MCPcopy Create free account

hub / github.com/ZhaoJ9014/face.evoLVe / functions

Functions423 in github.com/ZhaoJ9014/face.evoLVe

⨍ Functions 423◇ Types & classes 103
↓ 13 callersMethod__init__
(self, feat_dim, num_classes)
head/metrics.py:329
↓ 12 callersFunctionget_block
(in_channel, depth, num_units, stride=2)
backbone/model_irse.py:98
↓ 12 callersFunctionget_block
(in_channel, depth, num_units, stride=2)
paddle/backbone/model_irse.py:95
↓ 9 callersFunctionschedule_lr
(optimizer, )
paddle/utils.py:73
↓ 9 callersMethodupdate
(self, val, n=1)
paddle/utils.py:66
↓ 8 callersFunctionget_time
()
paddle/utils.py:8
↓ 7 callersMethod__init__
(self, block, ch_in, ch_out, count,
paddle/backbone/resnet_pp.py:367
↓ 7 callersFunctionbuffer_val
(writer, db_name, acc, best_threshold, roc_curve_tensor, epoch)
backup/utils.py:195
↓ 7 callersFunctionbuffer_val
(writer, db_name, acc, best_threshold, roc_curve_tensor, epoch)
util/utils.py:217
↓ 7 callersFunctionget_same_padding_conv2d
Chooses static padding if you have specified an image size, and dynamic padding otherwise. Static padding is necessary for ONNX exporting of mo
backbone/EfficientNets.py:191
↓ 7 callersFunctionget_val_pair
(path, name)
backup/utils.py:55
↓ 7 callersFunctionget_val_pair
(path, name)
util/utils.py:55
↓ 7 callersFunctionperform_val
(multi_gpu, device, embedding_size, batch_size, backbone, carray, issame, nrof_folds = 10, tta = True)
backup/utils.py:158
↓ 7 callersFunctionperform_val
(multi_gpu, device, embedding_size, batch_size, backbone, carray, issame, nrof_folds = 10, tta = True)
util/utils.py:176
↓ 6 callersFunctionload_weight
(model, weight_path, optimizer=None)
paddle/backbone/resnet_pp.py:691
↓ 5 callersMethod__init__
(self, block_args, global_params, image_size=None)
backbone/EfficientNets.py:655
↓ 5 callersFunction_make_divisible
This function is taken from the original tf repo. It ensures that all layers have a channel number that is divisible by 8 It can be seen
backbone/GhostNet.py:26
↓ 5 callersMethodbackward
(ctx, grad_output)
backbone/EfficientNets.py:65
↓ 5 callersFunctionround_filters
Calculate and round number of filters based on width multiplier. Use width_coefficient, depth_divisor and min_depth of global_params. Args
backbone/EfficientNets.py:75
↓ 5 callersFunctiontformfwd
Function: ---------- apply affine transform 'trans' to uv Parameters: ---------- @trans: 3x3 np.array tr
applications/align/matlab_cp2tform.py:11
↓ 4 callersMethod__init__
(self, input_channels, output_channels, stride=1)
backbone/AttentionNets.py:16
↓ 4 callersMethod__init__
(self, width=1.0, drop_ratio=0.2, feat_dim=512, out_h=7, out_w=7)
backbone/GhostNet.py:174
↓ 4 callersMethod__init__
(self, c, num_block, groups, kernel=(3, 3), stride=(1, 1), padding=(1, 1))
backbone/MobileFaceNets.py:53
↓ 4 callersMethod__init__
(self, in_features, out_features)
paddle/head/metrics.py:17
↓ 4 callersMethod__init__
(self, in_features, out_features)
backup/metrics.py:19
↓ 4 callersMethod_make_layer
(self, block, planes, blocks, stride = 1)
backbone/model_resnet.py:131
↓ 4 callersMethod_make_layer
(self, block, planes, blocks, stride = 1)
paddle/backbone/model_resnet.py:132
↓ 4 callersFunctionccrop_batch
(imgs_tensor)
util/utils.py:153
↓ 4 callersFunctioncreat_predictor
返回face检测器
paddle/Paddle-Lite-Inference-demo/utils.py:4
↓ 4 callersFunctionget_time
()
backup/utils.py:20
↓ 4 callersFunctionget_time
()
util/utils.py:20
↓ 4 callersFunctionl2_norm
(input, axis = 1)
util/extract_feature_v1.py:10
↓ 4 callersFunctionl2_norm
(input, axis = 1)
util/utils.py:24
↓ 4 callersFunctionnms
Non-maximum suppression. Arguments: boxes: a float numpy array of shape [n, 5], where each row is (xmin, ymin, xmax, ymax, sc
applications/align/box_utils.py:5
↓ 4 callersFunctionpy_nms
