From c7063d7fd649ebda6a55018bee307d8caeb6992b Mon Sep 17 00:00:00 2001 From: pengzirong Date: Mon, 14 Mar 2022 10:49:55 +0800 Subject: [PATCH 1/2] =?UTF-8?q?=E5=A2=9E=E5=8A=A0torch=5Fnpu.=E5=89=8D?= =?UTF-8?q?=E7=BC=80?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- ...57\346\214\201\346\270\205\345\215\225.md" | 322 +++++++++--------- 1 file changed, 161 insertions(+), 161 deletions(-) diff --git "a/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" "b/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" index fbda2b9f5f..9fd2577488 100644 --- "a/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" +++ "b/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" @@ -1371,103 +1371,103 @@ torch_npu.npu.set_device()接口只支持在程序开始的位置通过set_devic ## NPU自定义算子 -| 序号 | 算子名称 | -| ---- | ---------------------------------------------- | -| 1 | npu_convolution_transpose | -| 2 | npu_conv_transpose2d | -| 3 | npu_convolution_transpose_backward | -| 4 | npu_conv_transpose2d_backward | -| 5 | npu_conv_transpose3d_backward | -| 6 | npu_convolution | -| 7 | npu_convolution_backward | -| 8 | npu_convolution_double_backward | -| 9 | npu_conv2d | -| 10 | npu_conv2d.out | -| 11 | npu_conv2d_backward | -| 12 | npu_conv3d | -| 13 | npu_conv3d.out | -| 14 | npu_conv3d_backward | -| 15 | one_ | -| 16 | npu_sort_v2.out | -| 17 | npu_sort_v2 | -| 18 | npu_format_cast | -| 19 | npu_format_cast_.acl_format | -| 20 | npu_format_cast_.src | -| 21 | npu_transpose_to_contiguous | -| 22 | npu_transpose | -| 23 | npu_transpose.out | -| 24 | npu_broadcast | -| 25 | npu_broadcast.out | -| 26 | npu_dtype_cast | -| 27 | npu_dtype_cast_.Tensor | -| 28 | npu_roi_alignbk | -| 29 | empty_with_format | -| 30 | empty_with_format.names | -| 31 | copy_memory_ | -| 32 | npu_one_hot | -| 33 | npu_stride_add | -| 34 | npu_softmax_cross_entropy_with_logits | -| 35 | npu_softmax_cross_entropy_with_logits_backward | -| 36 | npu_ps_roi_pooling | -| 37 | npu_ps_roi_pooling_backward | -| 38 | npu_roi_align | -| 39 | npu_nms_v4 | -| 40 | npu_lstm | -| 41 | npu_lstm_backward | -| 42 | npu_iou | -| 43 | npu_ptiou | -| 44 | npu_nms_with_mask | -| 45 | npu_pad | -| 46 | npu_bounding_box_encode | -| 47 | npu_bounding_box_decode | -| 48 | npu_gru | -| 49 | npu_gru_backward | -| 50 | npu_set_.source_Storage_storage_offset_format | -| 51 | npu_random_choice_with_mask | -| 52 | npu_batch_nms | -| 53 | npu_slice | -| 54 | npu_slice.out | -| 55 | npu_dropoutV2 | -| 56 | npu_dropoutV2_backward | -| 57 | _npu_dropout | -| 58 | _npu_dropout_inplace | -| 59 | npu_dropout_backward | -| 60 | npu_indexing | -| 61 | npu_indexing.out | -| 62 | npu_ifmr | -| 63 | npu_max.dim | -| 64 | npu_max.names_dim | -| 65 | npu_scatter | -| 66 | npu_max_backward | -| 67 | npu_apply_adam | -| 68 | npu_layer_norm_eval | -| 69 | npu_alloc_float_status | -| 70 | npu_get_float_status | -| 71 | npu_clear_float_status | -| 72 | npu_confusion_transpose | -| 73 | npu_confusion_transpose_backward | -| 74 | npu_bmmV2 | -| 75 | fast_gelu | -| 76 | fast_gelu_backward | -| 77 | npu_sub_sample | -| 78 | npu_deformable_conv2d | -| 79 | npu_deformable_conv2dbk | -| 80 | npu_mish | -| 81 | npu_anchor_response_flags | -| 82 | npu_yolo_boxes_encode | -| 83 | npu_grid_assign_positive | -| 84 | npu_mish_backward | -| 85 | npu_normalize_batch | -| 86 | npu_masked_fill_range | -| 87 | npu_linear | -| 88 | npu_linear_backward | -| 89 | npu_bert_apply_adam | -| 90 | npu_giou | -| 91 | npu_giou_backward | +| 序号 | 算子名称 | +| ---- | -------------------------------------------------------- | +| 1 | torch_npu.npu_convolution_transpose | +| 2 | torch_npu.npu_conv_transpose2d | +| 3 | torch_npu.npu_convolution_transpose_backward | +| 4 | torch_npu.npu_conv_transpose2d_backward | +| 5 | torch_npu.npu_conv_transpose3d_backward | +| 6 | torch_npu.npu_convolution | +| 7 | torch_npu.npu_convolution_backward | +| 8 | torch_npu.npu_convolution_double_backward | +| 9 | torch_npu.npu_conv2d | +| 10 | torch_npu.npu_conv2d.out | +| 11 | torch_npu.npu_conv2d_backward | +| 12 | torch_npu.npu_conv3d | +| 13 | torch_npu.npu_conv3d.out | +| 14 | torch_npu.npu_conv3d_backward | +| 15 | torch_npu.one_ | +| 16 | torch_npu.npu_sort_v2.out | +| 17 | torch_npu.npu_sort_v2 | +| 18 | torch_npu.npu_format_cast | +| 19 | torch_npu.npu_format_cast_.acl_format | +| 20 | torch_npu.npu_format_cast_.src | +| 21 | torch_npu.npu_transpose_to_contiguous | +| 22 | torch_npu.npu_transpose | +| 23 | torch_npu.npu_transpose.out | +| 24 | torch_npu.npu_broadcast | +| 25 | torch_npu.npu_broadcast.out | +| 26 | torch_npu.npu_dtype_cast | +| 27 | torch_npu.npu_dtype_cast_.Tensor | +| 28 | torch_npu.npu_roi_alignbk | +| 29 | torch_npu.empty_with_format | +| 30 | torch_npu.empty_with_format.names | +| 31 | torch_npu.copy_memory_ | +| 32 | torch_npu.npu_one_hot | +| 33 | torch_npu.npu_stride_add | +| 34 | torch_npu.npu_softmax_cross_entropy_with_logits | +| 35 | torch_npu.npu_softmax_cross_entropy_with_logits_backward | +| 36 | torch_npu.npu_ps_roi_pooling | +| 37 | torch_npu.npu_ps_roi_pooling_backward | +| 38 | torch_npu.npu_roi_align | +| 39 | torch_npu.npu_nms_v4 | +| 40 | torch_npu.npu_lstm | +| 41 | torch_npu.npu_lstm_backward | +| 42 | torch_npu.npu_iou | +| 43 | torch_npu.npu_ptiou | +| 44 | torch_npu.npu_nms_with_mask | +| 45 | torch_npu.npu_pad | +| 46 | torch_npu.npu_bounding_box_encode | +| 47 | torch_npu.npu_bounding_box_decode | +| 48 | torch_npu.npu_gru | +| 49 | torch_npu.npu_gru_backward | +| 50 | torch_npu.npu_set_.source_Storage_storage_offset_format | +| 51 | torch_npu.npu_random_choice_with_mask | +| 52 | torch_npu.npu_batch_nms | +| 53 | torch_npu.npu_slice | +| 54 | torch_npu.npu_slice.out | +| 55 | torch_npu.npu_dropoutV2 | +| 56 | torch_npu.npu_dropoutV2_backward | +| 57 | torch_npu._npu_dropout | +| 58 | torch_npu._npu_dropout_inplace | +| 59 | torch_npu.npu_dropout_backward | +| 60 | torch_npu.npu_indexing | +| 61 | torch_npu.npu_indexing.out | +| 62 | torch_npu.npu_ifmr | +| 63 | torch_npu.npu_max.dim | +| 64 | torch_npu.npu_max.names_dim | +| 65 | torch_npu.npu_scatter | +| 66 | torch_npu.npu_max_backward | +| 67 | torch_npu.npu_apply_adam | +| 68 | torch_npu.npu_layer_norm_eval | +| 69 | torch_npu.npu_alloc_float_status | +| 70 | torch_npu.npu_get_float_status | +| 71 | torch_npu.npu_clear_float_status | +| 72 | torch_npu.npu_confusion_transpose | +| 73 | torch_npu.npu_confusion_transpose_backward | +| 74 | torch_npu.npu_bmmV2 | +| 75 | torch_npu.fast_gelu | +| 76 | torch_npu.fast_gelu_backward | +| 77 | torch_npu.npu_sub_sample | +| 78 | torch_npu.npu_deformable_conv2d | +| 79 | torch_npu.npu_deformable_conv2dbk | +| 80 | torch_npu.npu_mish | +| 81 | torch_npu.npu_anchor_response_flags | +| 82 | torch_npu.npu_yolo_boxes_encode | +| 83 | torch_npu.npu_grid_assign_positive | +| 84 | torch_npu.npu_mish_backward | +| 85 | torch_npu.npu_normalize_batch | +| 86 | torch_npu.npu_masked_fill_range | +| 87 | torch_npu.npu_linear | +| 88 | torch_npu.npu_linear_backward | +| 89 | torch_npu.npu_bert_apply_adam | +| 90 | torch_npu.npu_giou | +| 91 | torch_npu.npu_giou_backward | 详细算子接口说明： -> npu_apply_adam(beta1_power, beta2_power, lr, beta1, beta2, epsilon, grad, use_locking, use_nesterov, out = (var, m, v)) +> torch_npu.npu_apply_adam(beta1_power, beta2_power, lr, beta1, beta2, epsilon, grad, use_locking, use_nesterov, out = (var, m, v)) count adam result. @@ -1493,7 +1493,7 @@ count adam result. None -> npu_convolution_transpose(input, weight, bias, padding, output_padding, stride, dilation, groups) -> Tensor +> torch_npu.npu_convolution_transpose(input, weight, bias, padding, output_padding, stride, dilation, groups) -> Tensor Applies a 2D or 3D transposed convolution operator over an input image composed of several input planes, sometimes also called “deconvolution”. @@ -1515,7 +1515,7 @@ Applies a 2D or 3D transposed convolution operator over an input image composed None -> npu_conv_transpose2d(input, weight, bias, padding, output_padding, stride, dilation, groups) -> Tensor +> torch_npu.npu_conv_transpose2d(input, weight, bias, padding, output_padding, stride, dilation, groups) -> Tensor Applies a 2D transposed convolution operator over an input image composed of several input planes, sometimes also called “deconvolution”. @@ -1537,7 +1537,7 @@ Applies a 2D transposed convolution operator over an input image composed of sev None -> npu_convolution(input, weight, bias, stride, padding, dilation, groups) -> Tensor +> torch_npu.npu_convolution(input, weight, bias, stride, padding, dilation, groups) -> Tensor Applies a 2D or 3D convolution over an input image composed of several input planes. @@ -1558,7 +1558,7 @@ Applies a 2D or 3D convolution over an input image composed of several input pla None -> npu_conv2d(input, weight, bias, stride, padding, dilation, groups) -> Tensor +> torch_npu.npu_conv2d(input, weight, bias, stride, padding, dilation, groups) -> Tensor Applies a 2D convolution over an input image composed of several input planes. @@ -1579,7 +1579,7 @@ Applies a 2D convolution over an input image composed of several input planes. None -> npu_conv3d(input, weight, bias, stride, padding, dilation, groups) -> Tensor +> torch_npu.npu_conv3d(input, weight, bias, stride, padding, dilation, groups) -> Tensor Applies a 3D convolution over an input image composed of several input planes. @@ -1600,7 +1600,7 @@ Applies a 3D convolution over an input image composed of several input planes. None -> one_(self) -> Tensor +> torch_npu.one_(self) -> Tensor Fills self tensor with ones. @@ -1624,7 +1624,7 @@ Fills self tensor with ones. [1., 1., 1.]], device='npu:0') ``` -> npu_sort_v2(self, dim=-1, descending=False, out=None) -> Tensor +> torch_npu.npu_sort_v2(self, dim=-1, descending=False, out=None) -> Tensor Sorts the elements of the