SYMBOL INDEX (546 symbols across 55 files)

FILE: 4G_memory_rag/langchain_demo.py
  function clean_text (line 68) | def clean_text(text):
  class MiniCPM_LLM (line 90) | class MiniCPM_LLM(LLM):
    method __init__ (line 94) | def __init__(self, model_path: str):
    method _call (line 121) | def _call(self, prompt, stop: Optional[List[str]] = None):
    method _llm_type (line 178) | def _llm_type(self) -> str:
  function load_documents (line 183) | def load_documents(file_paths):
  function load_models (line 213) | def load_models():
  function embed_documents (line 237) | def embed_documents(documents, embedding_models):
  function create_prompt_template (line 256) | def create_prompt_template():
  function create_rag_chain (line 276) | def create_rag_chain(llm, prompt):
  function analysis_links (line 282) | def analysis_links(docs):
  function main (line 309) | def main():
  function process_query (line 337) | def process_query(file, query):

FILE: MiniCPM-o-long_video_inference/infer.py
  function extract_frames_and_audio (line 22) | def extract_frames_and_audio(video_path, sample_fps=2, max_frames=None, ...
  class LongVideoAudioProcessor (line 72) | class LongVideoAudioProcessor:
    method __init__ (line 73) | def __init__(self,
    method preprocess_frame (line 117) | def preprocess_frame(self, frame):
    method update_memory_bank (line 133) | def update_memory_bank(self, frames, audio_segments, current_time):
    method calculate_time_weights (line 153) | def calculate_time_weights(self, current_time):
    method weighted_sampling (line 164) | def weighted_sampling(self, frames, weights, sample_count):
    method update_text_summary (line 172) | def update_text_summary(self, new_summary):
    method process_long_video (line 179) | def process_long_video(self, video_path, query):
    method _inference_with_memory (line 224) | def _inference_with_memory(self, frames, audio_segments, query, chunk_...
    method _merge_results (line 339) | def _merge_results(self, all_results, query):
  class AudioProcessor (line 396) | class AudioProcessor:
    method __init__ (line 397) | def __init__(self, sample_rate=16000):
    method extract_audio_from_video (line 400) | def extract_audio_from_video(self, video_path):
    method segment_audio (line 409) | def segment_audio(self, audio, duration, segment_duration=1.0):
    method extract_audio_features (line 426) | def extract_audio_features(self, audio_segment):
  function main (line 432) | def main():

FILE: MiniCPMV2_6_awq/modeling_minicpmv.py
  class MiniCPMVPreTrainedModel (line 17) | class MiniCPMVPreTrainedModel(Qwen2PreTrainedModel):
  class MiniCPMV (line 21) | class MiniCPMV(MiniCPMVPreTrainedModel):
    method __init__ (line 22) | def __init__(self, config):
    method init_vision_module (line 33) | def init_vision_module(self):
    method init_resampler (line 49) | def init_resampler(self, embed_dim, vision_dim):
    method get_input_embeddings (line 58) | def get_input_embeddings(self):
    method set_input_embeddings (line 61) | def set_input_embeddings(self, value):
    method get_output_embeddings (line 64) | def get_output_embeddings(self):
    method set_output_embeddings (line 67) | def set_output_embeddings(self, new_embeddings):
    method set_decoder (line 70) | def set_decoder(self, decoder):
    method prepare_inputs_for_generation (line 73) | def prepare_inputs_for_generation(
    method get_decoder (line 121) | def get_decoder(self):
    method get_vllm_embedding (line 124) | def get_vllm_embedding(self, data):
    method forward (line 217) | def forward(self, data, **kwargs):
    method _decode (line 236) | def _decode(self, inputs_embeds, tokenizer, attention_mask, decode_tex...
    method _decode_stream (line 249) | def _decode_stream(self, inputs_embeds, tokenizer, **kwargs):
    method _decode_text (line 265) | def _decode_text(self, result_ids, tokenizer):
    method generate (line 277) | def generate(
    method chat (line 321) | def chat(

FILE: MiniCPMV2_6_awq/quantize.py
  function copy_files_not_in_B (line 18) | def copy_files_not_in_B(A_path, B_path):
  function load_alpaca (line 46) | def load_alpaca():
  function load_wikitext (line 58) | def load_wikitext():

FILE: OCR_Multimodal_Search/finetune/dataset.py
  class SupervisedDataset (line 43) | class SupervisedDataset(Dataset):
    method __init__ (line 46) | def __init__(
    method __len__ (line 67) | def __len__(self):
    method __getitem__ (line 70) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  function data_collator (line 107) | def data_collator(examples, padding_value=0, max_length=2048):
  function conversation_to_ids (line 150) | def conversation_to_ids(conversation, tokenizer, llm_type=None, new_sche...
  function conversation_to_ids_minicpm (line 215) | def conversation_to_ids_minicpm(conversation, tokenizer):
  function conversation_to_ids_llama3 (line 247) | def conversation_to_ids_llama3(conversation, tokenizer):
  function conversation_to_ids_qwen2 (line 287) | def conversation_to_ids_qwen2(conversation, tokenizer):
  function preprocess (line 327) | def preprocess(
  function slice_image (line 414) | def slice_image(
  function ensure_divide (line 473) | def ensure_divide(length, patch_size):
  function find_best_resize (line 477) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 488) | def get_refine_size(
  function split_to_patches (line 512) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 529) | def get_grid_placeholder(tokenizer, grid, query_num, new_schema=False):
  function reshape_by_patch (line 550) | def reshape_by_patch(image_tensor, patch_size):

