崎 -11.7118
▁率先 -11.7121
▁特质 -11.7123
杨 -11.7125
▁0.6 -11.7126
si -11.7127
▁蟹 -11.7127
▁微生物 -11.7128
权益 -11.7129
▁贝尔 -11.7129
▁百分之 -11.7129
▁掀 -11.7129
▁除此 -11.713
▁破解 -11.7131
不觉 -11.7133
▁冻 -11.7135
LA -11.7135
▁侧面 -11.7136
禹 -11.7136
天下 -11.7136
▁四周 -11.7137
▁诡 -11.7142
▁析 -11.7143
▁iphone -11.7143
贸 -11.7145
▁彼 -11.7145
▁My -11.7147
▁励志 -11.7148
▁屋 -11.7149
腺 -11.7153
驰 -11.7154
▁交际 -11.7159
▁拍卖 -11.7159
▁77 -11.716
▁桃花 -11.7161
▁下雨 -11.7165
゚ -11.7167
▁92 -11.7168
▁光滑 -11.7169
▁魔术 -11.7169
▁小吃 -11.7169
▁仿真 -11.717
▁all -11.7174
▁1964 -11.7175
sp -11.7176
▁无敌 -11.7176
▁德语 -11.7176
▁宽容 -11.7179
艇 -11.718
城区 -11.7182
▁101 -11.7187
▁危 -11.7188
▁妥协 -11.7188
▁對 -11.7192
▁城区 -11.7192
▁不妨 -11.7196
NA -11.7197
▁劈 -11.7201
▁GPS -11.7201
き -11.7201
▁无视 -11.7202
ec -11.7202
▁侵权 -11.7202
▁磷 -11.7202
▁比喻 -11.7202
▁bug -11.7203
▁相继 -11.7203
▁流域 -11.721
▁翘 -11.7211
参考 -11.7212
▁but -11.7214
▁保温 -11.7217
▁雕刻 -11.7221
▁变革 -11.7221
▁圣经 -11.7223
▁斑 -11.7223
▁聚合 -11.7224
▁龟 -11.7225
▁噪音 -11.7225
▁调试 -11.7226
▁找出 -11.7226
▁二者 -11.7226
▁he -11.7227
▁土豪 -11.7229
▁面板 -11.7229
而出 -11.7231
▁流水 -11.7232
▁一组 -11.7233
本事 -11.7234
▁命中 -11.7245
▁web -11.7245
▁续航 -11.7246
努力 -11.7247
▁大批 -11.7249
▁遇上 -11.7249
最 -11.7249
斋 -11.725
les -11.725
人心 -11.7252
▁火锅 -11.7253
▁转身 -11.7254
▁帐 -11.7256
▁离子 -11.7257
▁去除 -11.7257
穆 -11.7257
适配 -11.7257
▁舔 -11.7259
庸 -11.726
努 -11.7263
随 -11.7263
▁二十多 -11.7264
▁七月 -11.7264
▁进度 -11.7266
▁480 -11.7266
魄 -11.7267
耐 -11.7267
▁108 -11.727
茎 -11.7273
▁调理 -11.7274
学好 -11.7274
▁1962 -11.7275
▁万物 -11.7276
» -11.7277
▁并发 -11.7282
提要 -11.7282
▁招标 -11.7283
被 -11.7283
▁并不 -11.7284
选择 -11.7288
▁1977 -11.7289
点儿 -11.7289
集体 -11.7291
衫 -11.7292
▁用水 -11.7293
尧 -11.7294
▁轨迹 -11.7295
▁更何 -11.7296
观点 -11.7302
▁调味 -11.7302
▁着急 -11.7304
▁长春 -11.7309
▁柔软 -11.7312
▁一幅 -11.7312
▁修炼 -11.7313
▁自家 -11.7314
▁愚蠢 -11.7314
▁选拔 -11.7314
▁第二十 -11.7317
窟 -11.7317
MO -11.732
▁酒精 -11.7325
▁修理 -11.7326
▁口号 -11.7326
▁口服 -11.7328
▁所得 -11.733
祠 -11.733
▁鲍 -11.733
▁跟随 -11.7332
▁前女 -11.7333
污染 -11.7333
700 -11.7336
▁额定 -11.7339
▁有意 -11.7344
▁持久 -11.7344
▁晋升 -11.7346
▁我来 -11.7346
▁初三 -11.7346
▁屡 -11.7347
▁添 -11.7352
矛盾 -11.7354
▁增值 -11.7358
▁一二 -11.7359
▁中外 -11.736
▁嘴里 -11.7361
杖 -11.7361
▁排放 -11.7362
衔 -11.7363
▁6.5 -11.737
歌手 -11.7371
▁考古 -11.7372
▁ク -11.7374
▁阳台 -11.7377
▁身后 -11.7378
AL -11.7378
▁掩 -11.7391
▁入住 -11.7392
▁联网 -11.7394
▁出行 -11.7394
兮 -11.7395
▁玩玩 -11.7395
肋 -11.7395
开朗 -11.7397
▁1963 -11.7397
▁支援 -11.7397
▁修订 -11.7398
病毒 -11.7398
恰当 -11.7399
庚 -11.7402
▁感应 -11.7403
▁葡萄牙 -11.7404
▁不顾 -11.7404
▁殿 -11.7404
▁闯 -11.7407
禽 -11.741
▁炫 -11.7411
态度 -11.7414
▁97871 -11.7417
SA -11.7417
▁感触 -11.7422
ho -11.7425
▁硅 -11.7425
▁溶解 -11.7427
▁反击 -11.7427
▁解说 -11.7428
▁沙漠 -11.7429
▁供给 -11.7429
▁泻 -11.7429
▁竣 -11.7429
▁岩石 -11.743
忆 -11.7431
air -11.7431
▁碰撞 -11.7432
▁这事 -11.7433
虑 -11.7434
扯 -11.7434
帘 -11.7436
▁二十四 -11.7438
卜 -11.7441
▁促销 -11.7442
▁必定 -11.7442
▁这座 -11.7444
▁标记 -11.7444
▁义务教育 -11.7445
▁矿产 -11.7447
▁逼格 -11.7449
▁睡不 -11.7452
侨 -11.7453
▁卤 -11.7453
▁疫 -11.7456
62 -11.7457
▁play -11.7459
钙 -11.7459
▁安置 -11.746
弧 -11.7461
颠 -11.7463
▁金刚 -11.7464
▁填写 -11.7465
熬 -11.7467
肺 -11.7468
投资 -11.7472
▁上涨 -11.7474
▁治安 -11.7475
▁羽 -11.7477
▁跳舞 -11.7477
合作 -11.7478
▁三月 -11.7478
▁市面 -11.7479
01 -11.748
▁丝毫 -11.7481
ome -11.7481
▁难忘 -11.7481
▁年薪 -11.7487
▁自习 -11.7489
辱 -11.749
需求 -11.749
▁处置 -11.7492
▁平淡 -11.7492
▁珠宝 -11.7492
▁痕迹 -11.7493
▁华东 -11.7494
99 -11.7495
▁粗糙 -11.7496
▁邵 -11.7496
ph -11.7497
▁一处 -11.75
ler -11.7501
意见 -11.7502
▁切实 -11.7502
▁失落 -11.7503
▁做些 -11.7503
▁------- -11.7503
▁主任医 -11.7504
▁背光 -11.7504
く -11.7506
强度 -11.7508
▁链 -11.751
▁offer -11.7514
冀 -11.7514
吗 -11.7515
猎 -11.7515
▁波兰 -11.7516
课题 -11.7516
ㄨ -11.7517
朋 -11.7517
▁闽 -11.7517
▁Do -11.7519
▁任命 -11.7527
▁西瓜 -11.7528
hu -11.7529
Cl -11.7529
▁带上 -11.753
腕 -11.7531
▁抵押 -11.7532
▁73 -11.7533
ā -11.7534
▁新城 -11.7536
公开 -11.7537
88 -11.7537
ium -11.7541
▁荷 -11.7542
▁告别 -11.7545
▁一头 -11.7547
▁尊严 -11.7548
▁激活 -11.755
▁心动 -11.755
▁五大 -11.7552
▁口袋 -11.7553
藻 -11.7553
喘 -11.7553
▁84 -11.7554
▁哼 -11.7554
▁打扰 -11.7555
▁1965 -11.7555
▁研 -11.7557
▁建模 -11.7561
▁终极 -11.7562
▁前列腺 -11.7562
▁空白 -11.7562
▁凝聚 -11.7563
▁185 -11.7563
□ -11.7564
▁出租 -11.7565
▁我爱 -11.7565
▁实业 -11.7568
▁有次 -11.7568
▁po -11.7569
▁1800 -11.757
▁后勤 -11.757
螺 -11.757
IE -11.7571
挑 -11.7572
誓 -11.7574
▁考证 -11.7576
SI -11.7576
▁介入 -11.7578
埃 -11.7578
▁社科 -11.7579
▁眉 -11.7581
会计师 -11.7583
▁购 -11.7583
▁简化 -11.7585
▁见证 -11.7587
▁XP -11.7589
▁甘心 -11.7593
▁统筹 -11.7596
对手 -11.7598
秒 -11.7598
▁发给 -11.7603
CE -11.7603
▁责 -11.7604
▁自豪 -11.7609
▁2500 -11.7614
▁脑海 -11.7615
▁baike -11.7617
蓉 -11.7618
▁寿 -11.7618
生产力 -11.7618
▁诏 -11.7619
▁加密 -11.7619
滞 -11.7621
▁大片 -11.7623
▁雌 -11.7624
▁剧中 -11.7625
领域 -11.7625
▁热心 -11.7628
▁不合理 -11.7629
▁推销 -11.763
形式 -11.763
▁商量 -11.763
▁退役 -11.7633
▁百合 -11.7633
党委 -11.7636
▁防晒 -11.7636
寂 -11.7637
▁大桥 -11.7642
▁传达 -11.7644
▁太小 -11.7646
▁三千 -11.7646
▁总理 -11.7647
枣 -11.7647
▁笑容 -11.765
▁出入 -11.7651
▁71 -11.7651
▁大海 -11.7652
▁气势 -11.7655
▁仓库 -11.7656
▁室外 -11.7656
▁雷达 -11.7659
高度 -11.766
▁诺基亚 -11.7661
▁孤单 -11.7661
sch -11.7664
▁繁 -11.7665
▁留言 -11.7667
▁背叛 -11.7672
▁抵抗 -11.7672
▁市场营销 -11.7674
▁阵容 -11.7675
离子 -11.7675
▁力求 -11.7676
▁试点 -11.7681
▁千里 -11.7682
分明 -11.7683
log -11.7683
▁旋 -11.7683
乔 -11.7687
ka -11.7691
耻 -11.7692
晴 -11.7693
━ -11.7695
qu -11.7695
▁无锡 -11.7698
▁甄 -11.7699
▁光荣 -11.77
▁1955 -11.7702
▁燃料 -11.7702
▁冠 -11.7703
▁柱 -11.7703
▁管理局 -11.7711
▁共鸣 -11.7712
▁妹纸 -11.7712
▁扁 -11.7715
▁175 -11.7716
▁左边 -11.7717
▁耶 -11.7721
▁炎症 -11.7722
借 -11.7726
弘 -11.7726
鼻 -11.7726
▁省委 -11.7726
▁激素 -11.7727
▁An -11.7727
仿 -11.773
▁大幅 -11.7732
▁博弈 -11.7733
禅 -11.7733
▁出品 -11.7734
▁愚 -11.7735
tri -11.7738
▁Di -11.7738
▁侠 -11.7742
却 -11.7745
▁肠 -11.7745
开始 -11.7745
宠 -11.7748
晒 -11.7751
▁金庸 -11.7751
▁扬声 -11.7752
▁抬头 -11.7752
が -11.7754
▁堪称 -11.7755
▁奢 -11.7755
▁废话 -11.7756
▁锦标 -11.7757
▁get -11.7758
客场 -11.7758
.06 -11.7761
▁0.4 -11.7766
▁炼 -11.777
DA -11.7771
▁81 -11.7772
规范 -11.7773
臭 -11.7773
卵 -11.7777
▁脱发 -11.7778
▁扩 -11.778
▁网址 -11.7781
▁125 -11.7783
▁伸 -11.7785
▁中医药 -11.7786
▁救助 -11.7786
逛 -11.779
▁监理 -11.7792
▁电子书 -11.7793
聪 -11.7794
ui -11.7794
蛙 -11.7794
▁后人 -11.7795
line -11.7796
▁植 -11.7796
卢 -11.7797
谣 -11.7798
▁防范 -11.78
▁嘿嘿 -11.7802
▁科学研究 -11.7803
▁全套 -11.7804
▁谋 -11.7804
▁故乡 -11.7806
▁委 -11.7806
▁省市 -11.7806
拟 -11.7808
▁联络 -11.7808
窑 -11.7814
▁示 -11.7815
▁菲律宾 -11.7815
抽 -11.7815
嫂 -11.7817
▁右边 -11.7819
▁健全 -11.7822
▁书店 -11.7824
▁乾 -11.7826
ts -11.7827
▁越大 -11.7828
▁乞 -11.7829
▁坏人 -11.7829
▁标识 -11.783
▁村里 -11.7831
▁恰恰 -11.7832
▁扛 -11.7832
▁虫 -11.7833
医科大学 -11.7836
▁111 -11.7838
樱 -11.7839
▁装甲 -11.7839
▁拱 -11.7839
▁险 -11.7841
ey -11.7841
景观 -11.7843
▁折叠 -11.7843
卦 -11.7845
▁社会保障 -11.7845
▁熏 -11.7847
麟 -11.785
▁Microsoft -11.7851
▁运转 -11.7852
▁哎呀 -11.7853
ver -11.7854
▁奶粉 -11.7855
▁火山 -11.7858
▁单机 -11.786
▁贱 -11.7861
▁那句 -11.7862
▁晶体 -11.7863
▁空军 -11.7864
删 -11.7866
食品 -11.7867
▁一节 -11.7872
▁道教 -11.7873
▁修行 -11.7875
▁看不出 -11.7876
▁长约 -11.7877
▁遥控 -11.7878
▁造就 -11.788
▁两只 -11.788
并 -11.7881
▁赖 -11.7884
▁启示 -11.7886
利益 -11.7886
▁查找 -11.7887
▁自制 -11.7889
▁膨胀 -11.789
▁恰好 -11.7897
▁两张 -11.7897
▁编著 -11.7898
▁8.2 -11.7899
▁想来 -11.79
兼 -11.7903
▁铅 -11.7904
▁主打 -11.7906
▁震 -11.7908
▁和尚 -11.7911
▁未成年 -11.7916
▁看病 -11.7917
xi -11.7919
▁复活 -11.792
▁历经 -11.792
▁纳入 -11.7922
▁致命 -11.7923
▁回合 -11.7924
▁牛仔 -11.7925
▁决议 -11.7925
▁应付 -11.7925
▁诊疗 -11.7926
◇ -11.7928
▁瓷 -11.7928
蟹 -11.7929
▁专职 -11.7933
▁富裕 -11.7933
▁display -11.7936
捧 -11.7937
▁湿度 -11.7938
▁液晶 -11.794
▁弯曲 -11.794
▁底下 -11.7941
▁1959 -11.7944
摇 -11.7946
▁消极 -11.7947
▁木材 -11.7948
▁凝 -11.7948
粥 -11.7955
一致 -11.7955
▁小镇 -11.7958
▁演艺 -11.7958
▁协同 -11.7959
▁光电 -11.7961
际关系 -11.7962
▁It -11.7962
▁全班 -11.7963
▁平原 -11.7963
▁最有 -11.7963
▁托福 -11.7963
纬 -11.7966
▁优异 -11.7971
▁动脉 -11.7972
▁畸形 -11.7978
▁阿根廷 -11.7982
▁冶金 -11.7983
▁一说 -11.7985
▁昭 -11.7985
▁缩小 -11.7987
惧 -11.7989
▁外来 -11.7989
▁应试 -11.7992
▁高山 -11.7993
合物 -11.7993
浴 -11.7993
EX -11.7994
▁Ar -11.7995
▁精细 -11.7996
▁打字 -11.7999
90 -11.7999
您 -11.8
▁墨西哥 -11.8
▁文人 -11.8
▁诠释 -11.8003
▁最深 -11.8004
联系 -11.8005
▁威力 -11.8005
厕 -11.8005
▁楼梯 -11.8007
▁卓 -11.801
▁int -11.8011
鳍 -11.8012
▁中年 -11.8013
