MCPcopy Create free account

hub / github.com/cure-lab/deep-active-learning / functions

Functions479 in github.com/cure-lab/deep-active-learning

⨍ Functions 479◇ Types & classes 111
↓ 31 callersMethodmax_accuracy
(self, istrain)
utils.py:71
↓ 18 callersMethodstep
(self)
query_strategies/util.py:168
↓ 15 callersMethodupdate
(self, val, n=1)
utils.py:35
↓ 13 callersMethodpredict
(self, X, Y)
query_strategies/strategy.py:155
↓ 13 callersFunctionprint_log
(string, log)
utils.py:10
↓ 11 callersFunctionadjust_learning_rate
Sets the learning rate to the initial LR decayed by 10 every 30 epochs
utils.py:156
↓ 11 callersFunctionsample_level
(n)
query_strategies/augmentations.py:55
↓ 11 callersMethodupdate
(self, idx, train_loss, train_acc, val_loss, val_acc)
utils.py:61
↓ 10 callersFunctionconvert_secs2time
(epoch_time)
utils.py:142
↓ 10 callersFunctiontime_string
()
utils.py:16
↓ 8 callersMethodpredict_prob
(self, X, Y)
query_strategies/strategy.py:195
↓ 7 callersMethodsave_model
(self)
query_strategies/strategy.py:297
↓ 6 callersFunctionfloat_parameter
Helper function to scale `val` between 0 and maxval. Args: level: Level of the operation that will be between [0, `PARAMETER_MAX`]. maxval:
query_strategies/augmentations.py:41
↓ 6 callersMethodget_embedding
get last layer embedding from current model
query_strategies/strategy.py:249
↓ 6 callersMethodget_embedding_dim
(self)
models/vgg.py:39
↓ 6 callersMethodplot_curve
(self, save_path)
utils.py:76
↓ 6 callersFunctionset_requires_grad
Used in training adversarial approach :param model: :param requires_grad: :return:
query_strategies/wasserstein_adversarial.py:15
↓ 5 callersMethod__init__
(self, z_dim=32, nc=3)
query_strategies/vaal.py:108
↓ 5 callersMethod__init__
(self, block, num_blocks, n_class=10, bayesian=False)
models/resnet64.py:122
↓ 5 callersMethod__init__
(self, n_class=10, bayesian=False)
models/linear.py:98
↓ 5 callersMethod__init__
(self)
models/wa_model.py:26
↓ 5 callersMethod__init__
(self, block, num_blocks, n_class=10, bayesian=False)
models/resnet.py:123
↓ 5 callersFunctionint_parameter
Helper function to scale `val` between 0 and maxval . Args: level: Level of the operation that will be between [0, `PARAMETER_MAX`]. maxval
query_strategies/augmentations.py:27
↓ 5 callersMethodreset_parameters
(self)
models/gcn.py:25
↓ 5 callersMethodtrain
(self, alpha=0.1, n_epoch=10)
query_strategies/semi_strategy.py:183
↓ 4 callersMethod_make_layer
(self, block, planes, num_blocks, stride)
models/resnet64.py:80
↓ 4 callersMethod_make_layer
(self, block, planes, num_blocks, stride)
models/resnet.py:80
↓ 4 callersMethodtrain
(self, alpha=0, n_epoch=80)
query_strategies/vaal.py:437
↓ 3 callersMethod__init__
(self)
query_strategies/learning_loss_for_al.py:58
↓ 3 callersMethod_make_layer
(self, out_planes, num_blocks, groups)
models/shufflenet.py:70
↓ 3 callersMethod_wide_layer
(self, block, planes, num_blocks, dropout_rate, stride)
models/wideresnet.py:65
↓ 3 callersMethodpredict_prob_dropout
(self, X, Y, n_drop)
query_strategies/strategy.py:212
↓ 3 callersMethodtrain
(self, alpha=0.1, n_epoch=10)
query_strategies/strategy.py:92
↓ 3 callersMethodupdate_distances
Update min distances given cluster centers. Args: cluster_centers: indices of cluster centers only_new: only calculate dis
query_strategies/coreGCN.py:101
↓ 2 callersMethodBrightness
(self, img, v, max_v, bias=0)
query_strategies/semi_fixmatch.py:39
↓ 2 callersMethodColor
(self, img, v, max_v, bias=0)
query_strategies/semi_fixmatch.py:43
↓ 2 callersMethodContrast