FILE: OCR_Multimodal_Search/finetune/dataset_original.py
  class SupervisedDataset (line 18) | class SupervisedDataset(Dataset):
    method __init__ (line 21) | def __init__(
    method __len__ (line 42) | def __len__(self):
    method __getitem__ (line 45) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  function data_collator (line 70) | def data_collator(examples, padding_value=0, max_length=2048):
  function conversation_to_ids (line 108) | def conversation_to_ids(conversation, tokenizer, llm_type=None, new_sche...
  function conversation_to_ids_minicpm (line 173) | def conversation_to_ids_minicpm(conversation, tokenizer):
  function conversation_to_ids_llama3 (line 205) | def conversation_to_ids_llama3(conversation, tokenizer):
  function conversation_to_ids_qwen2 (line 245) | def conversation_to_ids_qwen2(conversation, tokenizer):
  function preprocess (line 285) | def preprocess(
  function slice_image (line 372) | def slice_image(
  function ensure_divide (line 431) | def ensure_divide(length, patch_size):
  function find_best_resize (line 435) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 446) | def get_refine_size(
  function split_to_patches (line 470) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 487) | def get_grid_placeholder(tokenizer, grid, query_num, new_schema=False):
  function reshape_by_patch (line 508) | def reshape_by_patch(image_tensor, patch_size):

FILE: OCR_Multimodal_Search/finetune/finetune.py
  class ModelArguments (line 30) | class ModelArguments:
  class DataArguments (line 35) | class DataArguments:
  class TrainingArguments (line 45) | class TrainingArguments(transformers.TrainingArguments):
  class LoraArguments (line 66) | class LoraArguments:
  function rank0_print (line 80) | def rank0_print(*args):
  function safe_save_model_for_hf_trainer (line 85) | def safe_save_model_for_hf_trainer(trainer, output_dir: str, bias="none"):
  function make_supervised_data_module (line 91) | def make_supervised_data_module(
  function build_transform (line 142) | def build_transform():
  function get_parameter_number (line 154) | def get_parameter_number(model):
  function train (line 172) | def train():

FILE: OCR_Multimodal_Search/finetune/trainer.py
  class CPMTrainer (line 41) | class CPMTrainer(Trainer):
    method original_loss (line 42) | def original_loss(self, model, inputs, return_outputs=False):
    method coloss (line 73) | def coloss(self, query_embeddings, doc_embeddings):
    method compute_loss (line 104) | def compute_loss(self,model, inputs):
    method prediction_step (line 165) | def prediction_step(
    method training_step (line 273) | def training_step(self, model: nn.Module, inputs: Dict[str, Union[torc...
    method _save (line 315) | def _save(self, output_dir: Optional[str] = None, state_dict=None):

FILE: OCR_Multimodal_Search/infer/app.py
  function search (line 17) | def search(query: str, ds, images):
  function index (line 39) | def index(file, ds):

FILE: OCR_Multimodal_Search/infer/cli_demo.py
  function search (line 15) | def search(query: str, ds, images):
  function index (line 38) | def index(file):

FILE: OCR_Multimodal_Search/infer/dataset.py
  class ImageDataset (line 19) | class ImageDataset(Dataset):
    method __init__ (line 21) | def __init__(
    method __len__ (line 42) | def __len__(self):
    method __getitem__ (line 45) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  class QueryDataset (line 72) | class QueryDataset(Dataset):
    method __init__ (line 75) | def __init__(
    method __len__ (line 86) | def __len__(self):
    method __getitem__ (line 89) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  function data_collator_query (line 103) | def data_collator_query(examples, padding_value=0, max_length=2048):
  function data_collator (line 113) | def data_collator(examples, padding_value=0, max_length=2048,device='cpu'):
  function load_from_pdf (line 150) | def load_from_pdf(pdf_path: str):
  function load_from_json (line 155) | def load_from_json(json_path: str):
  function conversation_to_ids (line 161) | def conversation_to_ids(conversation, tokenizer, llm_type=None, new_sche...
  function conversation_to_ids_minicpm (line 226) | def conversation_to_ids_minicpm(conversation, tokenizer):
  function conversation_to_ids_llama3 (line 258) | def conversation_to_ids_llama3(conversation, tokenizer):
  function conversation_to_ids_qwen2 (line 298) | def conversation_to_ids_qwen2(conversation, tokenizer):
  function preprocess (line 338) | def preprocess(
  function slice_image (line 434) | def slice_image(
  function ensure_divide (line 493) | def ensure_divide(length, patch_size):
  function find_best_resize (line 497) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 508) | def get_refine_size(
  function split_to_patches (line 532) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 549) | def get_grid_placeholder(tokenizer, grid, query_num, new_schema=False):
  function reshape_by_patch (line 570) | def reshape_by_patch(image_tensor, patch_size):

FILE: OCR_Multimodal_Search/infer/inference.py
  function main (line 12) | def main() -> None:

FILE: OCR_Multimodal_Search/infer/utils.py
  function evaluate_colbert (line 4) | def evaluate_colbert( qs, ps, batch_size=128) -> torch.Tensor:
  function build_transform (line 22) | def build_transform():