▁逝世 -11.8016
▁扭曲 -11.8019
▁说不出 -11.802
▁穴 -11.8025
▁候选 -11.8025
▁一个多 -11.8027
▁伊斯兰教 -11.8027
栗 -11.8028
▁1957 -11.8029
涌 -11.803
▁吉林省 -11.8032
珊 -11.8033
▁感人 -11.8036
▁育 -11.8037
▁白云 -11.8037
▁风味 -11.8041
粗 -11.8043
ga -11.8045
相同 -11.8046
▁家园 -11.8049
▁甜蜜 -11.8049
.03 -11.8053
▁捐 -11.8054
▁序列 -11.8054
▁资格证 -11.8055
蕊 -11.8057
▁器材 -11.8057
土地 -11.8059
那 -11.8062
▁0.7 -11.8062
▁1972 -11.8063
▁玩儿 -11.8068
▁阻碍 -11.8069
▁Apple -11.807
▁256 -11.8071
▁文物保护 -11.8072
剧院 -11.8074
▁力学 -11.8074
▁岳 -11.8075
▁脉 -11.8077
▁敷 -11.8078
挺 -11.8078
方形 -11.8079
▁国语 -11.808
以后 -11.808
.04 -11.8083
▁万能 -11.8083
文章 -11.8084
▁教导 -11.8085
多彩 -11.8086
▁align -11.8086
▁中华民族 -11.8087
检查 -11.8092
厢 -11.8092
▁矫正 -11.8092
▁连载 -11.8094
▁热闹 -11.8096
▁宿 -11.8096
▁可信 -11.8097
▁大楼 -11.8098
▁萝卜 -11.81
▁胶版 -11.8101
▁谈话 -11.8102
▁亲切 -11.8103
▁感性 -11.8108
▁发售 -11.8108
灌 -11.8108
络 -11.8109
完善 -11.8109
▁行使 -11.811
を -11.811
▁以人为 -11.8112
tic -11.8113
▁7.3 -11.8114
▁人来 -11.8118
▁饱 -11.8119
▁加重 -11.8119
筑 -11.812
▁合金 -11.8122
▁景色 -11.8123
英语 -11.8123
red -11.8124
多年 -11.8129
▁350 -11.8129
ies -11.813
▁邱 -11.8135
谨 -11.814
▁行情 -11.814
使用 -11.814
▁Si -11.814
▁帅哥 -11.8143
▁原型 -11.8143
▁写下 -11.8145
▁宠 -11.8146
▁拦 -11.8147
▁1952 -11.8148
▁焦点 -11.8148
▁8.1 -11.8155
▁投降 -11.8156
孩 -11.8157
骚 -11.816
伞 -11.8162
▁诊 -11.8164
▁嗨 -11.8165
▁校友 -11.8165
▁法官 -11.8166
▁下课 -11.8166
▁签名 -11.8169
▁表彰 -11.817
▁附上 -11.8173
▁不理 -11.8173
ci -11.8174
卧 -11.8174
状态 -11.818
▁url -11.8184
▁子弹 -11.8184
▁离不开 -11.8184
部队 -11.8186
▁font -11.8186
汀 -11.8186
爪 -11.8187
▁病理 -11.8187
▁音响 -11.8189
▁中途 -11.8189
▁保湿 -11.819
绿色 -11.819
皱 -11.8191
▁1974 -11.8195
▁预告 -11.8195
沛 -11.8198
▁发光 -11.8199
▁莆田 -11.8199
贸易 -11.8199
▁网球 -11.82
▁稳产 -11.8202
同时 -11.8203
荐 -11.8204
▁google -11.8206
▁贯穿 -11.8208
▁往事 -11.8208
▁有权 -11.8209
▁区内 -11.821
▪ -11.821
▁吞 -11.821
疏 -11.8212
▁期末 -11.8216
▁浑身 -11.8218
▁山地 -11.822
▁芝麻 -11.8221
梗 -11.8223
OL -11.8226
展会 -11.8228
70 -11.8228
▁成品 -11.8229
▁石家庄 -11.8229
▁if -11.8229
▁121 -11.823
78 -11.8232
▁残忍 -11.8232
▁农场 -11.8233
▁孙子 -11.8234
DI -11.8235
TS -11.8238
▁排出 -11.8241
▁亚马逊 -11.8242
▁间隔 -11.8243
▁老鼠 -11.8246
▁总的来 -11.8246
▁库存 -11.825
▁不远 -11.825
▁迪 -11.8251
不然 -11.8254
步步 -11.8255
份额 -11.8257
▁宁夏 -11.826
焕 -11.8262
▁膝盖 -11.8263
楠 -11.8267
▁蚊 -11.8268
单单 -11.827
550 -11.827
ub -11.8272
▁炒作 -11.8272
pp -11.8272
▁抢救 -11.8273
▁破产 -11.8274
▁评判 -11.8275
KG -11.8276
▁同理 -11.8277
▁花卉 -11.8278
襄 -11.8282
land -11.8284
▁离家 -11.8287
▁低下 -11.8288
带走 -11.8288
癖 -11.8289
葵 -11.8289
▁完后 -11.8292
▁1.0 -11.8294
寡 -11.8295
SO -11.8297
斧 -11.8297
▁哈利 -11.8299
▁高尚 -11.83
▁挥 -11.8302
忽 -11.8304
▁名誉 -11.8304
抖 -11.8305
▁五六 -11.8305
三角 -11.8307
▁调用 -11.8308
▁阿里巴巴 -11.8309
▁起义 -11.8311
4.8 -11.8311
蚕 -11.8312
▁1973 -11.8312
▁力气 -11.8314
叨 -11.8314
趋 -11.8315
let -11.8317
▁1968 -11.8317
屎 -11.8318
▁锡 -11.8319
≥ -11.832
▁没钱 -11.8321
▁at -11.8323
ny -11.8324
返 -11.8325
嗣 -11.8326
▁联邦 -11.8328
斯基 -11.8328
▁协商 -11.8329
偿 -11.8329
▁山上 -11.8329
▁紧紧 -11.833
学者 -11.833
▁十个 -11.8332
▁恶劣 -11.8333
▁数百 -11.8333
▁一瓶 -11.8333
▁锂电 -11.8337
政法 -11.8339
▁胸部 -11.834
do -11.834
交易 -11.8344
庵 -11.8345
▁腐 -11.8345
▁协会会 -11.835
IA -11.8351
弓 -11.8352
▁传销 -11.8357
▁750 -11.8358
▁搅 -11.8358
▁柠檬 -11.836
前方 -11.8361
▁终点 -11.8366
▁联系方 -11.8368
新闻 -11.8368
拘 -11.837
▁粤 -11.8371
▁群里 -11.8373
▁切除 -11.8373
tion -11.8374
▁不当 -11.8374
▁夫妇 -11.8374
简单 -11.8376
▁亏损 -11.8376
堵 -11.8378
▁军团 -11.8379
▁递 -11.838
▁吼 -11.838
▁预约 -11.838
▁珠 -11.8381
邮 -11.8381
▁亲爱 -11.8382
▁家教 -11.8383
▁一波 -11.8383
日报 -11.8385
▁十余 -11.8385
▁79 -11.8387
▁愉悦 -11.8388
调查 -11.8388
▁桂林 -11.8388
▁1961 -11.8389
ッ -11.8391
▁大国 -11.8392
▁暨 -11.8394
▁酶 -11.8396
▁短短 -11.8396
噢 -11.8396
▁DC -11.8397
▁发了 -11.8397
▁焊 -11.8397
ost -11.8399
du -11.8399
纺 -11.8401
冤 -11.8402
▁黑龙江省 -11.8402
职业 -11.8402
▁搜集 -11.8402
▁windows -11.8402
▁枕 -11.8403
▁八月 -11.8403
650 -11.8405
向上 -11.8407
▁鸿 -11.8408
380 -11.8409
▁世俗 -11.8409
▁长远 -11.8409
▁老是 -11.841
▁线上 -11.8416
复杂 -11.8417
▁have -11.842
▁捷 -11.8423
▁浴 -11.8424
▁插件 -11.8426
▁一站 -11.8426
卓 -11.8426
999 -11.8427
ED -11.8429
酮 -11.8429
55 -11.8429
▁卧室 -11.843
▁享 -11.8431
▁谎 -11.8432
▁国立 -11.8437
▁燃气 -11.844
▁梵 -11.8441
愧 -11.8441
▁早年 -11.8442
甚 -11.8445
▁凭证 -11.8447
▁屠 -11.8448
医疗 -11.8448
▁胡椒 -11.8449
▁这款 -11.845
▁禁忌 -11.8452
▁8000 -11.8452
▁坎 -11.8454
▁谈论 -11.8455
▁观音 -11.8456
妥 -11.8458
奸 -11.8459
MC -11.8459
▁同名 -11.8459
帐 -11.846
▁唐朝 -11.846
私 -11.8461
▁少儿 -11.8462
yl -11.8462
ca -11.8462
▁慈 -11.8463
▁地带 -11.8463
▁III -11.8464
表现 -11.8468
▁颁奖 -11.847
翰 -11.8473
▁1966 -11.8473
▁欧盟 -11.8474
▁6.4 -11.8475
艘 -11.8477
で -11.8481
▁was -11.8481
▁元件 -11.8482
▁美剧 -11.8483
悍 -11.8483
▁字幕 -11.8484
▁星际 -11.8486
▁风暴 -11.8487
阱 -11.8487
▁感知 -11.8491
▁妊娠 -11.8491
▁路口 -11.8493
猴 -11.8494
▁圣母 -11.8496
▁一秒 -11.8497
▁很多很多 -11.8498
▁邹 -11.85
▁稻 -11.8501
▁九十 -11.8503
▁昔 -11.8504
▁合唱 -11.8504
效应 -11.8505
▁Part -11.8505
▁窗户 -11.8506
痴 -11.8508
谐 -11.8509
▁备份 -11.8509
▁一碗 -11.8509
寇 -11.8511
▁有空 -11.8511
▁停下 -11.8512
▁隆 -11.8513
▁青海 -11.8514
▁好莱 -11.8514
CO -11.8516
▁不合 -11.8517
▁不服 -11.8518
04 -11.8518
贱 -11.8519
▁精神病 -11.8519
▁召 -11.8521
▁除去 -11.8521
又 -11.8522
▁耽误 -11.8523
▁双眼 -11.8523
▁沒 -11.8524
OR -11.8524
艾 -11.8529
▁有无 -11.853
he -11.8532
▁涂料 -11.8533
▁自助 -11.8533
▁而后 -11.8533
尼亚 -11.8534
▁家务 -11.8537
▁发烧 -11.8537
▁旺 -11.8538
▁录像 -11.8542
▁操场 -11.8542
2.5 -11.8543
▁海水 -11.8543
▁环球 -11.8545
▁人有 -11.8545
▁颠覆 -11.8546
▁微波 -11.8547
指标 -11.8548
▁密集 -11.8549
▁尽力 -11.855
割 -11.8552
▁电阻 -11.8553
▁168 -11.8557
开展 -11.8558
▁捐赠 -11.856
▁卡尔 -11.8562
▁相见 -11.8563
▁四月 -11.8564
▁字数 -11.8567
▁民事 -11.857
▁John -11.857
▁商业银行 -11.8572
▁女权 -11.8575
▁六月 -11.8576
▁担当 -11.8576
▁低温 -11.8577
ei -11.8577
▁漆 -11.8578
▁生化 -11.8578
让人 -11.8578
钞 -11.858
严重 -11.8581
可否 -11.8581
交往 -11.8582
抵 -11.8584
▁二十一 -11.8588
▁教育局 -11.8588
▁第二十一 -11.8588
▁火焰 -11.8593
究 -11.8594
cl -11.8599
▁儒家 -11.8599
工学 -11.8601
▁维权 -11.8602
▁习俗 -11.8603
滚 -11.8603
▁信赖 -11.8603
▁三点 -11.8605
证书 -11.8606
▁庆祝 -11.8607
▁履历 -11.8607
妞 -11.861
烯 -11.8612
▁不屑 -11.8613
絮 -11.8617
ding -11.8617
报道 -11.8617
荫 -11.8619
▁琼 -11.8621
▁华北 -11.8623
atch -11.8627
▁变速 -11.8628
ool -11.8629
▁读过 -11.8629
▁点点 -11.8629
▁java -11.863
tor -11.8632
▁风扇 -11.8632
▁New -11.8632
▁文书 -11.8632
▁谓 -11.8633
▁改正 -11.8635
▁切割 -11.8637
▁林地 -11.8639
▁自古 -11.8639
▁日内 -11.8639
▁拆迁 -11.8639
一体 -11.864
畅 -11.8642
▁中东 -11.8643
▁无线电 -11.8647
商标 -11.8649
▁男士 -11.8651
MM -11.8652
▁压迫 -11.8653
脾气 -11.8653
▁哺乳 -11.8656
挠 -11.8656
▁诈骗 -11.8657
▁图标 -11.866
▁曼 -11.866
单词 -11.8662
▁可行性 -11.8663
▁譬如 -11.8663
▁媒介 -11.8665
▁供水 -11.8667
▁实物 -11.8668
▁3.6 -11.8669
▁不负 -11.8669
52 -11.8672
丙 -11.8672
▁浓厚 -11.8673
▁SATA -11.8673
揽 -11.8673
编制 -11.8673
▁催化 -11.8675
书法家 -11.8678
所用 -11.868
▁will -11.868
▁酥 -11.8681
▁一切都 -11.8685
▁头痛 -11.8691
▁深化 -11.8694
▁仪表 -11.8695
も -11.8695
攻击 -11.8695
▁九月 -11.8697
sk -11.8698
▁脾 -11.8698
▁詹 -11.8699
标注 -11.8699
▁另有 -11.87
▁师兄 -11.87
れ -11.8702
麻将 -11.8703
▁悄悄 -11.8706
▁__ -11.8706
▁制品 -11.8707
翅 -11.8708
广场 -11.8709
▁车间 -11.8711
▁稿 -11.8712
▁纠错 -11.8713
▁违背 -11.8715
▁廉价 -11.8716
▁娃娃 -11.8716
▁纤 -11.8716
姨妈 -11.8717
▁出土 -11.872
女生 -11.8723
▁三百 -11.8724
红色 -11.8727
矢 -11.8729
▁中西 -11.8732
飘 -11.8734
回族 -11.8734
磊 -11.8737
こ -11.8738
▁羽毛球 -11.8739
崩 -11.8741
▁大体 -11.8741
▁di -11.8743
▁电缆 -11.8744
▁残疾人 -11.8746
▁执政 -11.8747
▁112 -11.8747
ì -11.875
东西 -11.875
SH -11.875
▁不看 -11.875
▁解放军 -11.8751
▁屈 -11.8753
▁专业课 -11.8753
4.7 -11.8755
不足 -11.8755
▁流氓 -11.8757
shi -11.8758
▁GDP -11.876
▁one -11.8768
凳 -11.877
▁渔业 -11.8771
▁军训 -11.8772
▁炒股 -11.8772
28 -11.8775
▁农历 -11.8775
▁糊 -11.8777
一半 -11.8782
▁细分 -11.8783
▁归纳 -11.8784
▁机灵 -11.8785
决心 -11.8786
冻 -11.8787
▁新西兰 -11.8788
▁扩散 -11.8789
▁篇文 -11.8789
诵 -11.8791
▁天真 -11.8795
SC -11.8797
ID -11.8798
▁方程 -11.8799
fa -11.8801
▁商会 -11.8802
▁歌舞 -11.8802
▁工业园 -11.8805
武帝 -11.8808
楼梦 -11.8808
TI -11.8809
▁1954 -11.881
▁滋味 -11.8812
▁CEO -11.8813
▁因果 -11.8813
.05 -11.8813
措 -11.8816
▁桂 -11.8816
▁玛丽 -11.8818
▁文笔 -11.8819
▁根源 -11.882
▁重组 -11.8822
▁无形 -11.8823
BO -11.8823
判定 -11.8824
ns -11.8825
性化 -11.8826
铁路 -11.8827
▁1953 -11.8827
▁攒 -11.8828
02 -11.8833
ON -11.8833
▁复发 -11.8834
OD -11.8834
▁粒子 -11.8835
▁1969 -11.8835
宝强 -11.8836
EL -11.8837
▁更改 -11.8838
▁结晶 -11.8842
