SYMBOL INDEX (229 symbols across 36 files)

FILE: torchtools/lr_scheduler/delayed.py
  class DelayerScheduler (line 3) | class DelayerScheduler(_LRScheduler):
    method __init__ (line 12) | def __init__(self, optimizer, delay_epochs, after_scheduler):
    method get_lr (line 18) | def get_lr(self):
    method step (line 27) | def step(self, epoch=None):
  function DelayedCosineAnnealingLR (line 36) | def DelayedCosineAnnealingLR(optimizer, delay_epochs, cosine_annealing_e...

FILE: torchtools/lr_scheduler/inverse_sqrt.py
  class InverseSqrtLR (line 6) | class InverseSqrtLR(LRScheduler):
    method __init__ (line 7) | def __init__(self, optimizer, lr, warmup_steps, pre_warmup_lr=None, la...
    method _process_lr (line 14) | def _process_lr(self, _):
    method get_lr (line 19) | def get_lr(self):
    method _get_closed_form_lr (line 26) | def _get_closed_form_lr(self):

FILE: torchtools/nn/adain.py
  class AdaIN (line 4) | class AdaIN(nn.Module):
    method __init__ (line 5) | def __init__(self, n_channels):
    method forward (line 9) | def forward(self, image, style):

FILE: torchtools/nn/alias_free_activation.py
  class AliasFreeActivation (line 13) | class AliasFreeActivation(nn.Module):
    method __init__ (line 14) | def __init__(self, activation, level, max_levels, max_size, max_channe...
    method alias_level_params (line 46) | def alias_level_params(level, max_levels, max_size, max_channels, star...
    method _lowpass_filter (line 60) | def _lowpass_filter(self, n_taps, cutoff, band_half, sr):
    method _kaiser_window (line 67) | def _kaiser_window(self, n_taps, f_h, sr):
    method _kaiser_attenuation (line 72) | def _kaiser_attenuation(self, n_taps, f_h, sr):
    method _kaiser_beta (line 77) | def _kaiser_beta(self, n_taps, f_h, sr):
    method forward (line 87) | def forward(self, x):
    method _upsample (line 96) | def _upsample(self, x, kernel, factor, pad=(0, 0)):
    method _downsample (line 101) | def _downsample(self, x, kernel, factor, pad=(0, 0)):
    method extra_repr (line 106) | def extra_repr(self):

FILE: torchtools/nn/equal_layers.py
  class EqualLinear (line 12) | class EqualLinear(nn.Linear):
    method __init__ (line 13) | def __init__(self, *args, bias_init=0, lr_mul=1, **kwargs):
    method forward (line 23) | def forward(self, x):
  class EqualConv2d (line 27) | class EqualConv2d(nn.Conv2d):
    method __init__ (line 28) | def __init__(self, *args, **kwargs):
    method forward (line 38) | def forward(self, x):
  class EqualLeakyReLU (line 42) | class EqualLeakyReLU(nn.LeakyReLU):
    method __init__ (line 43) | def __init__(self, *args, scale=2**0.5, **kwargs):
    method forward (line 47) | def forward(self, x):

FILE: torchtools/nn/evonorm2d.py
  class SwishImplementation (line 6) | class SwishImplementation(torch.autograd.Function):
    method forward (line 8) | def forward(ctx, i):
    method backward (line 13) | def backward(ctx, grad_output):
  class MemoryEfficientSwish (line 18) | class MemoryEfficientSwish(nn.Module):
    method forward (line 19) | def forward(self, x):
  function instance_std (line 22) | def instance_std(x, eps=1e-5):
  function group_std (line 28) | def group_std(x, groups = 32, eps = 1e-5):
  class EvoNorm2D (line 34) | class EvoNorm2D(nn.Module):
    method __init__ (line 36) | def __init__(self, input, non_linear = True, version = 'S0', efficient...
    method reset_parameters (line 65) | def reset_parameters(self):
    method _check_input_dim (line 68) | def _check_input_dim(self, x):
    method forward (line 73) | def forward(self, x):