FILE: OCR_VG/chat.py
  function init_omni_lmm (line 21) | def init_omni_lmm(model_path):
  function expand_question_into_multimodal (line 61) | def expand_question_into_multimodal(question_text, image_token_len, im_s...
  function wrap_question_for_omni_lmm (line 70) | def wrap_question_for_omni_lmm(question, image_token_len, tokenizer):
  class OmniLMM12B (line 85) | class OmniLMM12B:
    method __init__ (line 86) | def __init__(self, model_path) -> None:
    method decode (line 94) | def decode(self, image, input_ids):
    method chat (line 115) | def chat(self, input):
  function img2base64 (line 133) | def img2base64(file_name):
  class MiniCPMV (line 138) | class MiniCPMV:
    method __init__ (line 139) | def __init__(self, model_path) -> None:
    method chat (line 144) | def chat(self, input):
  class MiniCPMV2_5 (line 162) | class MiniCPMV2_5:
    method __init__ (line 163) | def __init__(self, model_path) -> None:
    method chat (line 168) | def chat(self, input):
  class MiniCPMVChat (line 186) | class MiniCPMVChat:
    method __init__ (line 187) | def __init__(self, model_path) -> None:
    method chat (line 195) | def chat(self, input):

FILE: OCR_VG/gt_test.py
  function parse_text (line 11) | def parse_text(input_str):
  function cv2ImgAddText (line 25) | def cv2ImgAddText(img, text, left, top, textColor=(0, 255, 0), textSize=...
  function draw (line 38) | def draw(final_box,final_text,img,height,width):
  function get_pic_path (line 56) | def get_pic_path(dir):

FILE: OCR_VG/merge_box.py
  function cv2ImgAddText (line 15) | def cv2ImgAddText(img, text, left, top, textColor=(0, 255, 0), textSize=...
  function calculate_intersection_percentage (line 31) | def calculate_intersection_percentage(rect1, rect2):
  function get_threshold (line 70) | def get_threshold(boxes, text_list, width, height):
  function find_min_bounding_rectangle (line 101) | def find_min_bounding_rectangle(polygons,four_pint=False):
  function merge_boxes (line 133) | def merge_boxes(boxes, text_list, width, height):
  function calculate_angle (line 184) | def calculate_angle(v1, v2):
  function is_quadrilateral_angles_in_range (line 193) | def is_quadrilateral_angles_in_range(points):
  function is_rectangle (line 204) | def is_rectangle(points):
  function get_query_answer (line 221) | def get_query_answer(boxes, text_list, height, width):
  function load_json (line 262) | def load_json(file_path):
  function draw (line 267) | def draw(final_box,final_text,img):
  function save_to_json (line 288) | def save_to_json(output_data, output_path):
  function main (line 300) | def main():

FILE: OCR_VG/omnilmm/conversation.py
  class SeparatorStyle (line 6) | class SeparatorStyle(Enum):
  class Conversation (line 13) | class Conversation:
    method get_prompt (line 26) | def get_prompt(self):
    method append_message (line 51) | def append_message(self, role, message):
    method get_images (line 54) | def get_images(self, return_pil=False):
    method to_gradio_chatbot (line 110) | def to_gradio_chatbot(self):
    method copy (line 142) | def copy(self):
    method dict (line 152) | def dict(self):

FILE: OCR_VG/omnilmm/model/omnilmm.py
  class OmniLMMConfig (line 23) | class OmniLMMConfig(MistralConfig):
  class Identity (line 27) | class Identity(torch.nn.Identity):
    method forward (line 28) | def forward(self, input: Tensor, **kwargs) -> Tensor:
  function create_vision_module (line 32) | def create_vision_module(config):
  class OmniLMMModel (line 56) | class OmniLMMModel(MistralModel):
    method __init__ (line 59) | def __init__(self, config: OmniLMMConfig, mm_vision_tower=None, mm_hid...
    method initialize_vision_modules (line 75) | def initialize_vision_modules(self, vision_tower, no_randaug, num_quer...
    method get_vision_embedding (line 108) | def get_vision_embedding(self, pixel_values):
    method get_vllm_embedding (line 123) | def get_vllm_embedding(self, data):
    method forward (line 184) | def forward(
  class OmniLMMForCausalLM (line 269) | class OmniLMMForCausalLM(MistralForCausalLM):
    method __init__ (line 272) | def __init__(self, config, mm_vision_tower=None, tune_clip=True):
    method forward (line 283) | def forward(
    method prepare_inputs_for_generation (line 350) | def prepare_inputs_for_generation(
    method generate_vllm (line 372) | def generate_vllm(
    method initialize_vision_tokenizer (line 400) | def initialize_vision_tokenizer(self, mm_use_im_start_end, tokenizer, ...

FILE: OCR_VG/omnilmm/model/resampler.py
  function get_abs_pos (line 23) | def get_abs_pos(abs_pos, tgt_size):
  function get_2d_sincos_pos_embed (line 43) | def get_2d_sincos_pos_embed(embed_dim, grid_size, cls_token=False):
  function get_2d_sincos_pos_embed_from_grid (line 62) | def get_2d_sincos_pos_embed_from_grid(embed_dim, grid):
  function get_1d_sincos_pos_embed_from_grid (line 75) | def get_1d_sincos_pos_embed_from_grid(embed_dim, pos):
  class Resampler (line 96) | class Resampler(nn.Module):
    method __init__ (line 104) | def __init__(
    method _init_weights (line 140) | def _init_weights(self, m):
    method forward (line 149) | def forward(self, x, attn_mask=None):
    method _repeat (line 170) | def _repeat(self, query, N: int):