将军 -11.8844
▁归属 -11.8847
▁债券 -11.8848
Д -11.8848
棱 -11.8848
▁弓 -11.885
▁抚 -11.885
▁老头 -11.8851
▁right -11.8851
▁美军 -11.8853
▁城堡 -11.8854
▁102 -11.8857
▁梨 -11.8858
▁地名 -11.8858
▁新品 -11.8858
▁农药 -11.8862
▁第二十二 -11.8865
▁敢于 -11.8872
▁装有 -11.8873
▁束缚 -11.8873
▁178 -11.8873
▁评委 -11.8878
▁永不 -11.888
▁size -11.8883
▁广播电视 -11.8884
李 -11.8884
▁栖息 -11.8887
▁受过 -11.8888
CI -11.8888
风险 -11.8892
▁争夺 -11.8893
▁全校 -11.8894
▁心得 -11.8894
▁牧 -11.8895
▁随手 -11.8895
▁伊斯兰 -11.8896
落实 -11.8897
剔 -11.8897
▁埃及 -11.8905
▁寒假 -11.8906
▁信托 -11.8908
不大 -11.8909
变化 -11.8909
▁奏 -11.891
▁盒子 -11.8916
普查 -11.8917
▁闻名 -11.8919
西亚 -11.892
▁挫折 -11.8921
影院 -11.8923
▁物价 -11.8924
▁会展 -11.8924
▁水电 -11.8925
磅 -11.8928
▁门诊 -11.893
▁著作权 -11.893
▁赛车 -11.8931
▁怪兽 -11.8932
更 -11.8933
▁模样 -11.8934
amp -11.8935
涡 -11.894
▁被子 -11.8942
▁不带 -11.8944
▁另一方 -11.8944
▁苯 -11.8945
▁静静 -11.8946
▁GB -11.8946
▁限定 -11.8947
▁境外 -11.8948
▁兮 -11.8948
摆 -11.895
OP -11.8951
号码 -11.8952
此刻 -11.8953
▁炸弹 -11.8955
▁佛山 -11.8955
▁横向 -11.896
▁羊肉 -11.8961
▁开阔 -11.8965
▁藤 -11.8965
凝 -11.8967
▁诸位 -11.8967
▁一号 -11.8968
▁莲花 -11.8969
▁狠 -11.8969
▁雕 -11.8975
▁容纳 -11.8976
▁Lo -11.8979
▁徐州 -11.8982
cc -11.8982
tu -11.8987
▁廖 -11.8987
幼 -11.8989
▁奥地利 -11.8995
▁啃 -11.8996
贡 -11.8998
不小 -11.9005
▁2.8 -11.9007
削 -11.9008
▁污水 -11.9008
▁前沿 -11.9009
ins -11.901
▁怎 -11.9011
▁盲 -11.9014
▁We -11.9014
▁试着 -11.9015
▁观光 -11.9019
▁吉祥 -11.902
禾 -11.902
▁疾 -11.9021
▁全场 -11.9023
▁终结 -11.9024
▁薇 -11.9025
▁机箱 -11.9026
▁br -11.9026
▁单反 -11.9027
ber -11.9028
▁修养 -11.903
▁明清 -11.9032
OT -11.9034
▁到期 -11.9036
▁户通 -11.9037
▁烘 -11.9038
▁脚步 -11.904
▁报价 -11.9047
基地 -11.9049
▁审定 -11.9049
dy -11.9051
▁肿 -11.9056
▁三线 -11.9057
▁先天 -11.9058
规模 -11.9058
蓬 -11.9059
甫 -11.9061
大批 -11.9062
▁梦幻 -11.9062
闺 -11.9065
黛 -11.9065
▁之战 -11.9066
▁史上 -11.9067
渴 -11.9069
垂 -11.9071
▁腐蚀 -11.9071
▁外文 -11.9075
▁女票 -11.9076
国道 -11.9077
▁135 -11.908
▁润滑 -11.9081
▁脚本 -11.9082
▁九年 -11.9082
揍 -11.9085
▁190 -11.9086
▁1937 -11.9089
▁棉花 -11.9089
▁165 -11.9091
▁推崇 -11.9092
班子 -11.9093
▁like -11.9093
▁压制 -11.9094
rt -11.9095
MP -11.9099
▁VGA -11.9099
▁家电 -11.9099
han -11.9101
▁慌 -11.9101
MT -11.9102
▁云狗 -11.9104
▁话剧 -11.9104
▁Car -11.9105
芙 -11.9107
▁无情 -11.9114
▁日报 -11.9114
会儿 -11.9114
男孩 -11.9114
▁阅 -11.9114
▁遍布 -11.9115
▁如图 -11.9115
▁go -11.9118
▁红军 -11.912
▁高分 -11.9121
便宜 -11.9122
▁超人 -11.9124
▁尊敬 -11.9126
▁纵横 -11.9127
あ -11.9128
▁彩票 -11.913
一些 -11.9131
▁香蕉 -11.9132
▁借款 -11.9132
▁贼 -11.9133
▁满分 -11.9135
煮 -11.9136
▁一贯 -11.9137
ran -11.9138
bu -11.9138
拥 -11.9139
舒 -11.9144
▁诺贝尔 -11.9144
▁操纵 -11.9145
▁椅 -11.9145
▁听课 -11.9146
▁中央电视台 -11.9148
抄 -11.9149
Pa -11.915
▁卡片 -11.915
▁两面 -11.9154
▁TA -11.9157
▁蜘蛛 -11.9159
搞 -11.9159
▁动画片 -11.916
▁中日 -11.916
▁梦里 -11.916
▁德育 -11.9161
▁挽 -11.9164
原理 -11.9164
押 -11.9169
▁104 -11.9169
▁紫外 -11.9169
▁纯属 -11.9169
▁会同 -11.9169
▁尝 -11.917
▁蒸汽 -11.917
▁东风 -11.9173
▁莲 -11.9173
▁恐龙 -11.9179
▁评定 -11.918
▁排序 -11.9181
运输 -11.9182
▁名声 -11.9182
▁不安 -11.9183
拙 -11.9184
▁1948 -11.9186
▁民生 -11.9187
▁梅花 -11.9189
▁蒲 -11.9189
▁宝石 -11.919
▁广电 -11.9191
尾巴 -11.9191
▁礼品 -11.9193
▁7.5 -11.9196
寻 -11.9196
mer -11.9197
迟 -11.9198
平衡 -11.9201
▁蘑菇 -11.9201
gitextract_o5tgzk64/ ├── 3rd/ │ └── gdown.pl/ │ ├── LICENSE.txt │ ├── README.md │ └── gdown.pl ├── CPM-Generate/ │ ├── .gitignore │ ├── LICENSE │ ├── README.md │ ├── arguments.py │ ├── bpe_3w_new/ │ │ ├── chinese_vocab.model │ │ ├── chinese_vocab.vocab │ │ ├── merges.txt │ │ └── vocab.json │ ├── configure_data.py │ ├── data/ │ │ ├── Makefile │ │ ├── __init__.py │ │ ├── bert_dataset.py │ │ ├── dataset_utils.py │ │ ├── gpt2_dataset.py │ │ ├── helpers.cpp │ │ ├── indexed_dataset.py │ │ ├── samplers.py │ │ └── test/ │ │ ├── test_indexed_dataset.py │ │ └── test_preprocess_data.sh │ ├── data_utils/ │ │ ├── __init__.py │ │ ├── corpora.py │ │ ├── datasets.py │ │ ├── file_utils.py │ │ ├── lazy_loader.py │ │ ├── samplers.py │ │ ├── tf_dl.py │ │ ├── tokenization.py │ │ ├── tokenization_gpt2.py │ │ └── wordpiece.py │ ├── fp16/ │ │ ├── __init__.py │ │ ├── fp16.py │ │ ├── fp16util.py │ │ └── loss_scaler.py │ ├── generate_samples.py │ ├── model/ │ │ ├── __init__.py │ │ ├── distributed.py │ │ ├── gpt2_modeling.py │ │ ├── model.py │ │ └── modeling.py │ ├── mpu/ │ │ ├── __init__.py │ │ ├── cross_entropy.py │ │ ├── data.py │ │ ├── grads.py │ │ ├── initialize.py │ │ ├── layers.py │ │ ├── mappings.py │ │ ├── random.py │ │ ├── tests/ │ │ │ ├── __init__.py │ │ │ ├── commons.py │ │ │ ├── test_cross_entropy.py │ │ │ ├── test_data.py │ │ │ ├── test_initialize.py │ │ │ ├── test_layers.py │ │ │ └── test_random.py │ │ ├── transformer.py │ │ └── utils.py │ ├── requirements.txt │ ├── scripts/ │ │ └── generate_text.sh │ └── utils.py ├── ChineseAiDungeonColabDemo.ipynb ├── LICENSE ├── finetune.ipynb ├── labeled_data/ │ ├── advanture_translated/ │ │ ├── process_data.ipynb │ │ ├── processed_translated_story.txt │ │ ├── processed_translated_story_valid.txt │ │ ├── text_advanture_trans.txt │ │ ├── train_adventures_valid_translated.txt │ │ ├── truncated_advanture_train.json │ │ ├── truncated_advanture_valid.json │ │ └── valid_advantures_valid_translated.txt │ ├── raw_data/ │ │ ├── Jinyong/ │ │ │ ├── 书剑恩仇录.txt │ │ │ ├── 侠客行.txt │ │ │ ├── 倚天屠龙记.txt │ │ │ ├── 天龙八部.txt │ │ │ ├── 射雕英雄传.txt │ │ │ ├── 白马啸西风.txt │ │ │ ├── 碧血剑.txt │ │ │ ├── 神雕侠侣.txt │ │ │ ├── 笑傲江湖.txt │ │ │ ├── 越女剑.txt │ │ │ ├── 连城诀.txt │ │ │ ├── 雪山飞狐.txt │ │ │ ├── 飞狐外传.txt │ │ │ ├── 鸳鸯刀.txt │ │ │ └── 鹿鼎记.txt │ │ └── 邪气凛然.txt │ ├── 一次冒险.txt │ ├── 我从来都不主动.txt │ ├── 盗墓实录.txt │ └── 邪气凛然.txt ├── readme.md ├── requirements.txt ├── tf2gpt/ │ ├── .gitignore │ ├── README.md │ ├── loading.ipynb │ ├── model.py │ ├── predict.ipynb │ ├── test.ipynb │ ├── train.ipynb │ └── vocab.txt ├── utils/ │ ├── gpt2_tokenizer.py │ ├── progress_bar.py │ ├── story_helper.py │ └── story_util.py └── 标注.ipynb
SYMBOL INDEX (778 symbols across 48 files)
FILE: CPM-Generate/arguments.py
function add_model_config_args (line 23) | def add_model_config_args(parser):
function add_fp16_config_args (line 70) | def add_fp16_config_args(parser):
function add_training_args (line 99) | def add_training_args(parser):
function add_evaluation_args (line 177) | def add_evaluation_args(parser):
function add_text_generate_args (line 210) | def add_text_generate_args(parser):
function add_data_args (line 221) | def add_data_args(parser):
function get_args (line 311) | def get_args():
FILE: CPM-Generate/configure_data.py
class DataConfig (line 24) | class DataConfig:
method __init__ (line 26) | def __init__(self, defaults={}):
method apply (line 30) | def apply(self, args):
method set_defaults (line 36) | def set_defaults(self, **kwargs):
method apply_defaults (line 40) | def apply_defaults(self, args):
function make_data_loader (line 47) | def make_data_loader(dataset, batch_size, args):
function make_tfrecord_loaders (line 79) | def make_tfrecord_loaders(args):
function make_loaders (line 116) | def make_loaders(args):
function get_split (line 204) | def get_split(args):
function configure_data (line 228) | def configure_data():
FILE: CPM-Generate/data/bert_dataset.py
function build_train_valid_test_datasets (line 32) | def build_train_valid_test_datasets(data_prefix, data_impl, splits_string,
class BertDataset (line 102) | class BertDataset(Dataset):
method __init__ (line 104) | def __init__(self, name, indexed_dataset, data_prefix,
method __len__ (line 136) | def __len__(self):
method __getitem__ (line 139) | def __getitem__(self, idx):
function get_indexed_dataset_ (line 157) | def get_indexed_dataset_(data_prefix, data_impl, skip_warmup):
function get_train_valid_test_split_ (line 178) | def get_train_valid_test_split_(splits_string, size):
function get_samples_mapping_ (line 206) | def get_samples_mapping_(indexed_dataset,
FILE: CPM-Generate/data/dataset_utils.py
function compile_helper (line 25) | def compile_helper():
function build_training_sample (line 38) | def build_training_sample(sample,
function get_a_and_b_segments (line 100) | def get_a_and_b_segments(sample, np_rng):
function truncate_segments (line 132) | def truncate_segments(tokens_a, tokens_b, len_a, len_b, max_num_tokens, ...