贪婪策略选择人脸框 keep boxes overlap <= thresh rule out overlap > thresh :param dets: [[x1, y1, x2, y2 score]] :param thresh: retain over
paddle/PaddleInference-demo/utils.py:363
↓ 4 callersFunctionpy_nms
贪婪策略选择人脸框 keep boxes overlap <= thresh rule out overlap > thresh :param dets: [[x1, y1, x2, y2 score]] :param thresh: retain over
paddle/Paddle-Lite-Inference-demo/MTCNN.py:311
↓ 4 callersFunctionseparate_irse_bn_paras
(modules)
paddle/utils.py:12
↓ 4 callersFunctionseparate_resnet_bn_paras
(modules)
paddle/utils.py:81
↓ 3 callersMethod__init__
(self, channels, reduction)
backbone/model_irse.py:24
↓ 3 callersMethod__init__
(self)
applications/align/get_nets.py:29
↓ 3 callersMethod__init__
(self, channels, reduction)
paddle/backbone/model_irse.py:24
↓ 3 callersMethod__init__
(self)
paddle/align/get_nets.py:29
↓ 3 callersFunctionaccuracy
Computes the precision@k for the specified values of k
paddle/utils.py:32
↓ 3 callersFunctioncalculate_accuracy
(threshold, dist, actual_issame)
util/verification.py:94
↓ 3 callersFunctioncalculate_output_image_size
Calculates the output image size when using Conv2dSamePadding with a stride. Necessary for static padding. Thanks to mannatsingh for pointing t
backbone/EfficientNets.py:166
↓ 3 callersFunctioncalibrate_box
Transform bounding boxes to be more like true bounding boxes. 'offsets' is one of the outputs of the nets. Arguments: bboxes: a float
applications/align/box_utils.py:94
↓ 3 callersFunctioncalibrate_box
Transform bounding boxes to be more like true bounding boxes. 'offsets' is one of the outputs of the nets. Arguments: bboxes: a float
paddle/align/box_utils.py:94
↓ 3 callersFunctionconv1x1
1x1 convolution
backbone/model_resnet.py:15
↓ 3 callersFunctionconv1x1
1x1 convolution
paddle/backbone/model_resnet.py:16
↓ 3 callersFunctionconv3x3
3x3 convolution with padding
backbone/model_resnet.py:8
↓ 3 callersFunctionconv3x3
3x3 convolution with padding
paddle/backbone/model_resnet.py:9
↓ 3 callersFunctionfindNonreflectiveSimilarity
(uv, xy, options=None)
applications/align/matlab_cp2tform.py:60
↓ 3 callersFunctioninit_predictor
(model_dir)
paddle/PaddleInference-demo/utils.py:8
↓ 3 callersFunctionl2_norm
(input, axis = 1)
head/metrics.py:263
↓ 3 callersFunctionnms
Non-maximum suppression. Arguments: boxes: a float numpy array of shape [n, 5], where each row is (xmin, ymin, xmax, ymax, sc
paddle/align/box_utils.py:5
↓ 3 callersFunctionschedule_lr
(optimizer)
backup/utils.py:114
↓ 3 callersFunctionschedule_lr
(optimizer)
util/utils.py:114
↓ 3 callersMethodupdate
(self, val, n = 1)
backup/utils.py:212
↓ 3 callersMethodupdate
(self, val, n = 1)
util/utils.py:234
↓ 2 callersFunctionIR_101
Constructs a ir-101 model.
backbone/model_irse.py:200
↓ 2 callersFunctionIR_101
Constructs a ir-101 model.
paddle/backbone/model_irse.py:180
↓ 2 callersFunctionIR_152
Constructs a ir-152 model.
backbone/model_irse.py:208
↓ 2 callersFunctionIR_152
Constructs a ir-152 model.
paddle/backbone/model_irse.py:188
↓ 2 callersFunctionIR_50
Constructs a ir-50 model.
backbone/model_irse.py:192
↓ 2 callersFunctionIR_50
Constructs a ir-50 model.
paddle/backbone/model_irse.py:172
↓ 2 callersFunctionIR_SE_101
Constructs a ir_se-101 model.
backbone/model_irse.py:224
↓ 2 callersFunctionIR_SE_101
Constructs a ir_se-101 model.
paddle/backbone/model_irse.py:204
↓ 2 callersFunctionIR_SE_152
Constructs a ir_se-152 model.
backbone/model_irse.py:232
↓ 2 callersFunctionIR_SE_152
Constructs a ir_se-152 model.
paddle/backbone/model_irse.py:212
↓ 2 callersFunctionIR_SE_50
Constructs a ir_se-50 model.
backbone/model_irse.py:216
↓ 2 callersFunctionIR_SE_50
Constructs a ir_se-50 model.
paddle/backbone/model_irse.py:196
↓ 2 callersFunctionResNet_101
Constructs a ResNet-101 model.
backbone/model_resnet.py:175
↓ 2 callersFunctionResNet_101
Constructs a ResNet-101 model.