FILE: OCR_VG/omnilmm/model/utils.py
  function auto_upgrade (line 23) | def auto_upgrade(config):
  class KeywordsStoppingCriteria (line 42) | class KeywordsStoppingCriteria(StoppingCriteria):
    method __init__ (line 43) | def __init__(self, keywords, tokenizer, input_ids):
    method __call__ (line 49) | def __call__(self, output_ids: torch.LongTensor, scores: torch.FloatTe...
  function auto_upgrade (line 61) | def auto_upgrade(config):
  function identity_func (line 82) | def identity_func(img):
  function autocontrast_func (line 86) | def autocontrast_func(img, cutoff=0):
  function equalize_func (line 118) | def equalize_func(img):
  function rotate_func (line 142) | def rotate_func(img, degree, fill=(0, 0, 0)):
  function solarize_func (line 153) | def solarize_func(img, thresh=128):
  function color_func (line 163) | def color_func(img, factor):
  function contrast_func (line 185) | def contrast_func(img, factor):
  function brightness_func (line 198) | def brightness_func(img, factor):
  function sharpness_func (line 208) | def sharpness_func(img, factor):
  function shear_x_func (line 230) | def shear_x_func(img, factor, fill=(0, 0, 0)):
  function translate_x_func (line 238) | def translate_x_func(img, offset, fill=(0, 0, 0)):
  function translate_y_func (line 249) | def translate_y_func(img, offset, fill=(0, 0, 0)):
  function posterize_func (line 260) | def posterize_func(img, bits):
  function shear_y_func (line 268) | def shear_y_func(img, factor, fill=(0, 0, 0)):
  function cutout_func (line 276) | def cutout_func(img, pad_size, replace=(0, 0, 0)):
  function enhance_level_to_args (line 290) | def enhance_level_to_args(MAX_LEVEL):
  function shear_level_to_args (line 296) | def shear_level_to_args(MAX_LEVEL, replace_value):
  function translate_level_to_args (line 306) | def translate_level_to_args(translate_const, MAX_LEVEL, replace_value):
  function cutout_level_to_args (line 316) | def cutout_level_to_args(cutout_const, MAX_LEVEL, replace_value):
  function solarize_level_to_args (line 324) | def solarize_level_to_args(MAX_LEVEL):
  function none_level_to_args (line 331) | def none_level_to_args(level):
  function posterize_level_to_args (line 335) | def posterize_level_to_args(MAX_LEVEL):
  function rotate_level_to_args (line 342) | def rotate_level_to_args(MAX_LEVEL, replace_value):
  class RandomAugment (line 394) | class RandomAugment(object):
    method __init__ (line 396) | def __init__(self, N=2, M=10, isPIL=False, augs=[]):
    method get_random_ops (line 405) | def get_random_ops(self):
    method __call__ (line 409) | def __call__(self, img):
  function build_transform (line 421) | def build_transform(is_train, randaug=True, input_size=224, interpolatio...
  function img2b64 (line 465) | def img2b64(img_path):
  function str2b64 (line 475) | def str2b64(str):
  function b642str (line 479) | def b642str(b64):
  function is_dist_avail_and_initialized (line 483) | def is_dist_avail_and_initialized():
  function get_world_size (line 491) | def get_world_size():
  function get_rank (line 497) | def get_rank():
  function all_gather (line 503) | def all_gather(data):
  function mean (line 546) | def mean(lst):
  function stop_gradient_by_name (line 550) | def stop_gradient_by_name(name: str):

FILE: OCR_VG/omnilmm/train/train_utils.py
  function _tokenize_fn (line 22) | def _tokenize_fn(strings: Sequence[str],
  function omni_preprocess (line 50) | def omni_preprocess(sources,

FILE: OCR_VG/omnilmm/utils.py
  function build_logger (line 17) | def build_logger(logger_name, logger_filename):
  class StreamToLogger (line 60) | class StreamToLogger(object):
    method __init__ (line 65) | def __init__(self, logger, log_level=logging.INFO):
    method __getattr__ (line 71) | def __getattr__(self, attr):
    method write (line 74) | def write(self, buf):
    method flush (line 88) | def flush(self):
  function disable_torch_init (line 94) | def disable_torch_init():
  function violates_moderation (line 103) | def violates_moderation(text):
  function pretty_print_semaphore (line 124) | def pretty_print_semaphore(semaphore):

FILE: agent_auto_plan/autoplan/all_param_inference.py
  function format_outputs (line 12) | def format_outputs(outputs):
  function all_param_split_task (line 22) | def all_param_split_task(question,tokenizer,merge_model):
  function stopping_criteria (line 63) | def stopping_criteria(cur_len,output_so_far):

FILE: agent_auto_plan/autoplan/bing_search.py
  class SearchResult (line 24) | class SearchResult:
    method __init__ (line 25) | def __init__(self, title, url, snip) -> None:
    method dump (line 30) | def dump(self):
    method __str__ (line 37) | def __str__(self) -> str:
  class SearcherInterface (line 40) | class SearcherInterface:
    method search (line 41) | def search(self, query) -> List[SearchResult]:
  function generate_document (line 45) | def generate_document(url):
  function summarize_document (line 55) | def summarize_document(url, model_name):
  function fetch_webpage_content (line 67) | def fetch_webpage_content(url):
  function summarize_text (line 72) | def summarize_text(text,question,model,tokenizer):
  class Searcher (line 95) | class Searcher(SearcherInterface):
    method __init__ (line 96) | def __init__(self) -> None:
    method _parse (line 99) | def _parse(self, result) -> List[SearchResult]:
    method search (line 107) | def search(self, query) -> List[SearchResult]:
  function get_bing_search_raw_page (line 111) | def get_bing_search_raw_page(question: str):
  function query_bing (line 143) | def query_bing(question, max_tries=3,model=None,tokenizer=None):

FILE: agent_auto_plan/autoplan/fuctions.py
  function gpt_35_api_stream (line 11) | def gpt_35_api_stream(messages: list):
  function bm25 (line 43) | def bm25(query, corpus,model=None):
  function get_check_text (line 65) | def get_check_text(sub_task_list):
  function get_tools_description (line 70) | def get_tools_description(tools):
  function get_task_and_question (line 80) | def get_task_and_question(path):
  function task_text_split (line 102) | def task_text_split(text,question_orgin):#将任务分解后的输出变成任务列表，这里是用了qwen第一次th...
  function distance (line 115) | def distance(query,map_dict):#计算距离