function create_tokens_and_tokentypes (line 153) | def create_tokens_and_tokentypes(tokens_a, tokens_b, cls_id, sep_id):
function is_start_piece (line 183) | def is_start_piece(piece):
function create_masked_lm_predictions (line 192) | def create_masked_lm_predictions(tokens,
function pad_and_convert_to_numpy (line 376) | def pad_and_convert_to_numpy(tokens, tokentypes, masked_positions,
FILE: CPM-Generate/data/gpt2_dataset.py
function build_train_valid_test_datasets (line 30) | def build_train_valid_test_datasets(data_prefix, data_impl, splits_string,
function get_indexed_dataset_ (line 73) | def get_indexed_dataset_(data_prefix, data_impl, skip_warmup):
class GPT2Dataset (line 89) | class GPT2Dataset(torch.utils.data.Dataset):
method __init__ (line 91) | def __init__(self, name, data_prefix, documents, indexed_dataset,
method __len__ (line 106) | def __len__(self):
method __getitem__ (line 111) | def __getitem__(self, idx):
function _build_index_mappings (line 140) | def _build_index_mappings(name, data_prefix, documents, sizes,
function _num_tokens (line 231) | def _num_tokens(documents, sizes):
function _num_epochs (line 236) | def _num_epochs(tokens_per_epoch, seq_length, num_samples):
function _build_doc_idx (line 251) | def _build_doc_idx(documents, num_epochs, np_rng):
function _build_sample_idx (line 262) | def _build_sample_idx(sizes, doc_idx, seq_length,
function _build_shuffle_idx (line 311) | def _build_shuffle_idx(size, np_rng):
FILE: CPM-Generate/data/helpers.cpp
function build_sample_idx (line 36) | py::array build_sample_idx(const py::array_t<int32_t>& sizes_,
function get_target_sample_len (line 125) | inline int32_t get_target_sample_len(const int32_t short_seq_ratio,
function build_mapping_impl (line 138) | py::array build_mapping_impl(const py::array_t<int64_t>& docs_,
function build_mapping (line 370) | py::array build_mapping(const py::array_t<int64_t>& docs_,
function PYBIND11_MODULE (line 397) | PYBIND11_MODULE(helpers, m) {
FILE: CPM-Generate/data/indexed_dataset.py
function __best_fitting_dtype (line 24) | def __best_fitting_dtype(vocab_size=None):
function get_available_dataset_impl (line 31) | def get_available_dataset_impl():
function infer_dataset_impl (line 35) | def infer_dataset_impl(path):
function make_builder (line 51) | def make_builder(out_file, impl, vocab_size=None):
function make_dataset (line 58) | def make_dataset(path, impl, skip_warmup=False):
function dataset_exists (line 75) | def dataset_exists(path, impl):
function read_longs (line 82) | def read_longs(f, n):
function write_longs (line 88) | def write_longs(f, a):
function code (line 104) | def code(dtype):
function index_file_path (line 111) | def index_file_path(prefix_path):
function data_file_path (line 115) | def data_file_path(prefix_path):
function create_doc_idx (line 119) | def create_doc_idx(sizes):
class IndexedDataset (line 127) | class IndexedDataset(torch.utils.data.Dataset):
method __init__ (line 131) | def __init__(self, path):
method read_index (line 137) | def read_index(self, path):
method read_data (line 155) | def read_data(self, path):
method check_index (line 158) | def check_index(self, i):
method __del__ (line 162) | def __del__(self):
method __getitem__ (line 167) | def __getitem__(self, idx):
method __len__ (line 191) | def __len__(self):
method num_tokens (line 194) | def num_tokens(self, index):
method size (line 197) | def size(self, index):
method exists (line 201) | def exists(path):
method supports_prefetch (line 207) | def supports_prefetch(self):
class IndexedCachedDataset (line 211) | class IndexedCachedDataset(IndexedDataset):
method __init__ (line 213) | def __init__(self, path):
method supports_prefetch (line 219) | def supports_prefetch(self):
method prefetch (line 222) | def prefetch(self, indices):
method __getitem__ (line 247) | def __getitem__(self, idx):
class IndexedDatasetBuilder (line 264) | class IndexedDatasetBuilder(object):
method __init__ (line 275) | def __init__(self, out_file, dtype=np.int32):
method add_item (line 284) | def add_item(self, tensor):
method end_document (line 291) | def end_document(self):
method merge_file_ (line 294) | def merge_file_(self, another_file):
method finalize (line 314) | def finalize(self, index_file):
function _warmup_mmap_file (line 329) | def _warmup_mmap_file(path):
class MMapIndexedDataset (line 335) | class MMapIndexedDataset(torch.utils.data.Dataset):
class Index (line 336) | class Index(object):
method writer (line 340) | def writer(cls, path, dtype):
method __init__ (line 385) | def __init__(self, path, skip_warmup=False):
method __del__ (line 422) | def __del__(self):
method dtype (line 427) | def dtype(self):
method sizes (line 431) | def sizes(self):
method doc_idx (line 435) | def doc_idx(self):
method __getitem__ (line 439) | def __getitem__(self, i):
method __len__ (line 442) | def __len__(self):
method __init__ (line 445) | def __init__(self, path, skip_warmup=False):
method __getstate__ (line 454) | def __getstate__(self):
method __setstate__ (line 457) | def __setstate__(self, state):
method _do_init (line 460) | def _do_init(self, path, skip_warmup):
method __del__ (line 472) | def __del__(self):
method __len__ (line 477) | def __len__(self):
method __getitem__ (line 481) | def __getitem__(self, idx):
method get (line 500) | def get(self, idx, offset=0, length=None):
method sizes (line 515) | def sizes(self):
method doc_idx (line 519) | def doc_idx(self):
method get_doc_idx (line 522) | def get_doc_idx(self):
method set_doc_idx (line 525) | def set_doc_idx(self, doc_idx_):
method supports_prefetch (line 529) | def supports_prefetch(self):
method exists (line 533) | def exists(path):
class MMapIndexedDatasetBuilder (line 539) | class MMapIndexedDatasetBuilder(object):
method __init__ (line 540) | def __init__(self, out_file, dtype=np.int64):
method add_item (line 546) | def add_item(self, tensor):
method end_document (line 551) | def end_document(self):
method merge_file_ (line 554) | def merge_file_(self, another_file):
method finalize (line 566) | def finalize(self, index_file):
FILE: CPM-Generate/data/samplers.py
class RandomSampler (line 22) | class RandomSampler(data.sampler.Sampler):
method __init__ (line 35) | def __init__(self, data_source, replacement=False, num_samples=None):
method num_samples (line 55) | def num_samples(self):
method __iter__ (line 61) | def __iter__(self):
method __len__ (line 71) | def __len__(self):
method set_epoch (line 74) | def set_epoch(self, epoch):
class DistributedBatchSampler (line 78) | class DistributedBatchSampler(data.sampler.BatchSampler):
method __init__ (line 95) | def __init__(self, sampler, batch_size, drop_last, rank=-1,
method __iter__ (line 110) | def __iter__(self):
method data_iterator (line 132) | def data_iterator(self, _iter, wrap_around=False):
method _batch (line 142) | def _batch(self, batch):
FILE: CPM-Generate/data/test/test_indexed_dataset.py
function test_indexed_dataset (line 17) | def test_indexed_dataset(args):
function test_indexed_dataset_get (line 43) | def test_indexed_dataset_get(args):
function main (line 82) | def main():
FILE: CPM-Generate/data_utils/__init__.py
function should_split (line 29) | def should_split(split):
function get_ext (line 40) | def get_ext(path):
function get_dataset (line 44) | def get_dataset(path, **kwargs):
function supported_corpus (line 57) | def supported_corpus(corpus_name):
function make_dataset (line 61) | def make_dataset(path, seq_length, text_key, label_key, lazy=False, proc...
FILE: CPM-Generate/data_utils/corpora.py
class wikipedia (line 19) | class wikipedia(json_dataset):
method __init__ (line 27) | def __init__(self, **kwargs):
class webtext (line 37) | class webtext(json_dataset):
method __init__ (line 45) | def __init__(self, **kwargs):
FILE: CPM-Generate/data_utils/datasets.py
class ConcatDataset (line 37) | class ConcatDataset(data.Dataset):
method cumsum (line 48) | def cumsum(sequence):
method __init__ (line 56) | def __init__(self, datasets, **kwargs):
method SetTokenizer (line 66) | def SetTokenizer(self, tokenizer):
method GetTokenizer (line 70) | def GetTokenizer(self):
method __len__ (line 73) | def __len__(self):
method __getitem__ (line 76) | def __getitem__(self, idx):
method lens (line 85) | def lens(self):
method X (line 97) | def X(self):
method Y (line 105) | def Y(self):
method cummulative_sizes (line 114) | def cummulative_sizes(self):
class SplitDataset (line 119) | class SplitDataset(data.Dataset):
method __init__ (line 129) | def __init__(self, ds, split_inds, **kwargs):
method __len__ (line 138) | def __len__(self):
method __getitem__ (line 141) | def __getitem__(self, index):
method SetTokenizer (line 144) | def SetTokenizer(self, tokenizer):
method GetTokenizer (line 147) | def GetTokenizer(self):
method X (line 151) | def X(self):
method Y (line 157) | def Y(self):
method __iter__ (line 162) | def __iter__(self):
function split_ds (line 166) | def split_ds(ds, split=[.8,.2,.0], shuffle=True):
class csv_dataset (line 199) | class csv_dataset(data.Dataset):
method __init__ (line 217) | def __init__(self, path, tokenizer=None, preprocess_fn=None, delim=',',
method SetTokenizer (line 256) | def SetTokenizer(self, tokenizer):
method GetTokenizer (line 265) | def GetTokenizer(self):
method tokenizer (line 269) | def tokenizer(self):
method __len__ (line 274) | def __len__(self):
method __getitem__ (line 277) | def __getitem__(self, index):
method write (line 292) | def write(self, writer_gen=None, path=None, skip_header=False):
class json_dataset (line 315) | class json_dataset(data.Dataset):
method __init__ (line 330) | def __init__(self, path, tokenizer=None, preprocess_fn=None, binarize_...
method SetTokenizer (line 350) | def SetTokenizer(self, tokenizer):
method GetTokenizer (line 359) | def GetTokenizer(self):
method tokenizer (line 363) | def tokenizer(self):
method __getitem__ (line 368) | def __getitem__(self, index):
method __len__ (line 383) | def __len__(self):
method write (line 386) | def write(self, writer_gen=None, path=None, skip_header=False):
method save_json_stream (line 427) | def save_json_stream(self, save_path, json_stream):
method load_json_stream (line 440) | def load_json_stream(self, load_path):
class GPT2Dataset (line 456) | class GPT2Dataset(data.Dataset):
method __init__ (line 458) | def __init__(self, ds,
method init_weighting (line 479) | def init_weighting(self):
method get_weighted_samples (line 491) | def get_weighted_samples(self, np_rng):
method __len__ (line 498) | def __len__(self):
method __getitem__ (line 501) | def __getitem__(self, idx):
method getidx (line 541) | def getidx(self, data_idx):
method pad_seq (line 551) | def pad_seq(self, seq):
method contains_sentence_end (line 557) | def contains_sentence_end(self, tok):
class bert_sentencepair_dataset (line 567) | class bert_sentencepair_dataset(data.Dataset):
method __init__ (line 579) | def __init__(self, ds, max_seq_len=512, mask_lm_prob=.15, max_preds_pe...
method get_weighting (line 600) | def get_weighting(self):
method get_weighted_samples (line 611) | def get_weighted_samples(self, np_rng):
method __len__ (line 618) | def __len__(self):
method __getitem__ (line 621) | def __getitem__(self, idx):
method sentence_split (line 648) | def sentence_split(self, document):
method sentence_tokenize (line 659) | def sentence_tokenize(self, sent, sentence_num=0, beginning=False, end...
method get_doc (line 666) | def get_doc(self, idx):
method create_random_sentencepair (line 673) | def create_random_sentencepair(self, target_seq_length, rng, np_rng):
method truncate_seq_pair (line 754) | def truncate_seq_pair(self, a, b, max_seq_len, rng):
method mask_token (line 785) | def mask_token(self, idx, tokens, types, vocab_words, rng):
method pad_seq (line 803) | def pad_seq(self, seq):
method create_masked_lm_predictions (line 810) | def create_masked_lm_predictions(self, a, b, mask_lm_prob, max_preds_p...
FILE: CPM-Generate/data_utils/file_utils.py
function url_to_filename (line 43) | def url_to_filename(url, etag=None):
function filename_to_url (line 61) | def filename_to_url(filename, cache_dir=None):
function cached_path (line 87) | def cached_path(url_or_filename, cache_dir=None):
function split_s3_path (line 117) | def split_s3_path(url):
function s3_request (line 130) | def s3_request(func):
function s3_etag (line 150) | def s3_etag(url):
function s3_get (line 159) | def s3_get(url, temp_file):
function http_get (line 166) | def http_get(url, temp_file):
function get_from_cache (line 178) | def get_from_cache(url, cache_dir=None):
function read_set_from_file (line 238) | def read_set_from_file(filename):
function get_file_extension (line 250) | def get_file_extension(path, dot=True, lower=True):
FILE: CPM-Generate/data_utils/lazy_loader.py
function get_lazy_path (line 25) | def get_lazy_path(path):
function exists_lazy (line 31) | def exists_lazy(path, data_type='data'):
function make_lazy (line 44) | def make_lazy(path, strs, data_type='data'):
function split_strings (line 70) | def split_strings(strings, start, chr_lens):
class ProcessorTokenizer (line 76) | class ProcessorTokenizer:
method __init__ (line 81) | def __init__(self, tokenizer, process_fn=None):
method __call__ (line 85) | def __call__(self, string):
class lazy_array_loader (line 92) | class lazy_array_loader(object):
method __init__ (line 110) | def __init__(self, path, data_type='data', mem_map=False, map_fn=None):
method SetTokenizer (line 129) | def SetTokenizer(self, tokenizer):
method GetTokenizer (line 141) | def GetTokenizer(self):
method __getitem__ (line 144) | def __getitem__(self, index):
method __len__ (line 171) | def __len__(self):
method file_read (line 174) | def file_read(self, start=0, end=None):
FILE: CPM-Generate/data_utils/samplers.py
class RandomSampler (line 24) | class RandomSampler(data.sampler.Sampler):
method __init__ (line 36) | def __init__(self, data_source, replacement=False, num_samples=None):
method num_samples (line 54) | def num_samples(self):
method __iter__ (line 60) | def __iter__(self):
method __len__ (line 69) | def __len__(self):
method set_epoch (line 72) | def set_epoch(self, epoch):
class DistributedBatchSampler (line 75) | class DistributedBatchSampler(data.sampler.BatchSampler):
method __init__ (line 81) | def __init__(self, sampler, batch_size, drop_last, rank=-1, world_size...
method __iter__ (line 93) | def __iter__(self):
method data_iterator (line 125) | def data_iterator(self, _iter, wrap_around=False):
method _batch (line 135) | def _batch(self, batch):
FILE: CPM-Generate/data_utils/tf_dl.py
class TFRecordDataLoader (line 25) | class TFRecordDataLoader(object):
method __init__ (line 26) | def __init__(self, records, batch_size, max_seq_len, max_preds_per_seq...
method __iter__ (line 65) | def __iter__(self):
class Record2Example (line 75) | class Record2Example(object):
method __init__ (line 76) | def __init__(self, feature_map):
method __call__ (line 79) | def __call__(self, record):
function convert_tf_example_to_torch_tensors (line 87) | def convert_tf_example_to_torch_tensors(example):
class MultiprocessLoader (line 100) | class MultiprocessLoader(object):
method __init__ (line 101) | def __init__(self, dataloader, num_workers=2):
method __iter__ (line 105) | def __iter__(self):
function _multiproc_iter (line 117) | def _multiproc_iter(dl, output_queue):
FILE: CPM-Generate/data_utils/tokenization.py
function make_tokenizer (line 31) | def make_tokenizer(tokenizer_type, corpus, model_path=None, vocab_size=N...
class Tokenization (line 46) | class Tokenization(object):
method __init__ (line 57) | def __init__(self, tokenization, text=None, original_text=None, comman...
method set_command_tokens (line 69) | def set_command_tokens(self, command_tokens):
method parse_command_tokens (line 73) | def parse_command_tokens(self):
method __getitem__ (line 82) | def __getitem__(self, index):
method __len__ (line 85) | def __len__(self):
method insert (line 88) | def insert(self, idx, other):
method append (line 102) | def append(self, other):
method extend (line 115) | def extend(self, other):
function prep_command_tokens (line 137) | def prep_command_tokens(tokenlist, token_format=token_format):
class CommandToken (line 140) | class CommandToken(object):
method __init__ (line 141) | def __init__(self, name, token, Id):
method __str__ (line 146) | def __str__(self):
function prep_type_tokens (line 165) | def prep_type_tokens(tokenlist, token_format=token_format):
class TypeToken (line 168) | class TypeToken(object):
method __init__ (line 169) | def __init__(self, name, token, Id):
method __str__ (line 174) | def __str__(self):
class Tokenizer (line 192) | class Tokenizer(object):
method __init__ (line 202) | def __init__(self, text_tokenizer, command_tokens=None, type_tokens=No...