FILE: torchtools/nn/fourier_features.py
  class FourierFeatures2d (line 5) | class FourierFeatures2d(nn.Module):
    method __init__ (line 6) | def __init__(self, size, dim, cutoff, affine_eps=1e-8, freq_range=[-0....
    method forward (line 28) | def forward(self, affine):
    method extra_repr (line 60) | def extra_repr(self):

FILE: torchtools/nn/functional/gradient_penalty.py
  function gradient_penalty (line 9) | def gradient_penalty(netD, real_data, fake_data, l=10):

FILE: torchtools/nn/functional/magnitude_preserving.py
  function mp_cat (line 3) | def mp_cat(*args, dim=1, t=0.5):
  function mp_sum (line 13) | def mp_sum(*args, t=0.5):

FILE: torchtools/nn/functional/perceptual.py
  function total_variation (line 3) | def total_variation(X, reduction='sum'):

FILE: torchtools/nn/functional/vq.py
  class vector_quantize (line 4) | class vector_quantize(Function):
    method forward (line 6) | def forward(ctx, x, codebook):
    method backward (line 21) | def backward(ctx, grad_output, grad_indices):
  class binarize (line 36) | class binarize(Function):
    method forward (line 38) | def forward(ctx, x, threshold=0.5):
    method backward (line 46) | def backward(ctx, grad_output):

FILE: torchtools/nn/gp_loss.py
  class GPLoss (line 5) | class GPLoss(nn.Module):
    method __init__ (line 6) | def __init__(self, discriminator, l=10):
    method forward (line 11) | def forward(self, real_data, fake_data):

FILE: torchtools/nn/haar_dwt.py
  class HaarForward (line 5) | class HaarForward(nn.Module):
    method __init__ (line 10) | def __init__(self, beta=2):
    method forward (line 15) | def forward(self, x: torch.Tensor) -> torch.Tensor:
  class HaarInverse (line 33) | class HaarInverse(nn.Module):
    method __init__ (line 38) | def __init__(self, beta=2):
    method forward (line 43) | def forward(self, x: torch.Tensor) -> torch.Tensor:

FILE: torchtools/nn/magnitude_preserving.py
  class MP_GELU (line 4) | class MP_GELU(nn.GELU):
    method forward (line 5) | def forward(self, x):
  class MP_SiLU (line 8) | class MP_SiLU(nn.SiLU):
    method forward (line 9) | def forward(self, x):
  class Gain (line 12) | class Gain(nn.Module):
    method __init__ (line 13) | def __init__(self, init_w=0.0):
    method forward (line 17) | def forward(self, x):

FILE: torchtools/nn/mish.py
  class Mish (line 15) | class Mish(nn.Module):
    method __init__ (line 16) | def __init__(self):
    method forward (line 19) | def forward(self, x):

FILE: torchtools/nn/modulation.py
  class ModulatedConv2d (line 11) | class ModulatedConv2d(nn.Conv2d):
    method __init__ (line 12) | def __init__(self,  *args, demodulate=True, ema_decay=1.0, **kwargs):
    method forward (line 25) | def forward(self, x, w):

FILE: torchtools/nn/perceptual.py
  class TVLoss (line 5) | class TVLoss(nn.Module):
    method __init__ (line 6) | def __init__(self, reduction='sum', alpha=1e-4):
    method forward (line 11) | def forward(self, x):

FILE: torchtools/nn/pixel_normalzation.py
  class PixelNorm (line 4) | class PixelNorm(nn.Module):
    method __init__ (line 5) | def __init__(self, dim=1, eps=1e-4):
    method forward (line 10) | def forward(self, x):