FILE: agent_auto_plan/autoplan/load_model.py
  function get_model (line 6) | def get_model(args):

FILE: agent_auto_plan/autoplan/lora_inference_nomerge.py
  function get_merge_model (line 9) | def get_merge_model(base_path,adapter_path):
  function split_task (line 27) | def split_task(question,tokenizer,merge_model,prompt=None):
  function format_outputs (line 51) | def format_outputs(outputs):#将输出格式化
  function test (line 61) | def test():
  function stopping_criteria (line 94) | def stopping_criteria(cur_len,output_so_far):

FILE: agent_auto_plan/autoplan/main.py
  function _get_args (line 72) | def _get_args():
  function llm_with_plugin (line 126) | def llm_with_plugin(
  function build_input_text (line 315) | def build_input_text(chat_history, list_of_plugin_info) -> str:
  function text_completion (line 370) | def text_completion(input_text: str, stop_words) -> str:  # 作为一个文本续写模型来使用
  function parse_latest_plugin_call (line 394) | def parse_latest_plugin_call(text):
  function test (line 426) | def test():

FILE: agent_auto_plan/autoplan/prompt_plamte.py
  function check_action_inputs (line 97) | def check_action_inputs(question,action,list_of_plugin_info,history=None...
  function prompt_task_split (line 144) | def prompt_task_split(question,tools,args,write_file):#任务分解函数

FILE: agent_auto_plan/autoplan/tools_introduction.py
  function call_plugin (line 199) | def call_plugin(plugin_name: str, plugin_args: str,write_file,embeding_m...

FILE: agent_auto_plan/finetune_language/dataset.py
  class SupervisedDataset (line 23) | class SupervisedDataset(Dataset):
    method __init__ (line 26) | def __init__(
    method __len__ (line 49) | def __len__(self):
    method __getitem__ (line 52) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  function data_collator (line 94) | def data_collator(examples, padding_value=0, max_length=2048):
  function conversation_to_ids (line 132) | def conversation_to_ids(conversation, tokenizer, llm_type=None, new_sche...
  function conversation_to_ids_minicpm (line 206) | def conversation_to_ids_minicpm(conversation, tokenizer):
  function conversation_to_ids_llama3 (line 238) | def conversation_to_ids_llama3(conversation, tokenizer):
  function conversation_to_ids_qwen2 (line 278) | def conversation_to_ids_qwen2(conversation, tokenizer):
  function preprocess (line 317) | def preprocess(
  function slice_image (line 418) | def slice_image(
  function ensure_divide (line 477) | def ensure_divide(length, patch_size):
  function find_best_resize (line 481) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 492) | def get_refine_size(
  function split_to_patches (line 516) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 533) | def get_grid_placeholder(tokenizer, grid, query_num, new_schema=False):
  function reshape_by_patch (line 559) | def reshape_by_patch(image_tensor, patch_size):

FILE: agent_auto_plan/finetune_language/finetune.py
  class ModelArguments (line 27) | class ModelArguments:
  class DataArguments (line 32) | class DataArguments:
  class TrainingArguments (line 42) | class TrainingArguments(transformers.TrainingArguments):
  class LoraArguments (line 59) | class LoraArguments:
  function rank0_print (line 73) | def rank0_print(*args):
  function safe_save_model_for_hf_trainer (line 78) | def safe_save_model_for_hf_trainer(trainer, output_dir: str, bias="none"):
  function make_supervised_data_module (line 84) | def make_supervised_data_module(
  function build_transform (line 137) | def build_transform():
  function get_parameter_number (line 149) | def get_parameter_number(model):
  function train (line 167) | def train():

FILE: agent_auto_plan/finetune_language/merge_lora.py
  function copy_files_not_in_B (line 11) | def copy_files_not_in_B(A_path, B_path):

FILE: agent_auto_plan/finetune_language/replace_file/modeling_minicpmv.py
  class MiniCPMVPreTrainedModel (line 18) | class MiniCPMVPreTrainedModel(Qwen2PreTrainedModel):
  class MiniCPMV (line 22) | class MiniCPMV(MiniCPMVPreTrainedModel):
    method __init__ (line 23) | def __init__(self, config):
    method init_vision_module (line 37) | def init_vision_module(self):
    method init_resampler (line 53) | def init_resampler(self, embed_dim, vision_dim):
    method get_input_embeddings (line 62) | def get_input_embeddings(self):
    method set_input_embeddings (line 65) | def set_input_embeddings(self, value):
    method get_output_embeddings (line 68) | def get_output_embeddings(self):
    method set_output_embeddings (line 71) | def set_output_embeddings(self, new_embeddings):
    method set_decoder (line 74) | def set_decoder(self, decoder):
    method prepare_inputs_for_generation (line 77) | def prepare_inputs_for_generation(
    method get_decoder (line 125) | def get_decoder(self):
    method get_vllm_embedding (line 128) | def get_vllm_embedding(self, data):
    method forward (line 236) | def forward(self, data, **kwargs):
    method _decode (line 257) | def _decode(self, inputs_embeds, tokenizer, attention_mask, decode_tex...
    method _decode_stream (line 270) | def _decode_stream(self, inputs_embeds, tokenizer, **kwargs):
    method _decode_text (line 286) | def _decode_text(self, result_ids, tokenizer):
    method generate (line 298) | def generate(
    method chat (line 342) | def chat(

