Functions — github.com/AtlasAnalyticsLab/AdaFisher

hub / github.com/AtlasAnalyticsLab/AdaFisher / functions

Functions413 in github.com/AtlasAnalyticsLab/AdaFisher

⨍ Functions 413◇ Types & classes 90

Method__init__

(self, optimizer, milestones, gamma=0.1, last_epoch=-1)

optimizers/lr_scheduler.py:460

Method__init__

(self, optimizer, gamma, last_epoch=-1)

optimizers/lr_scheduler.py:491

Method__init__

(self, optimizer, T_max, eta_min=0, last_epoch=-1)

optimizers/lr_scheduler.py:548

Method__init__

(self, optimizer, mode='min', factor=0.1, patience=10, verbose=False, threshold=1e-4, thresho

optimizers/lr_scheduler.py:625

Method__init__

(self, optimizer, base_lr, max_lr, step_si

optimizers/lr_scheduler.py:839

Method__init__

(self, optimizer, T_0, T_mult=1, eta_min=0, last_epoch=-1)

optimizers/lr_scheduler.py:997

Method__init__

(self, o)

optimizers/lr_scheduler.py:1071

Method__init__

(self, optimizer, max_lr, total_steps=None,

optimizers/lr_scheduler.py:1180

Method__init__

(self, optimizer, start_factor=1.0 / 3, end_factor=1.0, total_iters=5, last_epoch=-1)

optimizers/lr_scheduler.py:1350

Method__init__

( self, config_path: Path, device: str, output_path: Path,

Image_Classification/src/train.py:104

Method__init__

(self)

Image_Classification/src/train.py:467

Method__init__

(self, n_holes, length)

Image_Classification/src/utils/data.py:20

Method__init__

(self, patience: int = 10, threshold: float = 1e-2)

Image_Classification/src/utils/early_stop.py:7

Method__init__

(self, num_input_features, growth_rate, bn_size, drop_rate, memory_efficient=False)

Image_Classification/src/models/densenet.py:47

Method__init__

(self, num_layers, num_input_features, bn_size, growth_rate, drop_rate, memory_efficient=False)

Image_Classification/src/models/densenet.py:120

Method__init__

(self, num_input_features, num_output_features)

Image_Classification/src/models/densenet.py:141

Method__init__

(self, inplace=False)

Image_Classification/src/models/mobilenetv3.py:36

Method__init__

(self, in_channels_num, reduction_ratio=4)

Image_Classification/src/models/mobilenetv3.py:63

Method__init__

use_SE: True or False -- use SE Module or not NL: nonlinearity, 'RE' or 'HS'

Image_Classification/src/models/mobilenetv3.py:89

Method__init__

configs: setting of the model mode: type of the model, 'large' or 'small'

Image_Classification/src/models/mobilenetv3.py:165

Method__init__

(self, dim, num_heads=8, attention_dropout=0.1, projection_dropout=0.1)

Image_Classification/src/models/cct.py:91

Method__init__

(self, d_model, nhead, dim_feedforward=2048, dropout=0.1, attention_dropout=0.1, drop_path_ra

Image_Classification/src/models/cct.py:126

Method__init__

(self, drop_prob=None)

Image_Classification/src/models/cct.py:153

Method__init__

(self, kernel_size, stride, padding, pooling_kernel_size=3, pooling_stride=2

Image_Classification/src/models/cct.py:172

Method__init__

(self, seq_pool=True, embedding_dim=768, num_layers=12,

Image_Classification/src/models/cct.py:219

Method__init__

(self, displacement)

Image_Classification/src/models/swin.py:10

Method__init__

(self, fn)

Image_Classification/src/models/swin.py:19

Method__init__

(self, dim, hidden_dim)

Image_Classification/src/models/swin.py:38

Method__init__

(self, dim, heads, head_dim, shifted, window_size, relative_pos_embedding)

Image_Classification/src/models/swin.py:73

Method__init__

(self, dim, heads, head_dim, mlp_dim, shifted, window_size, relative_pos_embedding)

Image_Classification/src/models/swin.py:140

Method__init__

(self, in_channels, out_channels, downscaling_factor)

Image_Classification/src/models/swin.py:157

Method__init__

(self, in_channels, hidden_dimension, layers, downscaling_factor, num_heads, head_dim, window_size,