FILE: torchtools/nn/pos_embeddings.py
  class RotaryEmbedding (line 8) | class RotaryEmbedding(nn.Module):
    method __init__ (line 9) | def __init__(self, dim, base=10000):
    method forward (line 17) | def forward(self, x, seq_dim=1):
  function rotate_half (line 29) | def rotate_half(x):
  function apply_rotary_pos_emb (line 34) | def apply_rotary_pos_emb(q, k, cos, sin):

FILE: torchtools/nn/simple_self_attention.py
  function conv1d (line 9) | def conv1d(ni, no, ks=1, stride=1, padding=0, bias=False):
  class SimpleSelfAttention (line 18) | class SimpleSelfAttention(nn.Module):
    method __init__ (line 20) | def __init__(self, n_in, ks=1, sym=False):
    method forward (line 27) | def forward(self, x):

FILE: torchtools/nn/stylegan2/upfirdn2d.cpp
  function upfirdn2d (line 17) | torch::Tensor upfirdn2d(const torch::Tensor &input, const torch::Tensor ...
  function PYBIND11_MODULE (line 29) | PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) {

FILE: torchtools/nn/stylegan2/upfirdn2d.py
  class UpFirDn2dBackward (line 20) | class UpFirDn2dBackward(Function):
    method forward (line 22) | def forward(
    method backward (line 64) | def backward(ctx, gradgrad_input):
  class UpFirDn2d (line 89) | class UpFirDn2d(Function):
    method forward (line 91) | def forward(ctx, input, kernel, up, down, pad):
    method backward (line 128) | def backward(ctx, grad_output):
  function upfirdn2d (line 149) | def upfirdn2d(input, kernel, up=1, down=1, pad=(0, 0)):
  function upfirdn2d_native (line 168) | def upfirdn2d_native(

FILE: torchtools/nn/transformers.py
  class GPTTransformerEncoderLayer (line 5) | class GPTTransformerEncoderLayer(nn.Module):
    method __init__ (line 6) | def __init__(self, d_model, nhead, dim_feedforward=2048, dropout=0):
    method forward (line 18) | def forward(self, x, src_mask=None, src_key_padding_mask=None):

FILE: torchtools/nn/vq.py
  class VectorQuantize (line 6) | class VectorQuantize(nn.Module):
    method __init__ (line 7) | def __init__(self, embedding_size, k, ema_decay=0.99, ema_loss=False):
    method _laplace_smoothing (line 27) | def _laplace_smoothing(self, x, epsilon):
    method _updateEMA (line 31) | def _updateEMA(self, z_e_x, indices):
    method idx2vq (line 42) | def idx2vq(self, idx, dim=-1):
    method forward (line 48) | def forward(self, x, get_losses=True, dim=-1):
  class Binarize (line 67) | class Binarize(nn.Module):
    method __init__ (line 68) | def __init__(self, threshold=0.5):
    method forward (line 78) | def forward(self, x):
  class FSQ (line 82) | class FSQ(nn.Module):
    method __init__ (line 83) | def __init__(self, bins, dim=-1, eps=1e-1):
    method _round (line 93) | def _round(self, x, quantize):
    method vq_to_idx (line 102) | def vq_to_idx(self, x, is_sigmoid=False):
    method idx_to_vq (line 110) | def idx_to_vq(self, x):
    method forward (line 118) | def forward(self, x, quantize=True):

FILE: torchtools/optim/lamb.py
  function log_lamb_rs (line 14) | def log_lamb_rs(optimizer: Optimizer, event_writer: SummaryWriter, token...
  class Lamb (line 29) | class Lamb(Optimizer):
    method __init__ (line 47) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-6,
    method step (line 62) | def step(self, closure=None):

FILE: torchtools/optim/lookahead.py
  class Lookahead (line 16) | class Lookahead(Optimizer):
    method __init__ (line 17) | def __init__(self, base_optimizer, alpha=0.5, k=6):
    method update_slow (line 33) | def update_slow(self, group):
    method sync_lookahead (line 45) | def sync_lookahead(self):
    method step (line 49) | def step(self, closure=None):
    method state_dict (line 59) | def state_dict(self):
    method load_state_dict (line 73) | def load_state_dict(self, state_dict):
  function LookaheadAdam (line 100) | def LookaheadAdam(params, alpha=0.5, k=6, *args, **kwargs):