FILE: agent_auto_plan/qwen_vllm.py
  function save_to_jsonl (line 17) | def save_to_jsonl(data, output_file):
  function save_to_excel (line 21) | def save_to_excel(data, output_file):
  function process_prompt (line 30) | def process_prompt(prompt):

FILE: agent_demo/agent_demo.py
  class SequenceStoppingCriteria (line 29) | class SequenceStoppingCriteria(StoppingCriteria):
    method __init__ (line 30) | def __init__(self, sequence_ids):
    method check_sequences (line 33) | def check_sequences(self, current_tokens, sequences):
    method __call__ (line 46) | def __call__(self, input_ids, scores, **kwargs):
  function llm_with_plugin (line 114) | def llm_with_plugin(prompt: str, history, list_of_plugin_info=()):
  function build_input_text (line 141) | def build_input_text(chat_history, list_of_plugin_info) -> str:
  function text_completion (line 186) | def text_completion(input_text: str, stop_words) -> str:  # 作为一个文本续写模型来使用
  function parse_latest_plugin_call (line 210) | def parse_latest_plugin_call(text):
  function call_plugin (line 235) | def call_plugin(plugin_name: str, plugin_args: str) -> str:
  function test (line 310) | def test():

FILE: agent_demo/build_react_prompt.py
  function build_input_text (line 24) | def build_input_text(chat_history, list_of_plugin_info) -> str:
  function parse_latest_plugin_call (line 71) | def parse_latest_plugin_call(text):

FILE: agent_demo/get_react_data.py
  function save_cpm3_data (line 78) | def save_cpm3_data(cpm3_data_path,cpm3_data):
  function switch_cpm_tool (line 86) | def switch_cpm_tool(tools):
  function function_call (line 126) | def function_call(plugin_name, plugin_args):
  function split_react_data (line 152) | def split_react_data(react_str):
  function get_answer_from_output (line 170) | def get_answer_from_output(output):
  function get_tool_description (line 177) | def get_tool_description(tool):
  function get_question (line 191) | def get_question():
  function get_react_data (line 220) | def get_react_data():
  function get_cpm_function_call (line 259) | def get_cpm_function_call():

FILE: ft_language_replace_file/finetune/dataset.py
  class SupervisedDataset (line 23) | class SupervisedDataset(Dataset):
    method __init__ (line 26) | def __init__(
    method __len__ (line 47) | def __len__(self):
    method __getitem__ (line 50) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  function data_collator (line 87) | def data_collator(examples, padding_value=0, max_length=2048):
  function conversation_to_ids (line 125) | def conversation_to_ids(conversation, tokenizer, llm_type=None):
  function conversation_to_ids_minicpm (line 178) | def conversation_to_ids_minicpm(conversation, tokenizer):
  function conversation_to_ids_llama3 (line 210) | def conversation_to_ids_llama3(conversation, tokenizer):
  function preprocess (line 252) | def preprocess(
  function slice_image (line 352) | def slice_image(
  function ensure_divide (line 411) | def ensure_divide(length, patch_size):
  function find_best_resize (line 415) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 426) | def get_refine_size(
  function split_to_patches (line 450) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 467) | def get_grid_placeholder(tokenizer, grid, query_num):
  function reshape_by_patch (line 485) | def reshape_by_patch(image_tensor, patch_size):

FILE: ft_language_replace_file/finetune/finetune.py
  class ModelArguments (line 44) | class ModelArguments:
  class DataArguments (line 49) | class DataArguments:
  class TrainingArguments (line 59) | class TrainingArguments(transformers.TrainingArguments):
  class LoraArguments (line 76) | class LoraArguments:
  function maybe_zero_3 (line 89) | def maybe_zero_3(param):
  function get_peft_state_maybe_zero_3 (line 100) | def get_peft_state_maybe_zero_3(named_params, bias):
  function rank0_print (line 126) | def rank0_print(*args):
  function safe_save_model_for_hf_trainer (line 131) | def safe_save_model_for_hf_trainer(trainer, output_dir: str, bias="none"):
  function make_supervised_data_module (line 147) | def make_supervised_data_module(
  function get_parameter_number (line 197) | def get_parameter_number(model):
  function train (line 215) | def train():

FILE: ft_language_replace_file/finetune/only_language_web_demo.py
  function create_component (line 110) | def create_component(params, comp='Slider'):
  function chat (line 134) | def chat(img, msgs, ctx, params=None, vision_hidden_states=None):
  function upload_img (line 175) | def upload_img(image, _chatbot, _app_session):
  function respond (line 185) | def respond(_question, _chat_bot, _app_cfg, params_form, num_beams, repe...
  function regenerate_button_clicked (line 227) | def regenerate_button_clicked(_question, _chat_bot, _app_cfg, params_for...