method __call__ (line 245) | def __call__(self, text, process_fn=None):
method __len__ (line 249) | def __len__(self):
method get_command (line 253) | def get_command(self, name):
method get_type (line 257) | def get_type(self, name):
method tokens (line 262) | def tokens(self):
method vocab (line 267) | def vocab(self):
method token_types (line 272) | def token_types(self):
method token_type_vocab (line 277) | def token_type_vocab(self):
method command_tokens (line 282) | def command_tokens(self):
method command_token_vocab (line 287) | def command_token_vocab(self):
method text_tokens (line 292) | def text_tokens(self):
method text_token_vocab (line 297) | def text_token_vocab(self):
method EncodeAsIds (line 301) | def EncodeAsIds(self, text, process_fn=None):
method EncodeAsTokens (line 310) | def EncodeAsTokens(self, text, process_fn=None):
method IdToToken (line 318) | def IdToToken(self, Id, type_token=False):
method TokenToId (line 328) | def TokenToId(self, token, type_token=False):
method DecodeIds (line 338) | def DecodeIds(self, Ids, type_token=False):
method DecodeTokens (line 364) | def DecodeTokens(self, Tokens, type_token=False):
class TextTokenizer (line 389) | class TextTokenizer(object):
method __init__ (line 393) | def __init__(self):
method __call__ (line 399) | def __call__(self, text, process_fn=None):
method __len__ (line 402) | def __len__(self):
method tokens (line 406) | def tokens(self):
method vocab (line 411) | def vocab(self):
method exists (line 416) | def exists(model_path):
method Train (line 420) | def Train(self, corpus):
method EncodeAsIds (line 424) | def EncodeAsIds(self, text, process_fn=None):
method EncodeAsTokens (line 431) | def EncodeAsTokens(self, text, process_fn=None):
method IdToToken (line 438) | def IdToToken(self, Id):
method TokenToId (line 442) | def TokenToId(self, token):
method DecodeIds (line 446) | def DecodeIds(self, Ids):
method DecodeTokens (line 450) | def DecodeTokens(self, Tokens):
class CharacterLevelTokenizer (line 455) | class CharacterLevelTokenizer(TextTokenizer):
method __init__ (line 459) | def __init__(self, **kwargs):
method __len__ (line 465) | def __len__(self):
method exists (line 469) | def exists(model_path):
method Train (line 472) | def Train(self, corpus):
method tokens (line 476) | def tokens(self):
method vocab (line 480) | def vocab(self):
method EncodeAsIds (line 483) | def EncodeAsIds(self, text, process_fn=None):
method EncodeAsTokens (line 492) | def EncodeAsTokens(self, text, process_fn=None):
method IdToToken (line 501) | def IdToToken(self, Id):
method TokenToId (line 505) | def TokenToId(self, token):
method DecodeIds (line 509) | def DecodeIds(self, Ids):
method DecodeTokens (line 515) | def DecodeTokens(self, Tokens):
function get_corpus_freq (line 524) | def get_corpus_freq(dataset, filepath, filetype='tsv'):
class SentencePieceTokenizer (line 576) | class SentencePieceTokenizer(TextTokenizer):
method __init__ (line 578) | def __init__(self, model_type='bpe', vocab_size=None, corpus=None, mod...
method __len__ (line 592) | def __len__(self):
method tokens (line 596) | def tokens(self):
method vocab (line 600) | def vocab(self):
method exists (line 604) | def exists(model_path):
method load_spm_model (line 614) | def load_spm_model(self):
method Train (line 624) | def Train(self, corpus, num_text_tokens):
method EncodeAsIds (line 651) | def EncodeAsIds(self, text, process_fn=None):
method EncodeAsTokens (line 659) | def EncodeAsTokens(self, text, process_fn=None):
method IdToToken (line 667) | def IdToToken(self, Id):
method TokenToId (line 671) | def TokenToId(self, token):
method DecodeIds (line 675) | def DecodeIds(self, Ids):
method DecodeTokens (line 681) | def DecodeTokens(self, Tokens):
class BertWordPieceTokenizer (line 687) | class BertWordPieceTokenizer(Tokenizer):
method __init__ (line 692) | def __init__(self, tokenizer_model_type=None, cache_dir=None, **kwargs):
method EncodeAsIds (line 745) | def EncodeAsIds(self, text, process_fn=None):
method EncodeAsTokens (line 754) | def EncodeAsTokens(self, text, process_fn=None):
method IdToToken (line 762) | def IdToToken(self, Id, type_token=False):
method TokenToId (line 770) | def TokenToId(self, token, type_token=False):
method DecodeIds (line 778) | def DecodeIds(self, Ids, type_token=False):
method DecodeTokens (line 790) | def DecodeTokens(self, Tokens, type_token=False):
class GPT2BPETokenizer (line 799) | class GPT2BPETokenizer(Tokenizer):
method __init__ (line 800) | def __init__(self, cache_dir=None, **kwargs):
method EncodeAsIds (line 839) | def EncodeAsIds(self, text, process_fn=None):
method EncodeAsTokens (line 850) | def EncodeAsTokens(self, text, process_fn=None):
method IdToToken (line 863) | def IdToToken(self, Id, type_token=False):
method TokenToId (line 870) | def TokenToId(self, token, type_token=False):
method DecodeIds (line 877) | def DecodeIds(self, Ids, type_token=False):
method DecodeTokens (line 884) | def DecodeTokens(self, Tokens, type_token=False):
FILE: CPM-Generate/data_utils/tokenization_gpt2.py
function lru_cache (line 33) | def lru_cache():
function bytes_to_unicode (line 54) | def bytes_to_unicode():
function get_pairs (line 76) | def get_pairs(word):
class GPT2Tokenizer (line 88) | class GPT2Tokenizer(object):
method from_pretrained (line 94) | def from_pretrained(cls, pretrained_model_name_or_path, cache_dir=None...
method __init__ (line 146) | def __init__(self, vocab_file, merges_file, model_file, errors='replac...
method vocab_size (line 173) | def vocab_size(self):
method __len__ (line 176) | def __len__(self):
method eod (line 180) | def eod(self):
method set_special_tokens (line 183) | def set_special_tokens(self, special_tokens):
method bpe (line 196) | def bpe(self, token):
method tokenize (line 237) | def tokenize(self, text):
method convert_tokens_to_ids (line 246) | def convert_tokens_to_ids(self, tokens):
method convert_ids_to_tokens (line 267) | def convert_ids_to_tokens(self, ids, skip_special_tokens=False):
method encode (line 278) | def encode(self, text):
method decode (line 282) | def decode(self, tokens):
method save_vocabulary (line 288) | def save_vocabulary(self, vocab_path):
FILE: CPM-Generate/data_utils/wordpiece.py
function load_vocab (line 50) | def load_vocab(vocab_file):
function whitespace_tokenize (line 65) | def whitespace_tokenize(text):
class BertTokenizer (line 74) | class BertTokenizer(object):
method __init__ (line 77) | def __init__(self, vocab_file, do_lower_case=True, max_len=None, do_ba...
method tokenize (line 107) | def tokenize(self, text):
method convert_tokens_to_ids (line 117) | def convert_tokens_to_ids(self, tokens):
method convert_ids_to_tokens (line 130) | def convert_ids_to_tokens(self, ids):
method from_pretrained (line 138) | def from_pretrained(cls, pretrained_model_name_or_path, cache_dir=None...
class BasicTokenizer (line 176) | class BasicTokenizer(object):
method __init__ (line 179) | def __init__(self,
method tokenize (line 190) | def tokenize(self, text):
method _run_strip_accents (line 211) | def _run_strip_accents(self, text):
method _run_split_on_punc (line 222) | def _run_split_on_punc(self, text):
method _tokenize_chinese_chars (line 244) | def _tokenize_chinese_chars(self, text):
method _is_chinese_char (line 257) | def _is_chinese_char(self, cp):
method _clean_text (line 279) | def _clean_text(self, text):
class WordpieceTokenizer (line 293) | class WordpieceTokenizer(object):
method __init__ (line 296) | def __init__(self, vocab, unk_token="[UNK]", max_input_chars_per_word=...
method tokenize (line 301) | def tokenize(self, text):
function _is_whitespace (line 353) | def _is_whitespace(char):
function _is_control (line 365) | def _is_control(char):
function _is_punctuation (line 377) | def _is_punctuation(char):
FILE: CPM-Generate/fp16/fp16.py
function conversion_helper (line 28) | def conversion_helper(val, conversion):
function fp32_to_fp16 (line 37) | def fp32_to_fp16(val):
function fp16_to_fp32 (line 48) | def fp16_to_fp32(val):
class FP16_Module (line 59) | class FP16_Module(nn.Module):
method __init__ (line 60) | def __init__(self, module):
method forward (line 64) | def forward(self, *inputs, **kwargs):
method state_dict (line 67) | def state_dict(self, destination=None, prefix='', keep_vars=False):
method load_state_dict (line 70) | def load_state_dict(self, state_dict, strict=True):
class FP16_Optimizer (line 74) | class FP16_Optimizer(object):
method __init__ (line 168) | def __init__(self,
method maybe_print (line 242) | def maybe_print(self, msg):
method __getstate__ (line 246) | def __getstate__(self):
method __setstate__ (line 249) | def __setstate__(self, state):
method zero_grad (line 252) | def zero_grad(self, set_grads_to_None=False):
method _check_overflow (line 278) | def _check_overflow(self):
method _update_scale (line 288) | def _update_scale(self, has_overflow=False):
method _master_params_to_model_params (line 291) | def _master_params_to_model_params(self):
method _model_params_to_master_params (line 295) | def _model_params_to_master_params(self):
method _model_grads_to_master_grads (line 301) | def _model_grads_to_master_grads(self):
method _downscale_master (line 305) | def _downscale_master(self):
method clip_master_grads (line 312) | def clip_master_grads(self, max_norm, norm_type=2):
method state_dict (line 336) | def state_dict(self):
method load_state_dict (line 357) | def load_state_dict(self, state_dict):
method step (line 399) | def step(self, closure=None): # could add clip option.
method _step_with_closure (line 455) | def _step_with_closure(self, closure):
method backward (line 494) | def backward(self, loss, update_master_grads=True, retain_graph=False):
method update_master_grads (line 556) | def update_master_grads(self):
method inspect_master_grad_data (line 569) | def inspect_master_grad_data(self):
method _get_loss_scale (line 604) | def _get_loss_scale(self):
method _set_loss_scale (line 607) | def _set_loss_scale(self, value):
method _get_state (line 613) | def _get_state(self):
method _set_state (line 616) | def _set_state(self, value):
method _get_param_groups (line 623) | def _get_param_groups(self):
method _set_param_groups (line 626) | def _set_param_groups(self, value):
FILE: CPM-Generate/fp16/fp16util.py
class tofp16 (line 24) | class tofp16(nn.Module):
method __init__ (line 32) | def __init__(self):
method forward (line 35) | def forward(self, input):
function BN_convert_float (line 39) | def BN_convert_float(module):
function network_to_half (line 52) | def network_to_half(network):
function convert_module (line 61) | def convert_module(module, dtype):
function convert_network (line 77) | def convert_network(network, dtype):
class FP16Model (line 88) | class FP16Model(nn.Module):
method __init__ (line 93) | def __init__(self, network):
method forward (line 97) | def forward(self, *inputs):
function backwards_debug_hook (line 102) | def backwards_debug_hook(grad):
function prep_param_lists (line 105) | def prep_param_lists(model, flat_master=False):
function model_grads_to_master_grads (line 151) | def model_grads_to_master_grads(model_params, master_params, flat_master...
function master_params_to_model_params (line 173) | def master_params_to_model_params(model_params, master_params, flat_mast...
function to_python_float (line 191) | def to_python_float(t):
FILE: CPM-Generate/fp16/loss_scaler.py
function to_python_float (line 20) | def to_python_float(t):
class LossScaler (line 26) | class LossScaler:
method __init__ (line 38) | def __init__(self, scale=1):
method has_overflow (line 42) | def has_overflow(self, params):
method _has_inf_or_nan (line 46) | def _has_inf_or_nan(x):
method update_scale (line 49) | def update_scale(self, overflow):
method loss_scale (line 53) | def loss_scale(self):
method scale_gradient (line 56) | def scale_gradient(self, module, grad_in, grad_out):
method backward (line 59) | def backward(self, loss, retain_graph=False):
class DynamicLossScaler (line 63) | class DynamicLossScaler:
method __init__ (line 89) | def __init__(self,
method has_overflow_serial (line 107) | def has_overflow_serial(self, params):
method has_overflow (line 114) | def has_overflow(self, params):
method _has_inf_or_nan (line 127) | def _has_inf_or_nan(x):
method update_scale (line 148) | def update_scale(self, overflow):
method loss_scale (line 175) | def loss_scale(self):
method scale_gradient (line 178) | def scale_gradient(self, module, grad_in, grad_out):
method backward (line 181) | def backward(self, loss, retain_graph=False):
FILE: CPM-Generate/generate_samples.py
function get_masks_and_position_ids (line 39) | def get_masks_and_position_ids(data,
function initialize_distributed (line 92) | def initialize_distributed(args):
function set_random_seed (line 113) | def set_random_seed(seed):
function get_batch (line 122) | def get_batch(context_tokens, device, args):
function top_k_logits (line 136) | def top_k_logits(logits, top_k=0, top_p=0.0, filter_value=-float('Inf')):
function generate_samples (line 164) | def generate_samples(model, tokenizer, args, device):
function prepare_tokenizer (line 261) | def prepare_tokenizer(args):
function get_model (line 285) | def get_model(args):
function setup_model (line 324) | def setup_model(args):
function main (line 333) | def main():
FILE: CPM-Generate/model/distributed.py
class DistributedDataParallel (line 24) | class DistributedDataParallel(Module):
method __init__ (line 26) | def __init__(self, module):
method forward (line 76) | def forward(self, *inputs, **kwargs):
method state_dict (line 80) | def state_dict(self, destination=None, prefix='', keep_vars=False):
method load_state_dict (line 89) | def load_state_dict(self, state_dict, strict=True):
FILE: CPM-Generate/model/gpt2_modeling.py
function init_method_normal (line 24) | def init_method_normal(std=0.02):
class GPT2Model (line 35) | class GPT2Model(torch.nn.Module):
method __init__ (line 42) | def __init__(self,
method forward (line 83) | def forward(self, input_ids, position_ids, attention_mask):
function gpt2_get_params_for_weight_decay_optimization (line 108) | def gpt2_get_params_for_weight_decay_optimization(module):
FILE: CPM-Generate/model/model.py
function get_params_for_weight_decay_optimization (line 25) | def get_params_for_weight_decay_optimization(module):
class BertModel (line 45) | class BertModel(torch.nn.Module):
method __init__ (line 47) | def __init__(self, args):
method forward (line 78) | def forward(self, input_tokens, token_type_ids=None,
method state_dict (line 84) | def state_dict(self, destination=None, prefix='', keep_vars=False):
method load_state_dict (line 88) | def load_state_dict(self, state_dict, strict=True):
FILE: CPM-Generate/model/modeling.py
function normal_init_method (line 42) | def normal_init_method(mean, std):
function scaled_init_method (line 47) | def scaled_init_method(mean, std, num_layers):
function load_tf_weights_in_bert (line 70) | def load_tf_weights_in_bert(model, tf_checkpoint_path):
function gelu (line 131) | def gelu(x):
function swish (line 139) | def swish(x):
class BertConfig (line 145) | class BertConfig(object):
method __init__ (line 148) | def __init__(self,
method from_dict (line 216) | def from_dict(cls, json_object):
method from_json_file (line 224) | def from_json_file(cls, json_file):
method __repr__ (line 230) | def __repr__(self):
method to_dict (line 233) | def to_dict(self):
method to_json_string (line 238) | def to_json_string(self):
class BertLayerNorm (line 246) | class BertLayerNorm(nn.Module):
method __init__ (line 247) | def __init__(self, hidden_size, eps=1e-12):
method forward (line 255) | def forward(self, x):
class BertEmbeddings (line 261) | class BertEmbeddings(nn.Module):
method __init__ (line 264) | def __init__(self, config):
method forward (line 282) | def forward(self, input_ids, token_type_ids=None):
class BertSelfAttention (line 323) | class BertSelfAttention(nn.Module):
method __init__ (line 324) | def __init__(self, config):
method transpose_for_scores (line 340) | def transpose_for_scores(self, x):
method forward (line 345) | def forward(self, hidden_states, attention_mask):
class BertSelfOutput (line 376) | class BertSelfOutput(nn.Module):
method __init__ (line 377) | def __init__(self, config):
method forward (line 397) | def forward(self, hidden_states, input_tensor):
class BertAttention (line 410) | class BertAttention(nn.Module):
method __init__ (line 411) | def __init__(self, config):
method forward (line 422) | def forward(self, input_tensor, attention_mask):
class BertIntermediate (line 428) | class BertIntermediate(nn.Module):
method __init__ (line 429) | def __init__(self, config):
method forward (line 442) | def forward(self, hidden_states):
class BertOutput (line 448) | class BertOutput(nn.Module):
method __init__ (line 449) | def __init__(self, config):
method forward (line 469) | def forward(self, hidden_states, input_tensor):
class BertLayer (line 482) | class BertLayer(nn.Module):
method __init__ (line 483) | def __init__(self, config):
method forward (line 489) | def forward(self, hidden_states, attention_mask):
class BertEncoder (line 496) | class BertEncoder(nn.Module):
method __init__ (line 497) | def __init__(self, config):
method forward (line 512) | def forward(self, hidden_states, attention_mask, output_all_encoded_la...