FILE: torchtools/optim/novograd.py
  class AdamW (line 24) | class AdamW(Optimizer):
    method __init__ (line 47) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8,
    method __setstate__ (line 61) | def __setstate__(self, state):
    method step (line 66) | def step(self, closure=None):
  class Novograd (line 124) | class Novograd(Optimizer):
    method __init__ (line 143) | def __init__(self, params, lr=1e-3, betas=(0.95, 0), eps=1e-8,
    method __setstate__ (line 160) | def __setstate__(self, state):
    method step (line 165) | def step(self, closure=None):

FILE: torchtools/optim/over9000.py
  function Over9000 (line 17) | def Over9000(params, alpha=0.5, k=6, *args, **kwargs):

FILE: torchtools/optim/radam.py
  class RAdam (line 10) | class RAdam(Optimizer):
    method __init__ (line 11) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8, weig...
    method __setstate__ (line 29) | def __setstate__(self, state):
    method step (line 32) | def step(self, closure=None):
  class PlainRAdam (line 100) | class PlainRAdam(Optimizer):
    method __init__ (line 102) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8, weig...
    method __setstate__ (line 117) | def __setstate__(self, state):
    method step (line 120) | def step(self, closure=None):
  class AdamW (line 177) | class AdamW(Optimizer):
    method __init__ (line 179) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8, weig...
    method __setstate__ (line 193) | def __setstate__(self, state):
    method step (line 196) | def step(self, closure=None):

FILE: torchtools/optim/ralamb.py
  class Ralamb (line 9) | class Ralamb(Optimizer):
    method __init__ (line 11) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8, weig...
    method __setstate__ (line 16) | def __setstate__(self, state):
    method step (line 19) | def step(self, closure=None):

FILE: torchtools/optim/ranger.py
  function Ranger (line 14) | def Ranger(params, alpha=0.5, k=6, betas=(.95, 0.999), *args, **kwargs):

FILE: torchtools/transforms/smart_crop.py
  class ResBlock (line 8) | class ResBlock(nn.Module):
    method __init__ (line 9) | def __init__(self, channels):
    method forward (line 22) | def forward(self, x):
  class Upsample2d (line 26) | class Upsample2d(nn.Module):
    method __init__ (line 27) | def __init__(self, scale_factor):
    method forward (line 33) | def forward(self, x):
  class MicroResNet (line 37) | class MicroResNet(nn.Module):
    method __init__ (line 38) | def __init__(self):
    method forward (line 73) | def forward(self, x):
  class SmartCrop (line 80) | class SmartCrop(nn.Module):
    method __init__ (line 81) | def __init__(self, output_size, randomize_p=0.0, randomize_q=0.1, temp...
    method forward (line 92) | def forward(self, image):