FILE: ft_language_replace_file/finetune/replace_file/modeling_minicpmv.py
  class MiniCPMVPreTrainedModel (line 25) | class MiniCPMVPreTrainedModel(LlamaPreTrainedModel):
  class MiniCPMV (line 29) | class MiniCPMV(MiniCPMVPreTrainedModel):
    method __init__ (line 30) | def __init__(self, config):
    method init_vision_module (line 40) | def init_vision_module(self):
    method init_resampler (line 51) | def init_resampler(self, embed_dim, vision_dim,):
    method init_transform (line 60) | def init_transform(self):
    method get_input_embeddings (line 70) | def get_input_embeddings(self):
    method set_input_embeddings (line 73) | def set_input_embeddings(self, value):
    method get_vllm_embedding (line 76) | def get_vllm_embedding(self, data):
    method forward (line 184) | def forward(self, data, **kwargs):
    method _convert_to_tensors (line 206) | def _convert_to_tensors(
    method _process_list (line 231) | def _process_list(
    method _decode (line 246) | def _decode(self, inputs_embeds, tokenizer, **kwargs):
    method _decode_stream (line 259) | def _decode_stream(self, inputs_embeds, tokenizer, **kwargs):
    method _decode_text (line 278) | def _decode_text(self, result_ids, tokenizer):
    method slice_image (line 289) | def slice_image(self, image):
    method get_slice_image_placeholder (line 297) | def get_slice_image_placeholder(self, image, tokenizer):
    method reshape_by_patch (line 327) | def reshape_by_patch(self, image_tensor):
    method generate (line 344) | def generate(
    method chat (line 396) | def chat(
  class PreTrainedTokenizerFastWrapper (line 509) | class PreTrainedTokenizerFastWrapper(PreTrainedTokenizerFast):
    method __init__ (line 510) | def __init__(self, **kwargs):
    method eos_id (line 525) | def eos_id(self):
    method bos_id (line 529) | def bos_id(self):
    method unk_id (line 533) | def unk_id(self):
    method eot_id (line 537) | def eot_id(self):
    method im_start_id (line 541) | def im_start_id(self):
    method im_end_id (line 545) | def im_end_id(self):
    method escape (line 549) | def escape(text: str) -> str:
    method unescape (line 553) | def unescape(text: str) -> str:
  function pad (line 557) | def pad(orig_items, key, max_length=None, padding_value=0, padding_side=...
  function slice_image (line 605) | def slice_image(
  function ensure_divide (line 662) | def ensure_divide(length, patch_size):
  function find_best_resize (line 666) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 677) | def get_refine_size(
  function split_to_patches (line 701) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 718) | def get_grid_placeholder(tokenizer, grid, query_num):

FILE: ft_language_replace_file/finetune/replace_file/resampler.py
  function get_2d_sincos_pos_embed (line 17) | def get_2d_sincos_pos_embed(embed_dim, image_size):
  function get_2d_sincos_pos_embed_from_grid (line 37) | def get_2d_sincos_pos_embed_from_grid(embed_dim, grid):
  function get_1d_sincos_pos_embed_from_grid_new (line 48) | def get_1d_sincos_pos_embed_from_grid_new(embed_dim, pos):
  class Resampler (line 68) | class Resampler(nn.Module):
    method __init__ (line 76) | def __init__(
    method _set_2d_pos_cache (line 109) | def _set_2d_pos_cache(self, max_size, device='cpu'):
    method _adjust_pos_cache (line 115) | def _adjust_pos_cache(self, tgt_sizes, device):
    method _init_weights (line 122) | def _init_weights(self, m):
    method forward (line 131) | def forward(self, x, tgt_sizes=None):
    method _repeat (line 172) | def _repeat(self, query, N: int):
  class MultiheadAttention (line 176) | class MultiheadAttention(nn.MultiheadAttention):
    method __init__ (line 177) | def __init__(self, embed_dim, num_heads, dropout=0., bias=True, add_bi...
    method forward (line 184) | def forward(
    method multi_head_attention_forward (line 340) | def multi_head_attention_forward(
  function _mha_shape_check (line 624) | def _mha_shape_check(query: Tensor, key: Tensor, value: Tensor,
  function _canonical_mask (line 672) | def _canonical_mask(
  function _none_or_dtype (line 701) | def _none_or_dtype(input: Optional[Tensor]) -> Optional[DType]:
  function _in_projection_packed (line 708) | def _in_projection_packed(
  function _in_projection (line 768) | def _in_projection(

FILE: ft_language_replace_file/finetune/trainer.py
  class CPMTrainer (line 11) | class CPMTrainer(Trainer):
    method compute_loss (line 12) | def compute_loss(self, model, inputs, return_outputs=False):
    method prediction_step (line 44) | def prediction_step(
    method training_step (line 172) | def training_step(self, model: nn.Module, inputs: Dict[str, Union[torc...
    method _save (line 214) | def _save(self, output_dir: Optional[str] = None, state_dict=None):

FILE: get_minicpmv2.6_embeding/dataset.py
  class ImageDataset (line 19) | class ImageDataset(Dataset):
    method __init__ (line 21) | def __init__(
    method __len__ (line 42) | def __len__(self):
    method __getitem__ (line 45) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  class QueryDataset (line 72) | class QueryDataset(Dataset):
    method __init__ (line 75) | def __init__(
    method __len__ (line 86) | def __len__(self):
    method __getitem__ (line 89) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  function data_collator_query (line 116) | def data_collator_query(examples, padding_value=0, max_length=2048):
  function data_collator (line 126) | def data_collator(examples, padding_value=0, max_length=2048,device='cpu'):
  function load_from_pdf (line 162) | def load_from_pdf(pdf_path: str):
  function load_from_json (line 167) | def load_from_json(json_path: str):
  function conversation_to_ids (line 173) | def conversation_to_ids(conversation, tokenizer, llm_type=None, new_sche...
  function conversation_to_ids_minicpm (line 238) | def conversation_to_ids_minicpm(conversation, tokenizer):
  function conversation_to_ids_llama3 (line 270) | def conversation_to_ids_llama3(conversation, tokenizer):
  function conversation_to_ids_qwen2 (line 310) | def conversation_to_ids_qwen2(conversation, tokenizer):
  function preprocess (line 350) | def preprocess(
  function slice_image (line 446) | def slice_image(
  function ensure_divide (line 505) | def ensure_divide(length, patch_size):
  function find_best_resize (line 509) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 520) | def get_refine_size(
  function split_to_patches (line 544) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 561) | def get_grid_placeholder(tokenizer, grid, query_num, new_schema=False):
  function reshape_by_patch (line 582) | def reshape_by_patch(image_tensor, patch_size):