class BertPooler (line 543) | class BertPooler(nn.Module):
method __init__ (line 544) | def __init__(self, config):
method forward (line 549) | def forward(self, hidden_states):
class BertPredictionHeadTransform (line 558) | class BertPredictionHeadTransform(nn.Module):
method __init__ (line 559) | def __init__(self, config):
method forward (line 567) | def forward(self, hidden_states):
class BertLMPredictionHead (line 579) | class BertLMPredictionHead(nn.Module):
method __init__ (line 580) | def __init__(self, config, bert_model_embedding_weights):
method forward (line 602) | def forward(self, hidden_states):
class BertOnlyMLMHead (line 618) | class BertOnlyMLMHead(nn.Module):
method __init__ (line 619) | def __init__(self, config, bert_model_embedding_weights):
method forward (line 623) | def forward(self, sequence_output):
class BertOnlyNSPHead (line 628) | class BertOnlyNSPHead(nn.Module):
method __init__ (line 629) | def __init__(self, config):
method forward (line 633) | def forward(self, pooled_output):
class BertPreTrainingHeads (line 638) | class BertPreTrainingHeads(nn.Module):
method __init__ (line 639) | def __init__(self, config, bert_model_embedding_weights):
method forward (line 644) | def forward(self, sequence_output, pooled_output):
class PreTrainedBertModel (line 654) | class PreTrainedBertModel(nn.Module):
method __init__ (line 658) | def __init__(self, config, *inputs, **kwargs):
method init_bert_weights (line 669) | def init_bert_weights(self, module):
method from_pretrained (line 683) | def from_pretrained(cls, pretrained_model_name, state_dict=None, cache...
class BertModel (line 797) | class BertModel(PreTrainedBertModel):
method __init__ (line 841) | def __init__(self, config):
method forward (line 848) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
class BertForPreTraining (line 886) | class BertForPreTraining(PreTrainedBertModel):
method __init__ (line 936) | def __init__(self, config):
method forward (line 942) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
class BertForMaskedLM (line 958) | class BertForMaskedLM(PreTrainedBertModel):
method __init__ (line 1000) | def __init__(self, config):
method forward (line 1006) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
class BertForNextSentencePrediction (line 1019) | class BertForNextSentencePrediction(PreTrainedBertModel):
method __init__ (line 1062) | def __init__(self, config):
method forward (line 1068) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
class BertForSequenceClassification (line 1081) | class BertForSequenceClassification(PreTrainedBertModel):
method __init__ (line 1126) | def __init__(self, config, num_labels=2):
method forward (line 1134) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
class BertForMultipleChoice (line 1147) | class BertForMultipleChoice(PreTrainedBertModel):
method __init__ (line 1191) | def __init__(self, config, num_choices=2):
method forward (line 1199) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
class BertForTokenClassification (line 1216) | class BertForTokenClassification(PreTrainedBertModel):
method __init__ (line 1261) | def __init__(self, config, num_labels=2):
method forward (line 1277) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
class BertForQuestionAnswering (line 1291) | class BertForQuestionAnswering(PreTrainedBertModel):
method __init__ (line 1338) | def __init__(self, config):
method forward (line 1354) | def forward(self, input_ids, token_type_ids=None, attention_mask=None,...
FILE: CPM-Generate/mpu/cross_entropy.py
class _VocabParallelCrossEntropy (line 25) | class _VocabParallelCrossEntropy(torch.autograd.Function):
method forward (line 28) | def forward(ctx, vocab_parallel_logits, target):
method backward (line 84) | def backward(ctx, grad_output):
function vocab_parallel_cross_entropy (line 107) | def vocab_parallel_cross_entropy(vocab_parallel_logits, target):
FILE: CPM-Generate/mpu/data.py
function _check_data_types (line 26) | def _check_data_types(keys, data, target_dtype):
function _build_key_size_numel_dictionaries (line 33) | def _build_key_size_numel_dictionaries(keys, data):
function broadcast_data (line 76) | def broadcast_data(keys, data, datatype):
FILE: CPM-Generate/mpu/grads.py
function clip_grad_norm (line 28) | def clip_grad_norm(parameters, max_norm, norm_type=2):
FILE: CPM-Generate/mpu/initialize.py
function initialize_model_parallel (line 30) | def initialize_model_parallel(model_parallel_size_):
function model_parallel_is_initialized (line 81) | def model_parallel_is_initialized():
function get_model_parallel_group (line 88) | def get_model_parallel_group():
function get_data_parallel_group (line 95) | def get_data_parallel_group():
function get_model_parallel_world_size (line 102) | def get_model_parallel_world_size():
function get_model_parallel_rank (line 107) | def get_model_parallel_rank():
function get_model_parallel_src_rank (line 112) | def get_model_parallel_src_rank():
function get_data_parallel_world_size (line 120) | def get_data_parallel_world_size():
function get_data_parallel_rank (line 125) | def get_data_parallel_rank():
function destroy_model_parallel (line 130) | def destroy_model_parallel():
FILE: CPM-Generate/mpu/layers.py
function _initialize_affine_weight (line 42) | def _initialize_affine_weight(weight, output_size, input_size,
class VocabParallelEmbedding (line 77) | class VocabParallelEmbedding(torch.nn.Module):
method __init__ (line 87) | def __init__(self, num_embeddings, embedding_dim,
method forward (line 117) | def forward(self, input_):
class ParallelEmbedding (line 136) | class ParallelEmbedding(torch.nn.Module):
method __init__ (line 146) | def __init__(self, num_embeddings, embedding_dim,
method forward (line 175) | def forward(self, input_):
class ColumnParallelLinear (line 185) | class ColumnParallelLinear(torch.nn.Module):
method __init__ (line 205) | def __init__(self, input_size, output_size, bias=True, gather_output=T...
method forward (line 239) | def forward(self, input_):
class RowParallelLinear (line 252) | class RowParallelLinear(torch.nn.Module):
method __init__ (line 278) | def __init__(self, input_size, output_size, bias=True,
method forward (line 312) | def forward(self, input_):
FILE: CPM-Generate/mpu/mappings.py
function _reduce (line 22) | def _reduce(input_):
function _split (line 36) | def _split(input_):
function _gather (line 56) | def _gather(input_):
class _CopyToModelParallelRegion (line 79) | class _CopyToModelParallelRegion(torch.autograd.Function):
method forward (line 83) | def forward(ctx, input_):
method backward (line 87) | def backward(ctx, grad_output):
class _ReduceFromModelParallelRegion (line 91) | class _ReduceFromModelParallelRegion(torch.autograd.Function):
method forward (line 95) | def forward(ctx, input_):
method backward (line 99) | def backward(ctx, grad_output):
class _ScatterToModelParallelRegion (line 103) | class _ScatterToModelParallelRegion(torch.autograd.Function):
method forward (line 107) | def forward(ctx, input_):
method backward (line 111) | def backward(ctx, grad_output):
class _GatherFromModelParallelRegion (line 115) | class _GatherFromModelParallelRegion(torch.autograd.Function):
method forward (line 119) | def forward(ctx, input_):
method backward (line 123) | def backward(ctx, grad_output):
function copy_to_model_parallel_region (line 131) | def copy_to_model_parallel_region(input_):
function reduce_from_model_parallel_region (line 134) | def reduce_from_model_parallel_region(input_):
function scatter_to_model_parallel_region (line 137) | def scatter_to_model_parallel_region(input_):
function gather_from_model_parallel_region (line 140) | def gather_from_model_parallel_region(input_):
FILE: CPM-Generate/mpu/random.py
function see_memory_usage (line 35) | def see_memory_usage(message, force=False):
function detach_variable (line 62) | def detach_variable(inputs, device=None):
function _set_cuda_rng_state (line 85) | def _set_cuda_rng_state(new_state, device=-1):
class CudaRNGStatesTracker (line 119) | class CudaRNGStatesTracker:
method __init__ (line 127) | def __init__(self):
method reset (line 133) | def reset(self):
method get_states (line 138) | def get_states(self):
method set_states (line 146) | def set_states(self, states):
method add (line 151) | def add(self, name, seed):
method fork (line 169) | def fork(self, name=_MODEL_PARALLEL_RNG_TRACKER_NAME):
function get_cuda_rng_tracker (line 193) | def get_cuda_rng_tracker():
function model_parallel_cuda_manual_seed (line 198) | def model_parallel_cuda_manual_seed(seed):
function get_partition_start (line 236) | def get_partition_start(item):
function get_partition_size (line 242) | def get_partition_size(item):
function get_full_inputs (line 248) | def get_full_inputs(tensors):
class CheckpointFunction (line 273) | class CheckpointFunction(torch.autograd.Function):
method forward (line 281) | def forward(ctx, run_function, *args):
method backward (line 332) | def backward(ctx, *args):
function checkpoint (line 375) | def checkpoint(function, *args):
function partition_activations_in_checkpoint (line 380) | def partition_activations_in_checkpoint(partition_activation):
FILE: CPM-Generate/mpu/tests/commons.py
class IdentityLayer (line 25) | class IdentityLayer(torch.nn.Module):
method __init__ (line 26) | def __init__(self, size, scale=1.0):
method forward (line 29) | def forward(self):
function set_random_seed (line 33) | def set_random_seed(seed):
function initialize_distributed (line 41) | def initialize_distributed(backend='nccl'):
function print_separator (line 75) | def print_separator(message):
FILE: CPM-Generate/mpu/tests/test_cross_entropy.py
function torch_cross_entropy (line 31) | def torch_cross_entropy(batch_size, seq_length, vocab_size,
function mpu_cross_entropy (line 46) | def mpu_cross_entropy(batch_size, seq_length, vocab_size,
function test_cross_entropy (line 60) | def test_cross_entropy(model_parallel_size):
FILE: CPM-Generate/mpu/tests/test_data.py
function test_boradcast_data (line 29) | def test_boradcast_data(model_parallel_size):
FILE: CPM-Generate/mpu/tests/test_initialize.py
function test_initialize_model_parallel (line 26) | def test_initialize_model_parallel(model_parallel_size):
function test_get_model_parallel_src_rank (line 65) | def test_get_model_parallel_src_rank(model_parallel_size_):
FILE: CPM-Generate/mpu/tests/test_layers.py
function test_parallel_embedding (line 31) | def test_parallel_embedding(model_parallel_size):
function test_initialize_affine_weight (line 109) | def test_initialize_affine_weight(model_parallel_size):
class IdentityLayer2D (line 178) | class IdentityLayer2D(torch.nn.Module):
method __init__ (line 179) | def __init__(self, m , n):
method forward (line 183) | def forward(self):
function test_column_parallel_linear (line 187) | def test_column_parallel_linear(model_parallel_size):
function test_row_parallel_linear (line 254) | def test_row_parallel_linear(model_parallel_size):
class IdentityLayer3D (line 319) | class IdentityLayer3D(torch.nn.Module):
method __init__ (line 320) | def __init__(self, m , n, k):
method forward (line 324) | def forward(self):
function parallel_self_attention (line 328) | def parallel_self_attention(model_parallel_size, num_att_heads_per_parti...
function test_parallel_self_attention (line 361) | def test_parallel_self_attention(model_parallel_size):
function parallel_transformer (line 412) | def parallel_transformer(model_parallel_size, num_att_heads_per_partition,
function test_parallel_transformer_layer (line 448) | def test_parallel_transformer_layer(model_parallel_size):
FILE: CPM-Generate/mpu/tests/test_random.py
function test_set_cuda_rng_state (line 26) | def test_set_cuda_rng_state(model_parallel_size):
function test_cuda_rng_tracker (line 88) | def test_cuda_rng_tracker(model_parallel_size):
function test_model_parallel_cuda_manual_seed (line 159) | def test_model_parallel_cuda_manual_seed(model_parallel_size):
FILE: CPM-Generate/mpu/transformer.py
class GPT2ParallelSelfAttention (line 36) | class GPT2ParallelSelfAttention(torch.nn.Module):
method __init__ (line 62) | def __init__(self, hidden_size, num_attention_heads,
method _transpose_for_scores (line 94) | def _transpose_for_scores(self, tensor):
method forward (line 104) | def forward(self, hidden_states, ltor_mask):
function gelu_impl (line 153) | def gelu_impl(x):
function gelu (line 158) | def gelu(x):
class GPT2ParallelMLP (line 162) | class GPT2ParallelMLP(torch.nn.Module):
method __init__ (line 181) | def __init__(self, hidden_size, output_dropout_prob, init_method,
method forward (line 199) | def forward(self, hidden_states):
class GPT2ParallelTransformerLayer (line 210) | class GPT2ParallelTransformerLayer(torch.nn.Module):
method __init__ (line 238) | def __init__(self,
method forward (line 274) | def forward(self, hidden_states, ltor_mask):
function unscaled_init_method (line 294) | def unscaled_init_method(sigma):
function scaled_init_method (line 302) | def scaled_init_method(sigma, num_layers):
class GPT2ParallelTransformer (line 311) | class GPT2ParallelTransformer(torch.nn.Module):
method __init__ (line 345) | def __init__(self,
method forward (line 382) | def forward(self, hidden_states, attention_mask):
class BertParallelSelfAttention (line 411) | class BertParallelSelfAttention(torch.nn.Module):
method __init__ (line 436) | def __init__(self, hidden_size, num_attention_heads,
method _transpose_for_scores (line 462) | def _transpose_for_scores(self, tensor):
method forward (line 472) | def forward(self, hidden_states, attention_mask):
class BertParallelTransformerOutput (line 518) | class BertParallelTransformerOutput(torch.nn.Module):
method __init__ (line 521) | def __init__(self, input_size, output_size, dropout_prob,
method forward (line 533) | def forward(self, hidden_states, input_tensor):
class BertParallelTransformerLayer (line 541) | class BertParallelTransformerLayer(torch.nn.Module):
method __init__ (line 572) | def __init__(self,
method forward (line 607) | def forward(self, hidden_states, attention_mask):
FILE: CPM-Generate/mpu/utils.py
function ensure_divisibility (line 20) | def ensure_divisibility(numerator, denominator):
function divide (line 26) | def divide(numerator, denominator):
function split_tensor_along_last_dim (line 33) | def split_tensor_along_last_dim(tensor, num_partitions,
class VocabUtility (line 54) | class VocabUtility:
method vocab_range_from_per_partition_vocab_size (line 60) | def vocab_range_from_per_partition_vocab_size(per_partition_vocab_size,
method vocab_range_from_global_vocab_size (line 67) | def vocab_range_from_global_vocab_size(global_vocab_size, rank, world_...