FILE: torchtools/utils/diffusion.py
  class SimpleSampler (line 4) | class SimpleSampler():
    method __init__ (line 5) | def __init__(self, diffuzz):
    method __call__ (line 9) | def __call__(self, *args, **kwargs):
    method init_x (line 13) | def init_x(self, shape):
    method step (line 16) | def step(self, x, t, t_prev, noise):
  class DDPMSampler (line 19) | class DDPMSampler(SimpleSampler):
    method step (line 20) | def step(self, x, t, t_prev, noise):
  class DDIMSampler (line 29) | class DDIMSampler(SimpleSampler):
    method step (line 30) | def step(self, x, t, t_prev, noise):
  class DPMSolverPlusPlusSampler (line 38) | class DPMSolverPlusPlusSampler(SimpleSampler):  # FIXME: CURRENTLY NOT W...
    method __init__ (line 39) | def __init__(self, *args, **kwargs):
    method _get_coef (line 43) | def _get_coef(self, alpha_cumprod):
    method init_x (line 50) | def init_x(self, shape):
    method step (line 54) | def step(self, x, t, t_prev, noise):
  class Diffuzz (line 89) | class Diffuzz():
    method __init__ (line 90) | def __init__(self, s=0.008, device="cpu", cache_steps=None, scaler=1, ...
    method _alpha_cumprod (line 100) | def _alpha_cumprod(self, t):
    method diffuse (line 111) | def diffuse(self, x, t, noise=None): # t -> [0, 1]
    method undiffuse (line 117) | def undiffuse(self, x, t, t_prev, noise, sampler=None):
    method sample (line 122) | def sample(self, model, model_inputs, shape, mask=None, t_start=1.0, t...
    method p2_weight (line 152) | def p2_weight(self, t, k=1.0, gamma=1.0):

FILE: torchtools/utils/diffusion2.py
  class SimpleSampler (line 5) | class SimpleSampler():
    method __init__ (line 6) | def __init__(self, diffuzz, mode="v"):
    method __call__ (line 13) | def __call__(self, *args, **kwargs):
    method init_x (line 17) | def init_x(self, shape):
    method step (line 20) | def step(self, x, t, t_prev, noise):
  class DDIMSampler (line 24) | class DDIMSampler(SimpleSampler):
    method step (line 25) | def step(self, x, t, t_prev, pred, eta=0):
  class DDPMSampler (line 42) | class DDPMSampler(DDIMSampler):
    method step (line 43) | def step(self, x, t, t_prev, pred, eta=1):
  class Diffuzz2 (line 52) | class Diffuzz2():
    method __init__ (line 53) | def __init__(self, s=0.008, device="cpu", cache_steps=None, scaler=1, ...
    method _alpha_cumprod (line 63) | def _alpha_cumprod(self, t):
    method scale_t (line 73) | def scale_t(self, t, scaler):
    method diffuse (line 81) | def diffuse(self, x, t, noise=None): # t -> [0, 1]
    method get_v (line 87) | def get_v(self, x, t, noise):
    method x0_from_v (line 93) | def x0_from_v(self, noised, pred_v, t):
    method noise_from_v (line 97) | def noise_from_v(self, noised, pred_v, t):
    method undiffuse (line 101) | def undiffuse(self, x, t, t_prev, pred, sampler=None, **kwargs):
    method sample (line 106) | def sample(self, model, model_inputs, shape, mask=None, t_start=1.0, t...
    method p2_weight (line 158) | def p2_weight(self, t, k=1.0, gamma=1.0):
    method truncated_snr_weight (line 162) | def truncated_snr_weight(self, t, min=1.0, max=None):

FILE: torchtools/utils/gamma_parametrization.py
  class _GammaScaling (line 5) | class _GammaScaling(nn.Module):
    method __init__ (line 6) | def __init__(self):
    method forward (line 10) | def forward(self, w):
  function apply_gamma_reparam (line 13) | def apply_gamma_reparam(module, name="weight"): # this reparametrizes th...
  function gamma_reparam_model (line 18) | def gamma_reparam_model(model):
  function remove_gamma_reparam (line 26) | def remove_gamma_reparam(model):

FILE: torchtools/utils/weight_normalization.py
  class _WeigthNorm (line 4) | class _WeigthNorm(nn.Module):
    method __init__ (line 5) | def __init__(self, eps=1e-4):
    method _normalize (line 9) | def _normalize(self, w):
    method forward (line 15) | def forward(self, w):
  function apply_weight_norm (line 23) | def apply_weight_norm(module, name="weight", init_weight=True): # this r...
  function weight_norm_model (line 29) | def weight_norm_model(model, whitelist=None, init_weight=True):
  function remove_weight_norm (line 43) | def remove_weight_norm(model):