input tensor along a given dimension in ascending order by value without indices. If dim is not given, the last dimension of the input is chosen. @@ -1655,7 +1655,7 @@ If descending is True then the elements are sorted in descending order by value. [-2.7441, 0.1880, 0.7378, 1.3975]], device='npu:0') ``` -> npu_format_cast(self, acl_format) -> Tensor +> torch_npu.npu_format_cast(self, acl_format) -> Tensor Change the format of a npu tensor. @@ -1678,13 +1678,13 @@ Change the format of a npu tensor. 29 ``` -> npu_format_cast_ +> torch_npu.npu_format_cast_ -> npu_format_cast_.acl_format(self, acl_format) -> Tensor +> torch_npu.npu_format_cast_.acl_format(self, acl_format) -> Tensor In-place version of npu_format_cast() -> npu_format_cast_.src(self, src) -> Tensor +> torch_npu.npu_format_cast_.src(self, src) -> Tensor In-place Change the format of self, with the same format as src. @@ -1706,7 +1706,7 @@ Change the format of a npu tensor. 29 ``` -> npu_transpose(self, perm) -> Tensor +> torch_npu.npu_transpose(self, perm) -> Tensor Returns a view of the original tensor with its dimensions permuted, and make the result contiguous. @@ -1732,7 +1732,7 @@ Returns a view of the original tensor with its dimensions permuted, and make the torch.Size([5, 2, 3]) ``` -> npu_broadcast(self, perm) -> Tensor +> torch_npu.npu_broadcast(self, perm) -> Tensor Returns a new view of the self tensor with singleton dimensions expanded to a larger size, and make the result contiguous. @@ -1758,7 +1758,7 @@ Tensor can be also expanded to a larger number of dimensions, and the new ones w [3, 3, 3, 3]], device='npu:0') ``` -> npu_dtype_cast(input, dtype) -> Tensor +> torch_npu.npu_dtype_cast(input, dtype) -> Tensor Performs Tensor dtype conversion. @@ -1777,7 +1777,7 @@ Performs Tensor dtype conversion. tensor([ 0, 0, -1], device='npu:0', dtype=torch.int32) ``` -> empty_with_format(size, dtype, layout, device, pin_memory, acl_format) -> Tensor +> torch_npu.empty_with_format(size, dtype, layout, device, pin_memory, acl_format) -> Tensor Returns a tensor filled with uninitialized data. The shape of the tensor is defined by the variable argument size. The format of the tensor is defined by the variable argument acl_format. @@ -1806,7 +1806,7 @@ Returns a tensor filled with uninitialized data. The shape of the tensor is defi [1., 1., 1.]], device='npu:0') ``` -> copy_memory_(dst, src, non_blocking=False) -> Tensor +> torch_npu.copy_memory_(dst, src, non_blocking=False) -> Tensor Copies the elements from src into self tensor and returns self. @@ -1830,7 +1830,7 @@ Copies the elements from src into self tensor and returns self. tensor([1, 1, 1], device='npu:0', dtype=torch.int32) ``` -> npu_one_hot(input, num_classes=-1, depth=1, on_value=1, off_value=0) -> Tensor +> torch_npu.npu_one_hot(input, num_classes=-1, depth=1, on_value=1, off_value=0) -> Tensor Returns a one-hot tensor. The locations represented by index in "x" take value "on_value", while all other locations take value "off_value". @@ -1856,7 +1856,7 @@ Returns a one-hot tensor. The locations represented by index in "x" take value " [0., 1., 0., 0., 0.]], device='npu:0') ``` -> npu_stride_add(x1, x2, offset1, offset2, c1_len) -> Tensor +> torch_npu.npu_stride_add(x1, x2, offset1, offset2, c1_len) -> Tensor