Image_Classification/src/models/swin.py:172

Method__init__

(self, *, hidden_dim, layers, heads, channels=3, num_classes=1000, head_dim=32, window_size=7,

Image_Classification/src/models/swin.py:198

Method__init__

(self, in_planes, planes, stride=1)

Image_Classification/src/models/resnet_cifar.py:17

Method__init__

(self, in_planes, planes, stride=1)

Image_Classification/src/models/resnet_cifar.py:46

Method__init__

(self, in_planes, growth_rate)

Image_Classification/src/models/densenet_cifar.py:10

Method__init__

(self, in_planes, out_planes)

Image_Classification/src/models/densenet_cifar.py:28

Method__init__

(self, dim, focal_window, focal_level, focal_factor=2, bias=True, proj_drop=0., use_postln_in

Image_Classification/src/models/focalnet.py:35

Method__init__

(self, dim, input_resolution, mlp_ratio=4., drop=0., drop_path=0., act_layer=nn.GELU, norm_la

Image_Classification/src/models/focalnet.py:145

Method__init__

(self, dim, out_dim, input_resolution, depth, mlp_ratio=4., drop=0., drop_path=0., norm_layer

Image_Classification/src/models/focalnet.py:240

Method__init__

(self, img_size=(224, 224), patch_size=4, in_chans=3, embed_dim=96, use_conv_embed=False, nor

Image_Classification/src/models/focalnet.py:326

Method__init__

(self, img_size=32, patch_size=2, in_chans=3,

Image_Classification/src/models/focalnet.py:400

Method__init__

(self, inplanes, planes, stride=1, downsample=None, groups=1, base_width=64, dilation=1, norm

Image_Classification/src/models/resnet.py:50

Method__init__

(self, inplanes, planes, stride=1, downsample=None, groups=1, base_width=64, dilation=1, norm

Image_Classification/src/models/resnet.py:98

Method__init__

(self, model, lr=0.001, momentum=0.9, stat

Analysis/optimizers/kfac.py:9

Method__init__

Initializes the AdaFisher optimizer. Parameters: - model (Module): The neural network model to optimize.

Analysis/optimizers/AdaFisher.py:48

Method__init__

(self, params, lr=0.15, betas=(0.9, 0.999), eps=1e-4, weight_decay=0, hessian_power=1, single

Analysis/optimizers/AdaHessian.py:22

Method__init__

(self, params, lr=required, momentum=0, dampening=0, weight_decay=0, nesterov=False)

Analysis/optimizers/sgd.py:239

Method__init__

(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8, weight_decay=0, amsgrad=False)

Analysis/optimizers/Adam.py:102

Method__init__

(self, in_planes, planes, stride=1)

Analysis/models/resnet_cifar.py:16

Method__init__

(self, in_planes, planes, stride=1)

Analysis/models/resnet_cifar.py:45

Method__init__

(self, head_size, num_heads, sequence_length)

Language_Model/GPT1.py:10

Method__init__

( self, vocabulary_size=40479, embedding_size=768, sequence_le

Language_Model/GPT1.py:271

Method__init__

( self, root, split="train", break_mode="none", max_length=256, min_length=1 )

Language_Model/utils/wikitext2.py:63

Method__init__

( self, vocabulary_size, embedding_size, sequence_length, _tokens_embe

Language_Model/utils/embeddings.py:6

Method__len__

(self)

Language_Model/utils/wikitext2.py:122

Method__setstate__

(self, state)

optimizers/AdamW.py:121

Method__setstate__

(self, state)

optimizers/sgd.py:256

Method__setstate__

(self, state)

optimizers/Adam.py:121

Method__setstate__

(self, state)

Analysis/optimizers/sgd.py:256

Method__setstate__

(self, state)

Analysis/optimizers/Adam.py:121

Method_annealing_cos

Cosine anneal from `start` to `end` as pct goes from 0.0 to 1.0.

optimizers/lr_scheduler.py:1267

Method_annealing_linear

Linearly anneal from `start` to `end` as pct goes from 0.0 to 1.0.

optimizers/lr_scheduler.py:1272

Method_exp_range_scale_fn

(self, x)

optimizers/lr_scheduler.py:924

Method_get_closed_form_lr

(self)

optimizers/lr_scheduler.py:475

Method_get_closed_form_lr

(self)

optimizers/lr_scheduler.py:505

Method_get_closed_form_lr

(self)

optimizers/lr_scheduler.py:570

Method_get_closed_form_lr

(self)

optimizers/lr_scheduler.py:1378

Method_init_weights

(self, m)