FILE: CPM-Generate/utils.py
function print_rank_0 (line 30) | def print_rank_0(message):
function print_args (line 38) | def print_args(args):
function print_params_min_max_norm (line 47) | def print_params_min_max_norm(optimizer, iteration):
class Timers (line 67) | class Timers:
class Timer (line 70) | class Timer:
method __init__ (line 73) | def __init__(self, name):
method start (line 79) | def start(self):
method stop (line 86) | def stop(self):
method reset (line 93) | def reset(self):
method elapsed (line 98) | def elapsed(self, reset=True):
method __init__ (line 114) | def __init__(self):
method __call__ (line 117) | def __call__(self, name):
method log (line 122) | def log(self, names, normalizer=1.0, reset=True):
function report_memory (line 133) | def report_memory(name):
function get_checkpoint_name (line 148) | def get_checkpoint_name(checkpoints_path, iteration, release=False, zero...
function ensure_directory_exists (line 160) | def ensure_directory_exists(filename):
function get_checkpoint_tracker_filename (line 166) | def get_checkpoint_tracker_filename(checkpoints_path):
function save_zero_checkpoint (line 170) | def save_zero_checkpoint(args, iteration, optimizer):
function save_checkpoint (line 178) | def save_checkpoint(iteration, model, optimizer,
function save_ds_checkpoint (line 224) | def save_ds_checkpoint(iteration, model, args):
function get_checkpoint_iteration (line 240) | def get_checkpoint_iteration(args):
function load_checkpoint_model (line 267) | def load_checkpoint_model(model, args):
function load_weights (line 302) | def load_weights(src, dst, dst2src=False):
function load_mlp (line 322) | def load_mlp(our, oai, dst2src=False):
function load_attention (line 326) | def load_attention(our, oai, dst2src=False):
function load_transformer_layer (line 330) | def load_transformer_layer(our, oai, dst2src=False):
function move_weights (line 336) | def move_weights(our, oai, dst2src=False):
FILE: tf2gpt/model.py
function gelu (line 21) | def gelu(x):
function get_attention_mask (line 62) | def get_attention_mask(dim):
function get_dense (line 84) | def get_dense(units, name=None, stddev=0.02):
class Attention (line 93) | class Attention(tf.keras.layers.Layer):
method __init__ (line 94) | def __init__(self,
method call (line 116) | def call(self, x,kv_cache=None, **kwargs):
class Transformer (line 169) | class Transformer(tf.keras.layers.Layer):
method __init__ (line 171) | def __init__(self,
method call (line 199) | def call(self, x, kv_cache=None, **kwargs):
class PositionEmbedding (line 218) | class PositionEmbedding(tf.keras.layers.Layer):
method __init__ (line 220) | def __init__(self,
method call (line 230) | def call(self, x, kv_cache=None, **kwargs):
class GPT (line 240) | class GPT(tf.keras.Model):
method __init__ (line 241) | def __init__(self,
method call (line 270) | def call(self, x, kv_cache=None, use_cache=False, **kwargs):
method train_step (line 288) | def train_step(self, data,mask=None):
method eval_step (line 317) | def eval_step(self, data,mask=None):
FILE: utils/gpt2_tokenizer.py
function lru_cache (line 33) | def lru_cache():
function bytes_to_unicode (line 54) | def bytes_to_unicode():
function get_pairs (line 76) | def get_pairs(word):
class GPT2Tokenizer (line 88) | class GPT2Tokenizer(object):
method from_pretrained (line 94) | def from_pretrained(cls, pretrained_model_name_or_path, cache_dir=None...
method __init__ (line 146) | def __init__(self, vocab_file, merges_file, model_file, errors='replac...
method vocab_size (line 173) | def vocab_size(self):
method __len__ (line 176) | def __len__(self):
method eod (line 180) | def eod(self):
method set_special_tokens (line 183) | def set_special_tokens(self, special_tokens):
method bpe (line 196) | def bpe(self, token):
method tokenize (line 237) | def tokenize(self, text):
method convert_tokens_to_ids (line 246) | def convert_tokens_to_ids(self, tokens):
method convert_ids_to_tokens (line 267) | def convert_ids_to_tokens(self, ids, skip_special_tokens=False):
method encode (line 278) | def encode(self, text):
method decode (line 282) | def decode(self, tokens):
method save_vocabulary (line 289) | def save_vocabulary(self, vocab_path):
FILE: utils/progress_bar.py
class ProgressBar (line 7) | class ProgressBar():
method __init__ (line 8) | def __init__(self,worksum,info="",auto_display=True):
method startjob (line 13) | def startjob(self):
method complete (line 15) | def complete(self,num):
method display_progress_bar (line 20) | def display_progress_bar(self):
FILE: utils/story_helper.py
function print_warp (line 21) | def print_warp(instr):
function initize (line 25) | def initize(model_path="./tmp_weight"):
function sample_gpt (line 29) | def sample_gpt(tokenizer, gpt, sentence, number=1, length=20, p=0.9,k=40...
class Story (line 48) | class Story():
method __init__ (line 49) | def __init__(self,beginning,story_max_len=200,context_len=12):
method response_quality_ok (line 54) | def response_quality_ok(self,response):
method action (line 59) | def action(self,action):
method interactive (line 83) | def interactive(self):
FILE: utils/story_util.py
class Story (line 1) | class Story():
method __init__ (line 2) | def __init__(self,background,title):
method add_action (line 7) | def add_action(self,action,action_type,result,summary):
method to_str (line 10) | def to_str(self):
method to_dungeon_format (line 18) | def to_dungeon_format(self):
method to_normal_format (line 27) | def to_normal_format(self):
method from_file (line 36) | def from_file(self,fname):
method load_content (line 41) | def load_content(self,content):
class Stories (line 81) | class Stories():
method __init__ (line 82) | def __init__(self,fname):
Copy disabled (too large)
Download .json
Condensed preview — 108 files, each showing path, character count, and a content snippet. Download the .json file for the full structured content (25,157K chars).
[
{
"path": "3rd/gdown.pl/LICENSE.txt",
"chars": 35147,
"preview": " GNU GENERAL PUBLIC LICENSE\n Version 3, 29 June 2007\n\n Copyright (C) 2007 Free "
},
{
"path": "3rd/gdown.pl/README.md",
"chars": 1997,
"preview": "gdown.pl\n========\n\nGoogle Drive direct download of big files\n\nRequirements\n============\n\n*wget* and *Perl* must be in th"
},
{
"path": "3rd/gdown.pl/gdown.pl",
"chars": 3184,
"preview": "#!/usr/bin/env perl\n#\n# Google Drive direct download of big files\n# ./gdown.pl 'gdrive file url' ['desired file name']\n#"
},
{
"path": "CPM-Generate/.gitignore",
"chars": 1799,
"preview": "# Byte-compiled / optimized / DLL files\n__pycache__/\n*.py[cod]\n*$py.class\n\n# C extensions\n*.so\n\n# Distribution / packagi"
},
{
"path": "CPM-Generate/LICENSE",
"chars": 1064,
"preview": "MIT License\n\nCopyright (c) 2020 THU-PLM\n\nPermission is hereby granted, free of charge, to any person obtaining a copy\nof"
},
{
"path": "CPM-Generate/README.md",
"chars": 1640,
"preview": "# CPM-Generate\n\n为了促进中文自然语言处理研究的发展，本项目提供了 **CPM-LM** (2.6B) 模型的文本生成代码，可用于文本生成的本地测试，并以此为基础进一步研究零次学习/少次学习等场景。[[项目首页](https:"
},
{
"path": "CPM-Generate/arguments.py",
"chars": 18724,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/bpe_3w_new/chinese_vocab.vocab",
"chars": 375082,
"preview": "<unk>\t0\n<s>\t0\n</s>\t0\n<cls>\t0\n<sep>\t0\n<pad>\t0\n<mask>\t0\n<eod>\t0\n▁\t-1.72313\n,\t-3.12488\n▂\t-3.39822\n▁的\t-3.48588\n。\t-3.96942\n、\t"
},
{
"path": "CPM-Generate/bpe_3w_new/merges.txt",
"chars": 0,
"preview": ""
},
{
"path": "CPM-Generate/bpe_3w_new/vocab.json",
"chars": 715221,
"preview": "{\"<unk>\": 0, \"<s>\": 1, \"</s>\": 2, \"\\u2583\": 3, \"<sep>\": 4, \"<pad>\": 5, \"<mask>\": 6, \"<eod>\": 7, \"\\u2581\": 8, \",\": 9, \"\\u"
},
{
"path": "CPM-Generate/configure_data.py",
"chars": 9022,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data/Makefile",
"chars": 279,
"preview": "CXXFLAGS += -O3 -Wall -shared -std=c++11 -fPIC -fdiagnostics-color\nCPPFLAGS += $(shell python3 -m pybind11 --includes)\nL"
},
{
"path": "CPM-Generate/data/__init__.py",
"chars": 30,
"preview": "from . import indexed_dataset\n"
},
{
"path": "CPM-Generate/data/bert_dataset.py",
"chars": 11691,
"preview": "# coding=utf-8\n# Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data/dataset_utils.py",
"chars": 14864,
"preview": "# coding=utf-8\n# Copyright 2018 The Google AI Language Team Authors, and NVIDIA.\n#\n# Licensed under the Apache License, "
},
{
"path": "CPM-Generate/data/gpt2_dataset.py",
"chars": 13296,
"preview": "# coding=utf-8\n# Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data/helpers.cpp",
"chars": 14009,
"preview": "/*\n coding=utf-8\n Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.\n\n Licensed under the Apache License, Ver"
},
{
"path": "CPM-Generate/data/indexed_dataset.py",
"chars": 18532,
"preview": "# Copyright (c) Facebook, Inc. and its affiliates.\n#\n# This source code is licensed under the MIT license found in the\n#"
},
{
"path": "CPM-Generate/data/samplers.py",
"chars": 6055,
"preview": "# coding=utf-8\n# Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data/test/test_indexed_dataset.py",
"chars": 4619,
"preview": "# This file isn't really a formal automated test, it's just a place to\n# put some code used during development and manua"
},
{
"path": "CPM-Generate/data/test/test_preprocess_data.sh",
"chars": 241,
"preview": "#!/bin/bash\n\nIMPL=cached\npython ../preprocess_data.py \\\n --input test_samples.json \\\n --vocab vocab.txt \\\n "
},
{
"path": "CPM-Generate/data_utils/__init__.py",
"chars": 5354,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data_utils/corpora.py",
"chars": 2000,
"preview": "# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Version 2.0 (the "
},
{
"path": "CPM-Generate/data_utils/datasets.py",
"chars": 31718,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data_utils/file_utils.py",
"chars": 8441,
"preview": "# This file is provided as is from:\n# https://github.com/huggingface/pytorch-pretrained-BERT\n# Please refer to their r"
},
{
"path": "CPM-Generate/data_utils/lazy_loader.py",
"chars": 6895,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data_utils/samplers.py",
"chars": 5761,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data_utils/tf_dl.py",
"chars": 5443,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data_utils/tokenization.py",
"chars": 35987,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/data_utils/tokenization_gpt2.py",
"chars": 13640,
"preview": "# coding=utf-8\n# Copyright 2018 The Open AI Team Authors and The HuggingFace Inc. team.\n#\n# Licensed under the Apache Li"
},
{
"path": "CPM-Generate/data_utils/wordpiece.py",
"chars": 15599,
"preview": "# coding=utf-8\n# Copyright 2018 The Google AI Language Team Authors and The HuggingFace Inc. team.\n#\n# Licensed under th"
},
{
"path": "CPM-Generate/fp16/__init__.py",
"chars": 941,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/fp16/fp16.py",
"chars": 31717,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/fp16/fp16util.py",
"chars": 7681,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/fp16/loss_scaler.py",
"chars": 9776,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/generate_samples.py",
"chars": 13808,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/model/__init__.py",
"chars": 852,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/model/distributed.py",
"chars": 4895,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/model/gpt2_modeling.py",
"chars": 4524,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/model/model.py",
"chars": 3754,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/model/modeling.py",
"chars": 66897,
"preview": "# coding=utf-8\n# Copyright 2018 The Google AI Language Team Authors and The HugginFace Inc. team.\n# Copyright (c) 2018, "