Add the partial values of two tensors in format NC1HWC0. @@ -1894,7 +1894,7 @@ Add the partial values of two tensors in format NC1HWC0. [[[0.]]]]], device='npu:0') ``` -> npu_softmax_cross_entropy_with_logits(features, labels) -> Tensor +> torch_npu.npu_softmax_cross_entropy_with_logits(features, labels) -> Tensor Computes softmax cross entropy cost. @@ -1910,7 +1910,7 @@ Computes softmax cross entropy cost. None -> npu_ps_roi_pooling(x, rois, spatial_scale, group_size, output_dim) -> Tensor +> torch_npu.npu_ps_roi_pooling(x, rois, spatial_scale, group_size, output_dim) -> Tensor Performs Position Sensitive PS ROI Pooling. @@ -1945,7 +1945,7 @@ Performs Position Sensitive PS ROI Pooling. [0., 0.]]]], device='npu:0', dtype=torch.float16) ``` -> npu_roi_align(features, rois, spatial_scale, pooled_height, pooled_width, sample_num, roi_end_mode) -> Tensor +> torch_npu.npu_roi_align(features, rois, spatial_scale, pooled_height, pooled_width, sample_num, roi_end_mode) -> Tensor Obtains the ROI feature matrix from the feature map. It is a customized FasterRcnn operator. @@ -1978,7 +1978,7 @@ Obtains the ROI feature matrix from the feature map. It is a customized FasterRc [28.5000, 30.5000, 32.5000]]]], device='npu:0') ``` -> npu_nms_v4(boxes, scores, max_output_size, iou_threshold, scores_threshold, pad_to_max_output_size=False) -> (Tensor, Tensor) +> torch_npu.npu_nms_v4(boxes, scores, max_output_size, iou_threshold, scores_threshold, pad_to_max_output_size=False) -> (Tensor, Tensor) Greedily selects a subset of bounding boxes in descending order of score. @@ -2013,7 +2013,7 @@ Greedily selects a subset of bounding boxes in descending order of score. 54, 92], device='npu:0', dtype=torch.int32), tensor(20, device='npu:0', dtype=torch.int32)) ``` -> npu_nms_rotated(dets, scores, iou_threshold, scores_threshold=0, max_output_size=-1, mode=0) -> (Tensor, Tensor) +> torch_npu.npu_nms_rotated(dets, scores, iou_threshold, scores_threshold=0, max_output_size=-1, mode=0) -> (Tensor, Tensor) Greedy selects a subset of the rotated bounding boxes in descending fractional order. @@ -2049,7 +2049,7 @@ Greedy selects a subset of the rotated bounding boxes in descending fractional o tensor([62], device='npu:0', dtype=torch.int32) ``` -> npu_lstm(x, weight, bias, seq_len, h, c, has_biases, num_layers, dropout, train, bidirectional, batch_first, flag_seq, direction) +> torch_npu.npu_lstm(x, weight, bias, seq_len, h, c, has_biases, num_layers, dropout, train, bidirectional, batch_first, flag_seq, direction) DynamicRNN calculation. @@ -2087,8 +2087,8 @@ DynamicRNN calculation. None ->npu_iou(bboxes, gtboxes, mode=0) -> Tensor ->npu_ptiou(bboxes, gtboxes, mode=0) -> Tensor +>torch_npu.npu_iou(bboxes, gtboxes, mode=0) -> Tensor +>torch_npu.npu_ptiou(bboxes, gtboxes, mode=0) -> Tensor Computes the intersection over union (iou) or the intersection over. foreground (iof) based on the ground-truth and predicted regions. @@ -2117,7 +2117,7 @@ Computes the intersection over union (iou) or the intersection over. foreground [0.0000, 0.9961, 0.0000]], device='npu:0', dtype=torch.float16) ``` ->npu_pad(input, paddings) -> Tensor +>torch_npu.npu_pad(input, paddings) -> Tensor Pads a tensor @@ -2141,7 +2141,7 @@ Pads a tensor [ 0., 0., 0., 0., 0., 0.]], device='npu:0', dtype=torch.float16) ``` ->npu_nms_with_mask(input, iou_threshold) -> (Tensor, Tensor, Tensor) +>torch_npu.npu_nms_with_mask(input, iou_threshold) -> (Tensor, Tensor, Tensor) The value 01 is generated for the nms operator to determine the valid bit @@ -2175,7 +2175,7 @@ The value 01 is generated for the nms operator to determine the valid bit tensor([1, 1], device='npu:0', dtype=torch.uint8) ``` ->npu_bounding_box_encode(anchor_box, ground_truth_box, means0, means1, means2, means3, stds0, stds1, stds2, stds3) -> Tensor +>torch_npu.npu_bounding_box_encode(anchor_box, ground_truth_box, means0, means1, means2, means3, stds0, stds1, stds2, stds3) -> Tensor Computes the coordinate variations between bboxes and ground truth boxes. It is a customized FasterRcnn operator @@ -2207,7 +2207,7 @@ Computes the coordinate variations between bboxes and ground truth boxes. It is >>> ``` ->npu_bounding_box_decode(rois, deltas, means0, means1, means2, means3, stds0, stds1, stds2, stds3, max_shape, wh_ratio_clip) -> Tensor +>torch_npu.npu_bounding_box_decode(rois, deltas, means0, means1, means2, means3, stds0, stds1, stds2, stds3, max_shape, wh_ratio_clip) -> Tensor Generates bounding boxes based on "rois" and "deltas". It is a customized FasterRcnn operator . @@ -2240,7 +2240,7 @@ Generates bounding boxes based on "rois" and "deltas". It is a customized Faster [9.0000, 9.0000, 9.0000, 9.0000]], device='npu:0') ``` ->npu_gru(input, hx, weight_input, weight_hidden, bias_input, bias_hidden, seq_length, has_biases, num_layers, dropout, train, bidirectional, batch_first) -> (Tensor, Tensor, Tensor, Tensor, Tensor, Tensor) +>torch_npu.npu_gru(input, hx, weight_input, weight_hidden, bias_input, bias_hidden, seq_length, has_biases, num_layers, dropout, train, bidirectional, batch_first) -> (Tensor, Tensor, Tensor, Tensor, Tensor, Tensor) DynamicGRUV2 calculation. @@ -2276,7 +2276,7 @@ DynamicGRUV2 calculation. None ->npu_random_choice_with_mask(x, count=256, seed=0, seed2=0) -> (Tensor, Tensor) +>torch_npu.npu_random_choice_with_mask(x, count=256, seed=0, seed2=0) -> (Tensor, Tensor) Shuffle index of no-zero element @@ -2307,7 +2307,7 @@ Shuffle index of no-zero element tensor([True, True], device='npu:0') ``` ->npu_batch_nms(self, scores, score_threshold, iou_threshold, max_size_per_class, max_total_size, change_coordinate_frame=False, transpose_box=False) -> (Tensor, Tensor, Tensor, Tensor) +>torch_npu.npu_batch_nms(self, scores, score_threshold, iou_threshold, max_size_per_class, max_total_size, change_coordinate_frame=False, transpose_box=False) -> (Tensor, Tensor, Tensor, Tensor) Computes nms for input boxes and score, support multiple batch and classes. will do clip to window, score filter, top_k, and nms @@ -2344,7 +2344,7 @@ Computes nms for input boxes and score, support multiple batch and classes. will >>> nmsed_num ``` ->npu_slice(self, offsets, size) -> Tensor +>torch_npu.npu_slice(self, offsets, size) -> Tensor Extracts a slice from a tensor @@ -2369,7 +2369,7 @@ Extracts a slice from a tensor [6., 7.]], device='npu:0', dtype=torch.float16) ``` ->npu_dropoutV2(self, seed, p) -> (Tensor, Tensor, Tensor(a!)) +>torch_npu.npu_dropoutV2(self, seed, p) -> (Tensor, Tensor, Tensor(a!)) count