Image_Classification/src/models/focalnet.py:482

Method_save_grad_output

(self, module, grad_input, grad_output)

optimizers/kfac.py:61

Method_save_grad_output

Updates diagonal elements of gradient covariance matrix periodically. Args: module: Neural network layer gra

optimizers/AdaFisher.py:415

Method_save_grad_output

(self, module, grad_input, grad_output)

Analysis/optimizers/kfac.py:61

Method_save_grad_output

Updates the optimizer's internal state by capturing and maintaining the diagonal and full matrix elements of the gradient covariance

Analysis/optimizers/AdaFisher.py:154

Method_save_input

(self, module, input)

optimizers/kfac.py:53

Method_save_input

Updates diagonal elements of activation covariance matrix periodically. Args: module: Neural network layer i

optimizers/AdaFisher.py:396

Method_save_input

(self, module, input)

Analysis/optimizers/kfac.py:53

Method_save_input

Captures and updates the diagonal and full matrix elements of the activation covariance matrix for a given module at specified interv

Analysis/optimizers/AdaFisher.py:116

Method_triangular2_scale_fn

(self, x)

optimizers/lr_scheduler.py:921

Method_triangular_scale_fn

(self, x)

optimizers/lr_scheduler.py:918

Methodbatchnorm2d

Computes gradient covariance diagonal for BatchNorm2d. Args: s: Output gradients layer: BatchNorm2d layer

optimizers/AdaFisher.py:292

Methodbatchnorm2d

Computes the diagonal and the full matrix of the covariance matrix for activations from a BatchNorm2d layer. Parameters: - h

Analysis/optimizers/AdaFisher_utils.py:175

Functioncct_14

(*args, **kwargs)

Image_Classification/src/models/cct.py:52

Functioncct_16

(*args, **kwargs)

Image_Classification/src/models/cct.py:57

Functioncct_2

(*args, **kwargs)

Image_Classification/src/models/cct.py:27

Functioncct_4

(*args, **kwargs)

Image_Classification/src/models/cct.py:32

Functioncct_6

(*args, **kwargs)

Image_Classification/src/models/cct.py:37

Functioncct_7

(*args, **kwargs)

Image_Classification/src/models/cct.py:42

Functioncct_8

(*args, **kwargs)

Image_Classification/src/models/cct.py:47

Methodclosure

(*inputs)

Image_Classification/src/models/densenet.py:79

Methodcompute_G

Computes the diagonal and the full matrix of the gradient covariance matrix for the gradients of a given layer. Parameters:

Analysis/optimizers/AdaFisher_utils.py:224

Methodcompute_H_D

Computes diagonal of activation covariance matrix. Args: h: Input activations layer: PyTorch layer (Linear,

optimizers/AdaFisher.py:98

Methodcompute_H_bar

Computes the diagonal and the full matrix of the covariance matrix for the activations of a given layer. Parameters: - a (Te

Analysis/optimizers/AdaFisher_utils.py:91

Methodcompute_S_D

Computes gradient covariance matrix diagonal. Args: s: Output gradients layer: PyTorch layer (Conv2d, Linear

optimizers/AdaFisher.py:219

Methodcompute_cov_a

(cls, a, layer)

optimizers/kfac_utils.py:92

Methodcompute_cov_a

(cls, a, layer)

Analysis/optimizers/kfac_utils.py:92

Methodcompute_cov_g

:param g: gradient :param layer: the corresponding layer :param batch_averaged: if the gradient is already averaged with the

optimizers/kfac_utils.py:132

Methodcompute_cov_g

:param g: gradient :param layer: the corresponding layer :param batch_averaged: if the gradient is already averaged with the

Analysis/optimizers/kfac_utils.py:132

Methodconv2d

:param input: batch_size * in_c * in_h * in_w :param grad_output: batch_size * out_c * h * w :param layer: nn.module batch_si

optimizers/kfac_utils.py:69

Methodconv2d

(g, layer, batch_averaged)

optimizers/kfac_utils.py:154

Methodconv2d

Computes gradient covariance diagonal for Conv2d. Args: s: Gradients (batch_size, n_filters, out_h, out_w) l

optimizers/AdaFisher.py:257

← previous next →201–300 of 413, ranked by callers