(self, img, v, max_v, bias=0)
query_strategies/semi_fixmatch.py:47
↓ 2 callersMethodCutoutAbs
(self, img, v, **kwarg)
query_strategies/semi_flexmatch.py:57
↓ 2 callersMethodCutoutAbs
(self, img, v, **kwarg)
query_strategies/semi_fixmatch.py:58
↓ 2 callersFunctionResNet18
(n_class, bayesian=False)
models/resnet.py:177
↓ 2 callersMethodSharpness
(self, img, v, max_v, bias=0)
query_strategies/semi_fixmatch.py:92
↓ 2 callersMethod__init__
(self, in_planes, out_planes, stride, groups)
models/shufflenet.py:23
↓ 2 callersMethod_train
(self, epoch, loader_tr_labeled, loader_tr_unlabeled, optimizer, global_batch)
query_strategies/aug_uda.py:181
↓ 2 callersFunctionadjust_learning_rate
Sets the learning rate to the initial LR decayed by 10 every 30 epochs
query_strategies/learning_loss_for_al.py:283
↓ 2 callersFunctionaug
Perform AugMix augmentations and compute mixture. Args: image: PIL.Image input image preprocess: Preprocessing function which shou
query_strategies/util.py:201
↓ 2 callersMethodavg_cosine_distance
Input: @idx: the index of the sample to be measured density @listed: the index of the labeled samples Return:
query_strategies/mcadl.py:106
↓ 2 callersMethoddecoder
(self, z)
query_strategies/vaal.py:60
↓ 2 callersMethodencoder
(self, x)
query_strategies/vaal.py:50
↓ 2 callersFunctiongemm_func
Get appropriate blas function Parameters ---------- precision : dtype or str dtype or string signifying the datatype for whi
query_strategies/cpu_dist.py:107
↓ 2 callersMethodgetID
Input: @pred: the predicted label for the all samples @listID: the index of k neighbor to current sample Return:
query_strategies/mcadl.py:84
↓ 2 callersMethodget_grad_embedding
gradient embedding (assumes cross-entropy loss) of the last layer
query_strategies/strategy.py:269
↓ 2 callersFunctiongradient_penalty
Gradeitnt penalty approach
query_strategies/wasserstein_adversarial.py:28
↓ 2 callersFunctioninterleave
(xy, batch)
query_strategies/util.py:148
↓ 2 callersMethodkaiming_init
(self, m)
query_strategies/vaal.py:176
↓ 2 callersFunctionoutput_dtype
Get promoted datatype for A and B combined alongwith consideration for another input datatype. Parameters ---------- A : ndarray
query_strategies/cpu_dist.py:76
↓ 2 callersMethodpredict
(self,X,Y)
query_strategies/wasserstein_adversarial.py:254
↓ 2 callersMethodread_data
(self, dataloader, labels=True)
query_strategies/vaal.py:304
↓ 2 callersMethodtrain
Only training samples with labeled and unlabeled data-set alpha is the trade-off between the empirical loss and error, the more inter
query_strategies/wasserstein_adversarial.py:83
↓ 2 callersMethodtrain
(self, n_epoch=10, X=None, Y=None)
query_strategies/lal.py:227
↓ 2 callersMethodtrain
(self,alpha=0, n_epoch=80)
query_strategies/learning_loss_for_al.py:173
↓ 2 callersMethodvae_loss
(self, x, recon, mu, logvar, beta)
query_strategies/vaal.py:298
↓ 1 callersFunctionBCEAdjLoss
(scores, lbl, nlbl, l_adj)
query_strategies/uncertainGCN.py:36
↓ 1 callersFunctionBCEAdjLoss
(scores, lbl, nlbl, l_adj)
query_strategies/coreGCN.py:45
↓ 1 callersMethodL1_upper
Return the /|-log(proba)/|_1 :param probas: :return: # unlabeled \times 1
query_strategies/wasserstein_adversarial.py:360
↓ 1 callersMethodL2_upper
Return the /|-log(proba)/|_2 :param probas: :return: # unlabeled \times 1 (float tensor)
query_strategies/wasserstein_adversarial.py:347
↓ 1 callersFunctionLossPredLoss
(input, target, margin=1.0, reduction='mean')
query_strategies/learning_loss_for_al.py:102
↓ 1 callersFunctionShuffleNetG2
(channels=3,num_classes=10, dropout=False)