},
{
"path": "CPM-Generate/mpu/__init__.py",
"chars": 2047,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/cross_entropy.py",
"chars": 4716,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/data.py",
"chars": 4018,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/grads.py",
"chars": 3015,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/initialize.py",
"chars": 4883,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/layers.py",
"chars": 13402,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/mappings.py",
"chars": 4136,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/random.py",
"chars": 14562,
"preview": "# coding=utf-8\n#Modified by Samyam Rajbhandari\n#Used to partition the activations stored for backward propagation\n#There"
},
{
"path": "CPM-Generate/mpu/tests/__init__.py",
"chars": 0,
"preview": ""
},
{
"path": "CPM-Generate/mpu/tests/commons.py",
"chars": 2619,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/tests/test_cross_entropy.py",
"chars": 3870,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/tests/test_data.py",
"chars": 2919,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/tests/test_initialize.py",
"chars": 3498,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/tests/test_layers.py",
"chars": 19272,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/tests/test_random.py",
"chars": 6698,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/transformer.py",
"chars": 26265,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/mpu/utils.py",
"chars": 2711,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "CPM-Generate/requirements.txt",
"chars": 121,
"preview": "torch==1.6.0\nnltk>=3.4\nnumpy>=1.15.4\npandas>=0.24.0\nboto3==1.11.11\nregex==2020.1.8\nsentencepiece\njieba\npybind11\nrequests"
},
{
"path": "CPM-Generate/scripts/generate_text.sh",
"chars": 746,
"preview": "#!/bin/bash\n\nCHECKPOINT_PATH=$1\nMPSIZE=2\nNLAYERS=32\nNHIDDEN=2560\nNATT=32\nMAXSEQLEN=1024\n\n#SAMPLING ARGS\nTEMP=0.9\n#If TOP"
},
{
"path": "CPM-Generate/utils.py",
"chars": 12160,
"preview": "# coding=utf-8\n# Copyright (c) 2019, NVIDIA CORPORATION. All rights reserved.\n#\n# Licensed under the Apache License, Ve"
},
{
"path": "ChineseAiDungeonColabDemo.ipynb",
"chars": 2670,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"markdown\",\n \"id\": \"2a45dcbe\",\n \"metadata\": {},\n \"source\": [\n \"# 中文ai 地牢 cola"
},
{
"path": "LICENSE",
"chars": 18092,
"preview": " GNU GENERAL PUBLIC LICENSE\n Version 2, June 1991\n\n Copyright (C) 1989, 1991 Fr"
},
{
"path": "finetune.ipynb",
"chars": 47903,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": 1,\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n "
},
{
"path": "labeled_data/advanture_translated/process_data.ipynb",
"chars": 28658,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": 2,\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n "
},
{
"path": "labeled_data/advanture_translated/processed_translated_story.txt",
"chars": 1463814,
"preview": "<start>\n故事名称：\n1\n故事背景：\n你在平静的浅水池中醒来，在风的呼啸和冰冷的爱抚中醒来。月亮从上面满是星星的天空注视着你。它苍白的灯光照亮了你脚下光滑石头缝隙之间生长的苔藓和茂密的草。树根弯曲，长满苔藓，伸向天空，在这个奇怪的无墙"
},
{
"path": "labeled_data/advanture_translated/processed_translated_story_valid.txt",
"chars": 105415,
"preview": "<start>\n故事名称：\n1\n故事背景：\n黑暗的时代，萨利拉的土地急需光明。一群被称为黑手的黑暗巫师毫不留情地统治着这片土地。他们是邪恶的本质。黑手夺取了政权，谋杀了全国议会，在首都塞隆德拉公开处决了朱诺大师，让所有人都能看到。只有一个组"
},
{
"path": "labeled_data/advanture_translated/train_adventures_valid_translated.txt",
"chars": 2712335,
"preview": "<|startoftext|>你在平静的浅水池中醒来，在风的呼啸和冰冷的爱抚中醒来。月亮从上面满是星星的天空注视着你。它苍白的灯光照亮了你脚下光滑石头缝隙之间生长的苔藓和茂密的草。树根弯曲，长满苔藓，伸向天空，在这个奇怪的无墙花园中可以看到"
},
{
"path": "labeled_data/advanture_translated/truncated_advanture_valid.json",
"chars": 1700016,
"preview": "[{\"previous\": [\"\\u4f60\\u8fdb\\u5165\\u6559\\u5802\", \"\\u5a01\\u5c14\\u5f17\\u91cc\\u5fb7\\u795e\\u7236\\u8dea\\u5728\\u6559\\u5802\\u76"
},
{
"path": "labeled_data/advanture_translated/valid_advantures_valid_translated.txt",
"chars": 191021,
"preview": "<| startoftext |>欢迎并感谢您玩Cryptode！\n在比赛开始前，你需要注意和了解很多事情。\n \n时间-\n生活-\n其他-\n \n这三样东西控制着游戏的进行方式。如果你在游戏中浪费时间，并且你的时间计数器达到0，你将输掉游戏。生"
},
{
"path": "labeled_data/raw_data/Jinyong/书剑恩仇录.txt",
"chars": 164154,
"preview": "\r\n\r\nӹ齣¼\r\nһء ŵھԾ ΣͿ콣ʶ\r\nǬ¡ʮ£ܱԺһʮŮıĵ鷿ʦˡͨϡ֮սһ飬ڽЩ褵Ĺ¡ûΣŮȴδҪʦٽ¡ʢﾲĵأûһ˿硣Ů鷿֮⣬ʦ˯δѣȥ㣬Ƶ⣬ͷϽΣڴֽϴ˸Сף۹ȥֻʦϥϣ¶ЦһɵһôڰһȥֻϷżʮֻӬһʮ֣ע"
},
{
"path": "labeled_data/raw_data/Jinyong/侠客行.txt",
"chars": 124691,
"preview": "www.wmtxt.comtxtС˵վ\r\nѵעwww.wmtxt.com\r\n\r\n\r\nӹ\r\nһ \r\nԿϺӧ˪ѩհǡ\r\nʮɱһˣǧﲻС˷ȥ\r\nйѽϥǰᡣ캥Ȱ\r\nȻŵΪᡣۻȺ\r\nԻӽǧ׳ʿۻبݺմǡ\r\n㣬Ӣ˭£̫\r\nһסСŷ磬дս"
},
{
"path": "labeled_data/raw_data/Jinyong/倚天屠龙记.txt",
"chars": 317888,
"preview": "ѵϼעŹںţӰСվʮָл~\r\n\r\nӹǡ\r\nһ ˼\r\nκƵ꺮ʳ滨ʱڡѩҹ¡˼ϣϼͨƹˣ㣬߽ࡣβ˭ŵȺͬСӢɲѷֱ̨ȥ췽\r\nһסʵģĩһλѧңе֮ʿųӣȫ֮һȫг͵Ʒ۴˴ʵ֮νҲ֮ˡ״е滨ʵȴһλµòŮ˵ƹˣ㣬߽ࡱ˵ӢɲȺͬСŮ˹"
},
{
"path": "labeled_data/raw_data/Jinyong/天龙八部.txt",
"chars": 401313,
"preview": "ѵϼעŹںţӰСվʮָл~\r\n\r\n˲\r\nһշ » ¸Զ ˭ӵ˭Ժ ƻڶ Х ̺\r\nĻ Ӵ ˮ ָȺϷ ǧж ƽʱ Ӣ ǧ \r\nǧãã ˫ ţԼ о ç̤ѩгܲ 굴 ľ Ӣ\r\nġɸӮɰ "
},
{
"path": "labeled_data/raw_data/Jinyong/射雕英雄传.txt",
"chars": 304760,
"preview": "www.wmtxt.comtxtС˵վ\r\nѵעwww.wmtxt.com\r\n\r\nӢ۴\r\nӹ\r\nһ ѩ\r\nǮƺƽˮҹҹĴٰţҴƹ뺣һʮڰҶƻհ죬ǰʱǰҰݸʼƣһĨбӳ֮£˼ΧһѴŮŮʮСԾ۾һ˵\r\n˵ʮͣһ಼ϴʳɫֻƬ滨ľ˼£һСɹõ\r\nС"
},
{
"path": "labeled_data/raw_data/Jinyong/白马啸西风.txt",
"chars": 25454,
"preview": "\r\n\r\nХ\r\nõõãõõá\r\nõõãõõá\r\nڻɳççĻؽĮ֮ϣɳߣһǰһļ\r\n۶ǰƥȳİٸ§߰\r\nСƥϷǸݵĺӡ\r\nǺ߱ȴһ֦Ѫϣ\r\n£˻ɳ֮Сְμֻ֦һͻ֧\r\nסʱС˭أҲûʲᡣ˭ǰĽŮ\r\nᣬĵڽ١\r\nµʮأѽƣû۵ı\r\n֮£Ƶ"
},
{
"path": "labeled_data/raw_data/Jinyong/碧血剑.txt",
"chars": 160548,
"preview": "\r\n\r\nӹѪ\r\nһء Σ \r\nʵδϺⲳǼˣͬӡܡáӼ㳼ԡ裡Ϭǡ¡ʵ۴ããš\r\nޱˣֳࡢˡԼӢ£ϵͬһ֮룩Ľлγ̫ƽ˹꣬յйʷΪDzʹԡ̴\r\nǼ˹ۼ쳯Ϲ︷ν̻¹ߣϲ̾Ȼȥʮһ£һϣˮΡΪϧΪ֮ꡳգϾ⣨Ͼл۱ɽ´ĹַػطأϮ"
},
{
"path": "labeled_data/raw_data/Jinyong/神雕侠侣.txt",
"chars": 324502,
"preview": "ѵϼעŹںţӰСվʮָл~=====================================================================¡Ŀ¼=================================="
},
{
"path": "labeled_data/raw_data/Jinyong/笑傲江湖.txt",
"chars": 314934,
"preview": "ѵϼעŹںţӰСվʮָл~\r\n\r\nӹЦ\r\nһ \r\nͷѬˣϹüڡʡݸŴ֣ʯ·ֱչȥֱͨšһΰլ֮ǰʯ̳иһߵˣ˶Ʈ졣ϻɫ˿һͷצ̬͵ʨչԵʨʨͷһԺ˿չ衣šھ֡ĸ֣վǷլţϲ豭Сͭ⣬ŶҶдšھ֡ĸ֣顰ܺšС֡ŴųʣŰװĺӣͦԳһӢ֮\r\nͻ"
},
{
"path": "labeled_data/raw_data/Jinyong/越女剑.txt",
"chars": 5953,
"preview": "\r\n\r\nԽ Ů \r\n ӹ\r\n룡룡\r\nʿԵת⣬սֱִ\r\nδվֱͻȻһ죬˫ཻ˸\r\nһԹ˶ǡסһ\r\n½ʿʿһһ½ʿһ߸ȣϽ\r\nֱ£ƾʿֽýݣԾܹ⽣\r\nأӸŵˢˢֹȥ½ʿﲻDZ\r\nЦڣ\r\nʿͻȻ㼲½ʿתԽԽ졣\r\nʿӵֳ⣬нһӽ"
},
{
"path": "labeled_data/raw_data/Jinyong/连城诀.txt",
"chars": 78875,
"preview": "\r\n\r\nǾ\r\nӹ\r\nһ ˽\r\nУУУУ\r\nľ轻ײ֮ʱöϢʱײ֮飬\r\nϽϪ£С֮ǰɹȳϣһŮֳľڱԡ\r\nǰһͷҧһ̶̵ĺ̹ܣڴЬż̧ͷŮһۣDZЦʾһ̣ͷ֮ϣľƳһʱ⾼ȻʵҲϣƺʮҲ\r\nŮʮ߰ͣԲԲһ˫ۺģʱ۵öͷһˮֱС˲Ϻ¹ŵ"
},
{
"path": "labeled_data/raw_data/Jinyong/雪山飞狐.txt",
"chars": 43511,
"preview": "www.wmtxt.comtxtС˵վ\r\nѵעwww.wmtxt.com\r\n\r\nѩɽɺ\r\nһ\r\n쬵һһ֦Ӷɽ˳죬գһͷ㾱Сڿд˼ﶷѩء\r\nʮ⣬̤ѩϳ˿üԼͬһƥ߷ʱԣһգʱֹȾҲþѵһԵðϿ£áۼм£жһɣҪǷǺε\r\n˰Σɽʼ˳ȴһ죬֮˾"
},
{
"path": "labeled_data/raw_data/Jinyong/飞狐外传.txt",
"chars": 143160,
"preview": "ѵϼעŹںţӰСվʮָл~\r\n\r\nɺ\r\nӹ\r\nһ¡̼ұ\r\nһأ࣡\r\n˷ز֣Ϲȣ\r\nһ˻ƵɤӵͳؽšгԹͷŭݷбųǧꡢ磬ÿһͿѪͳޡ\r\nͻͻͻͻ죬ĵ֦鷢ľơ\r\nÿľƵ淴涼һȫΣһϻǸŨ״ֺĴעһ֣һϻǸݳӣע˷֣Ѩľһֽýݵ׳һƣߡ"
},
{
"path": "labeled_data/raw_data/Jinyong/鸳鸯刀.txt",
"chars": 13342,
"preview": "\r\n\r\nĸװĺӲڵ·\r\nǺڵɽկǿˣֻĸĪ֮Ул\r\n´֣ǼСƺƴڶӣԶ֮Ψֲĸ\r\nģ·ѵָ֣Լ\r\nˣһ˶С°ͼһԶüִ̡ڶ\r\nַʣһڵ£ǰһʯдǡȿƸ\r\n֮ĹһĹˣ֪ǰк⣿Ƴϱû˵\r\nһλǰְеģƤһͻһ磬һ\r\nͷ˰磬һλòӢõһǴұߵ"
},
{
"path": "labeled_data/raw_data/Jinyong/鹿鼎记.txt",
"chars": 420668,
"preview": "\r\n\r\n¹\r\nһ ݺṳ µ\r\n絶ر˪\r\nϽһ·ϣһִǹѺðС\r\nǰзֱӣ磬һǰߣ\r\nˡŮӣһǸٸбŸŮӤŮӤ\r\nݡĸǣŮӤֻǴޡһˣڳһţȵ\r\nٿޣٿޣ!ŮӤһøˡ\r\n뿪·ʮɴݣվһʿһʮһСʿ\r\n龰̾һۿҲˣ˵!\r\nũС"
},
{
"path": "labeled_data/raw_data/邪气凛然.txt",
"chars": 1147636,
"preview": "аȻ\r\n\r\n\r\n ع鶼⣬ٴа漣\r\n\r\n<h6style=\"font-size:2em;line-height:350%;\">ݿΪգһС⣬\r\n\r\n <spanstyle=\"color:red\">ٱ</span>֪ǻᾡлл֧֣\r\n\r\n "
},
{
"path": "labeled_data/一次冒险.txt",
"chars": 1263,
"preview": "<start>\n故事名称：\n一次冒险\n故事背景：\n你是公务员，夏天快要过完的时候，你的妻子想去爬山\n--------------------\n动作：\n做\n去爬山\n结果：\n天上的太阳已不再那么炽热。你和妻子开车从长沙市中心出发，一路奔向黑麋峰"
},
{
"path": "labeled_data/我从来都不主动.txt",
"chars": 6646,
"preview": "<start>\n故事名称：\n我从来都不主动\n故事背景：\n你是许安阳，刚入学南京大学的大一新生\n--------------------\n动作：\n做\n给妈妈打电话\n结果：\n你先掏出手机，给老妈打了个电话，报了平安，说自己到南京了。 “怎"
},
{
"path": "labeled_data/盗墓实录.txt",
"chars": 591,
"preview": "<start>\n故事名称：\n盗墓实录\n故事背景：\n你是青格勒图，刚刚从监狱出狱\n--------------------\n动作：\n做\n走出监狱\n结果：\n你带着行李步出监狱正门，来不及回头说声感谢的话，高大的铁门左下方的小角门就“砰”的一声锁"
},
{
"path": "labeled_data/邪气凛然.txt",
"chars": 1852,
"preview": "<start>\n故事名称：\n邪气凛然\n故事背景：\n你叫陈阳，年纪二十三岁，生活在南方的城市里，是个夜总会主管。\n--------------------\n动作：\n做\n在夜总会走走\n结果：\n你走出休息室，走廊里还没有什么人，不过一些服务员已经"
},
{
"path": "readme.md",
"chars": 7843,
"preview": "<pre>\n ▄████▄ ██░ ██ ██▓ ███▄ █ ▓█████ ██████ ▓█████ \n ▒██▀ ▀█ ▓██░ ██▒▓██▒ ██ ▀█ █ ▓█ ▀ ▒██ ▒"
},
{
"path": "requirements.txt",
"chars": 96,
"preview": "sentencepiece==0.1.96\njieba==0.42.1\nregex==2020.1.8\ntensorflow-gpu==2.3.0\ntensorflow-hub==0.12.0"
},
{
"path": "tf2gpt/.gitignore",
"chars": 55,
"preview": "*.py[cod]\ngpt_model_tf2\nmodel.ckpt-*\n.ipynb_checkpoints"
},
{
"path": "tf2gpt/README.md",
"chars": 1414,
"preview": "# TensorFlow 2.0 GPT\n\nStructure is as same as [imcaspar/gpt2-ml](https://github.com/imcaspar/gpt2-ml), not official work"
},
{
"path": "tf2gpt/loading.ipynb",
"chars": 19620,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": 1,\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n "
},
{
"path": "tf2gpt/model.py",
"chars": 11366,
"preview": "\"\"\"\nPaper:\n- https://d4mucfpksywv.cloudfront.net/better-language-models/language-models.pdf\n\nOfficial repo:\n- https://gi"
},
{
"path": "tf2gpt/predict.ipynb",
"chars": 2732,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": 1,\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n "
},
{
"path": "tf2gpt/test.ipynb",
"chars": 4605,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": 1,\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n "
},
{
"path": "tf2gpt/train.ipynb",
"chars": 4896,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": null,\n \"metadata\": {},\n \"outputs\": [],\n \"source\": "
},
{
"path": "tf2gpt/vocab.txt",
"chars": 75770,
"preview": "[PAD]\n[unused1]\n[unused2]\n[unused3]\n[unused4]\n[unused5]\n[unused6]\n[unused7]\n[unused8]\n[unused9]\n[unused10]\n[unused11]\n[u"
},
{
"path": "utils/gpt2_tokenizer.py",
"chars": 13738,
"preview": "# coding=utf-8\n# Copyright 2018 The Open AI Team Authors and The HuggingFace Inc. team.\n#\n# Licensed under the Apache Li"
},
{
"path": "utils/progress_bar.py",
"chars": 995,
"preview": "# define dataset class to feed the model\nimport numpy as np \nimport os\nimport sys\nimport time\n\nclass ProgressBar():\n "
},
{
"path": "utils/story_helper.py",
"chars": 2949,
"preview": "import re\nimport numpy as np\nimport tensorflow as tf\nimport os\nimport sys\nfrom tf2gpt.model import GPT\nfrom utils.story_"
},
{
"path": "utils/story_util.py",
"chars": 3589,
"preview": "class Story():\n def __init__(self,background,title):\n self.background = background.replace(\"\\n\",\"\")\n se"
},
{
"path": "标注.ipynb",
"chars": 9177,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"code\",\n \"execution_count\": 1,\n \"metadata\": {},\n \"outputs\": [],\n \"source\": [\n "
}
]
// ... and 3 more files (download for full content)
About this extraction
This page contains the full source code of the bupticybee/ChineseAiDungeon GitHub repository, extracted and formatted as plain text for AI agents and large language models (LLMs). The extraction includes 108 files (78.0 MB), approximately 3.0M tokens, and a symbol index with 778 extracted functions, classes, methods, constants, and types. Use this with OpenClaw, Claude, ChatGPT, Cursor, Windsurf, or any other AI tool that accepts text input. You can copy the full output to your clipboard or download it as a .txt file.
Extracted by GitExtract — free GitHub repo to text converter for AI. Built by Nikandr Surkov.