FILE: get_minicpmv2.6_embeding/inference.py
  function build_transform (line 9) | def build_transform():
  function main (line 20) | def main() -> None:

FILE: get_minicpmv2.6_embeding/modeling_minicpmv.py
  class MiniCPMVPreTrainedModel (line 18) | class MiniCPMVPreTrainedModel(Qwen2PreTrainedModel):
  class MiniCPMV (line 22) | class MiniCPMV(MiniCPMVPreTrainedModel):
    method __init__ (line 23) | def __init__(self, config):
    method init_vision_module (line 34) | def init_vision_module(self):
    method init_resampler (line 50) | def init_resampler(self, embed_dim, vision_dim):
    method get_input_embeddings (line 59) | def get_input_embeddings(self):
    method set_input_embeddings (line 62) | def set_input_embeddings(self, value):
    method get_output_embeddings (line 65) | def get_output_embeddings(self):
    method set_output_embeddings (line 68) | def set_output_embeddings(self, new_embeddings):
    method set_decoder (line 71) | def set_decoder(self, decoder):
    method get_decoder (line 74) | def get_decoder(self):
    method get_vllm_embedding (line 77) | def get_vllm_embedding(self, data):
    method forward (line 174) | def forward(self, data, **kwargs):
    method _decode (line 193) | def _decode(self, inputs_embeds, tokenizer, attention_mask, decode_tex...
    method _decode_stream (line 206) | def _decode_stream(self, inputs_embeds, tokenizer, **kwargs):
    method _decode_text (line 222) | def _decode_text(self, result_ids, tokenizer):
    method generate (line 234) | def generate(
    method chat (line 278) | def chat(

FILE: mbti_role_play/mbti_demo.py
  function check_model_v (line 44) | def check_model_v(img_file_path: str = None):
  function hf_gen (line 68) | def hf_gen(dialog: List, top_p: float, temperature: float, repetition_pe...
  function hf_v_gen (line 102) | def hf_v_gen(dialog: List, top_p: float, temperature: float, repetition_...
  function generate (line 134) | def generate(chat_history: List, query: str, I_E_choice: str, N_S_choice...
  function regenerate (line 174) | def regenerate(chat_history: List, I_E_choice, N_S_choice, T_F_choice, J...
  function clear_history (line 212) | def clear_history():
  function reverse_last_round (line 221) | def reverse_last_round(chat_history):
  function process_choice (line 232) | def process_choice(I_E, N_S, T_F, J_P):

FILE: mbti_role_play/mbti_sft_dpo_data/get_rank_data.py
  function load_json (line 8) | def load_json(file_path):

FILE: windows_minicpm3.0_agent/app.py
  function search (line 17) | def search(query: str, ds, images):
  function index (line 39) | def index(file, ds):

FILE: windows_minicpm3.0_agent/cli_demo.py
  function get_relevant_image (line 13) | def get_relevant_image(query,file_path):

FILE: windows_minicpm3.0_agent/dataset.py
  class ImageDataset (line 19) | class ImageDataset(Dataset):
    method __init__ (line 21) | def __init__(
    method __len__ (line 42) | def __len__(self):
    method __getitem__ (line 45) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  class QueryDataset (line 72) | class QueryDataset(Dataset):
    method __init__ (line 75) | def __init__(
    method __len__ (line 86) | def __len__(self):
    method __getitem__ (line 89) | def __getitem__(self, i) -> Dict[str, torch.Tensor]:
  function data_collator_query (line 103) | def data_collator_query(examples, padding_value=0, max_length=2048):
  function data_collator (line 113) | def data_collator(examples, padding_value=0, max_length=2048,device='cpu'):
  function load_from_pdf (line 150) | def load_from_pdf(pdf_path: str):
  function load_from_json (line 155) | def load_from_json(json_path: str):
  function conversation_to_ids (line 161) | def conversation_to_ids(conversation, tokenizer, llm_type=None, new_sche...
  function conversation_to_ids_minicpm (line 226) | def conversation_to_ids_minicpm(conversation, tokenizer):
  function conversation_to_ids_llama3 (line 258) | def conversation_to_ids_llama3(conversation, tokenizer):
  function conversation_to_ids_qwen2 (line 298) | def conversation_to_ids_qwen2(conversation, tokenizer):
  function preprocess (line 338) | def preprocess(
  function slice_image (line 434) | def slice_image(
  function ensure_divide (line 493) | def ensure_divide(length, patch_size):
  function find_best_resize (line 497) | def find_best_resize(original_size, scale_resolution, patch_size, allow_...
  function get_refine_size (line 508) | def get_refine_size(
  function split_to_patches (line 532) | def split_to_patches(image, grid):
  function get_grid_placeholder (line 549) | def get_grid_placeholder(tokenizer, grid, query_num, new_schema=False):
  function reshape_by_patch (line 570) | def reshape_by_patch(image_tensor, patch_size):

FILE: windows_minicpm3.0_agent/inference.py
  function main (line 12) | def main() -> None:

FILE: windows_minicpm3.0_agent/utils.py
  function evaluate_colbert (line 4) | def evaluate_colbert( qs, ps, batch_size=128) -> torch.Tensor:
  function build_transform (line 22) | def build_transform():

FILE: windows_minicpm3.0_agent/windows_agent.py
  function text_to_image_search (line 20) | def text_to_image_search(text_description,images_path):
  function image_anwer_question (line 23) | def image_anwer_question(image_path,query):
  function fake_tool_execute (line 40) | def fake_tool_execute(toolcalls):
  function clear_cuda_variables (line 114) | def clear_cuda_variables():