models/shufflenet.py:98
↓ 1 callersMethod__init__
(self, all_data_for_lal, all_labels_for_lal)
query_strategies/lal.py:23
↓ 1 callersMethod__init__
(self, init_points, cluster_id)
query_strategies/batch_active_learning_at_scale.py:13
↓ 1 callersMethod__init__
(self, transform1,transform2)
query_strategies/ssl_diff2augdirect.py:9
↓ 1 callersMethod__init__
(self, transform)
query_strategies/ssl_consistency.py:9
↓ 1 callersMethod__init__
(self, transform1,transform2)
query_strategies/ssl_diff2augkmeans.py:10
↓ 1 callersMethod__init__
r nfeat: input feature dimension nhid: the hidden layer dimension nclass: the output dimension
models/gcn.py:46
↓ 1 callersMethod__init__
(self, in_planes, out_planes, stride=1)
models/mobilenet.py:13
↓ 1 callersMethod__init__
(self, in_planes, planes, dropout_rate, stride=1)
models/wideresnet.py:23
↓ 1 callersMethod_decode
(self, z)
query_strategies/vaal.py:189
↓ 1 callersMethod_encode
(self, x)
query_strategies/vaal.py:186
↓ 1 callersMethod_getFeaturevector4LAL
(self, unknown_data, known_labels, nFeatures)
query_strategies/lal.py:111
↓ 1 callersMethod_make_layers
(self, in_planes)
models/mobilenet.py:39
↓ 1 callersMethod_make_layers
(self, cfg)
models/wa_model.py:111
↓ 1 callersMethod_make_layers
(self, cfg)
models/vgg.py:25
↓ 1 callersFunction_resize
(X,length)
query_strategies/mcadl.py:14
↓ 1 callersMethod_train
(self, epoch, loader_tr, optimizer)
query_strategies/strategy.py:55
↓ 1 callersMethod_train
(self, epoch, loader_tr_labeled, loader_tr_unlabeled, optimizer)
query_strategies/semi_flexmatch.py:214
↓ 1 callersMethod_train
(self, epoch, loader_tr_labeled, loader_tr_unlabeled, optimizer)
query_strategies/semi_pseudolabel.py:55
↓ 1 callersMethod_train
(self, epoch, loader_labeled, loader_unlabeled, optimizer, ema_optimizer, train_iteration)
query_strategies/semi_strategy.py:64
↓ 1 callersMethod_train
(self, epoch, loader_tr_labeled, loader_tr_unlabeled, optimizer)
query_strategies/semi_fixmatch.py:214
↓ 1 callersFunctionadjust_learning_rate
Sets the learning rate to the initial LR decayed by 10 every 30 epochs
query_strategies/vaal.py:526
↓ 1 callersFunctionaff_to_adj
(x, y=None)
query_strategies/uncertainGCN.py:24
↓ 1 callersFunctionaff_to_adj
(x, y=None)
query_strategies/coreGCN.py:33
↓ 1 callersFunctionauto_dtype
Get promoted datatype for A and B combined. Parameters ---------- A : ndarray B : ndarray Returns ------- precision
query_strategies/cpu_dist.py:50
↓ 1 callersFunctionaverage_linkage
(cluster_1, cluster_2)
query_strategies/batch_active_learning_at_scale.py:26
↓ 1 callersMethodbuiltModel
Fits the regressor with the parameters identifier as an input
query_strategies/lal.py:58
↓ 1 callersMethodcal_dis
(self, x)
query_strategies/adversarial_deepfool.py:14
↓ 1 callersMethodcal_dis
(self, x)
query_strategies/adversarial_bim.py:14
↓ 1 callersMethodcompute_NKC
(self, X,Y)
query_strategies/batch_BALD.py:27
↓ 1 callersFunctionconv3x3
(in_planes, out_planes, stride=1)
models/wideresnet.py:10
↓ 1 callersMethodcrossValidateLALmodel
Cross-validate the regressor model. input: possible_estimators -- list of possible number of estimators (trees) in Random Forest regression
query_strategies/lal.py:28
↓ 1 callersMethoddiff2_aug_kmeans
(self, unlabeled_index, Kmeans_list, k)
query_strategies/ssl_diff2augkmeans.py:68
↓ 1 callersFunctiondist_accum
Compute squared euclidean distance between two 2D arrays representing n-dimensional points using accumulation based approach. For more in
query_strategies/cpu_dist.py:153
next →1–100 of 479, ranked by callers