paddle/backbone/model_resnet.py:176
↓ 2 callersFunctionResNet_152
Constructs a ResNet-152 model.
backbone/model_resnet.py:183
↓ 2 callersFunctionResNet_152
Constructs a ResNet-152 model.
paddle/backbone/model_resnet.py:184
↓ 2 callersFunctionResNet_50
Constructs a ResNet-50 model.
backbone/model_resnet.py:167
↓ 2 callersFunctionResNet_50
Constructs a ResNet-50 model.
paddle/backbone/model_resnet.py:168
↓ 2 callersMethod__init__
(self, input_size, block, layers, zero_init_residual = True)
backbone/model_resnet.py:93
↓ 2 callersMethod__init__
(self, input_size, block, layers, zero_init_residual = True)
paddle/backbone/model_resnet.py:100
↓ 2 callersMethod_change_in_channels
Adjust model's first convolution layer to in_channels, if in_channels not equals 3. Args: in_channels (int): Input data's channel
backbone/EfficientNets.py:1024
↓ 2 callersMethod_check_model_name_is_valid
Validates model name. Args: model_name (str): Name for efficientnet. Returns: bool: Is a valid name or not.
backbone/EfficientNets.py:1012
↓ 2 callersMethod_paddle_where_more_than
(self, target, limit, x, y)
paddle/head/metrics.py:90
↓ 2 callersFunction_preprocess
Preprocessing step before feeding the network. Arguments: img: a float numpy array of shape [h, w, c]. Returns: a float nump
applications/align/box_utils.py:226
↓ 2 callersFunctioncalculate_val_far
(threshold, dist, actual_issame)
util/verification.py:151
↓ 2 callersFunctioncalibrate_box
校准box 参数： bbox:pnet生成的box reg:rnet生成的box偏移值 返回值： 调整后的box是针对原图的绝对坐标
paddle/PaddleInference-demo/utils.py:470
↓ 2 callersFunctioncalibrate_box
校准box 参数： bbox:pnet生成的box reg:rnet生成的box偏移值 返回值： 调整后的box是针对原图的绝对坐标
paddle/Paddle-Lite-Inference-demo/MTCNN.py:418
↓ 2 callersMethodcompute_theta
(self, p, g, label, batch_size)
head/metrics.py:651
↓ 2 callersFunctionconvert_to_square
Convert bounding boxes to a square form. Arguments: bboxes: a float numpy array of shape [n, 5]. Returns: a float numpy arra
applications/align/box_utils.py:71
↓ 2 callersFunctionconvert_to_square
将box转换成更大的正方形 参数： box：预测的box,[n,5] 返回值： 调整后的正方形box，[n,5]
paddle/PaddleInference-demo/utils.py:405
↓ 2 callersFunctionconvert_to_square
Convert bounding boxes to a square form. Arguments: bboxes: a float numpy array of shape [n, 5]. Returns: a float numpy arra
paddle/align/box_utils.py:71
↓ 2 callersFunctionconvert_to_square
将box转换成更大的正方形 参数： box：预测的box,[n,5] 返回值： 调整后的正方形box，[n,5]
paddle/Paddle-Lite-Inference-demo/MTCNN.py:353
↓ 2 callersFunctiondata_reader
()
paddle/quant/quant_post_static.py:16
↓ 2 callersMethoddetect_onet
将onet的选框继续筛选基本和rnet差不多但多返回了landmark
paddle/PaddleInference-demo/utils.py:197
↓ 2 callersMethoddetect_onet
将onet的选框继续筛选基本和rnet差不多但多返回了landmark
paddle/Paddle-Lite-Inference-demo/MTCNN.py:157
↓ 2 callersMethoddetect_pnet
(self, im, min_face_size, scale_factor, thresh)
paddle/PaddleInference-demo/utils.py:107
↓ 2 callersMethoddetect_pnet
(self, im, min_face_size, scale_factor, thresh)
paddle/Paddle-Lite-Inference-demo/MTCNN.py:67
↓ 2 callersMethoddetect_rnet
通过rent选择box 参数： im：输入图像 dets:pnet选择的box，是相对原图的绝对坐标 返回值： box绝对坐标
paddle/PaddleInference-demo/utils.py:145
↓ 2 callersMethoddetect_rnet
通过rent选择box 参数： im：输入图像 dets:pnet选择的box，是相对原图的绝对坐标 返回值： box绝对坐标
paddle/Paddle-Lite-Inference-demo/MTCNN.py:105
↓ 2 callersFunctiondownload
Download from url, save to path. url (str): download url path (str): download to given path
paddle/backbone/resnet_pp.py:736
↓ 2 callersFunctionevaluate
(embeddings, actual_issame, nrof_folds = 10, pca = 0)
util/verification.py:162
next →1–100 of 423, ranked by callers