dropout result with seed @@ -2413,7 +2413,7 @@ count dropout result with seed 0.0000, 0.0000, 0.0000, 0.0000, 0.0000], device='npu:0') ``` ->_npu_dropout(self, p) -> (Tensor, Tensor) +>torch_npu._npu_dropout(self, p) -> (Tensor, Tensor) count dropout result without seed @@ -2441,7 +2441,7 @@ count dropout result without seed 253, 255], device='npu:0', dtype=torch.uint8) ``` ->_npu_dropout_inplace(result, p) -> (Tensor(a!), Tensor) +>torch_npu._npu_dropout_inplace(result, p) -> (Tensor(a!), Tensor) count dropout result inplace. @@ -2471,7 +2471,7 @@ count dropout result inplace. 253, 255], device='npu:0', dtype=torch.uint8) ``` ->npu_indexing(self, begin, end, strides, begin_mask=0, end_mask=0, ellipsis_mask=0, new_axis_mask=0, shrink_axis_mask=0) -> Tensor +>torch_npu.npu_indexing(self, begin, end, strides, begin_mask=0, end_mask=0, ellipsis_mask=0, new_axis_mask=0, shrink_axis_mask=0) -> Tensor count indexing result by begin,end,strides array. @@ -2507,7 +2507,7 @@ count indexing result by begin,end,strides array. [5, 6]], device='npu:0', dtype=torch.int32) ``` ->npu_ifmr(Tensor data, Tensor data_min, Tensor data_max, Tensor cumsum, float min_percentile, float max_percentile, float search_start, float search_end, float search_step, bool with_offset) -> (Tensor, Tensor) +>torch_npu.npu_ifmr(Tensor data, Tensor data_min, Tensor data_max, Tensor cumsum, float min_percentile, float max_percentile, float search_start, float search_end, float search_step, bool with_offset) -> (Tensor, Tensor) count ifmr result by begin,end,strides array, Input Feature Map Reconstruction @@ -2597,7 +2597,7 @@ count ifmr result by begin,end,strides array, Input Feature Map Reconstruction tensor(0., device='npu:0') ``` ->npu_max.dim(self, dim, keepdim=False) -> (Tensor, Tensor) +>torch_npu.npu_max.dim(self, dim, keepdim=False) -> (Tensor, Tensor) count max result with dim. @@ -2648,7 +2648,7 @@ count max result with dim. [0, 0]]], device='npu:0', dtype=torch.int32) ``` ->npu_min.dim(self, dim, keepdim=False) -> (Tensor, Tensor) +>torch_npu.npu_min.dim(self, dim, keepdim=False) -> (Tensor, Tensor) count min result with dim. @@ -2698,7 +2698,7 @@ count min result with dim. [0, 1]]], device='npu:0', dtype=torch.int32) ``` ->npu_scatter(self, indices, updates, dim) -> Tensor +>torch_npu.npu_scatter(self, indices, updates, dim) -> Tensor count scatter result with dim. @@ -2734,7 +2734,7 @@ count scatter result with dim. [ 0.9041, -1.5247]], device='npu:0') ``` ->npu_layer_norm_eval(input, normalized_shape, weight=None, bias=None, eps=1e-05) -> Tensor +>torch_npu.npu_layer_norm_eval(input, normalized_shape, weight=None, bias=None, eps=1e-05) -> Tensor count layer norm result. @@ -2780,7 +2780,7 @@ count layer norm result. [ nan, -6.2792e-42, 1.7902e-20, 2.1050e-40]], device='npu:0') ``` ->npu_alloc_float_status(self) -> Tensor +>torch_npu.npu_alloc_float_status(self) -> Tensor Produces eight numbers with a value of zero @@ -2804,7 +2804,7 @@ Produces eight numbers with a value of zero tensor([0., 0., 0., 0., 0., 0., 0., 0.], device='npu:0') ``` -> npu_get_float_status(self) -> Tensor +> torch_npu.npu_get_float_status(self) -> Tensor Computes NPU get float status operator function. @@ -2824,7 +2824,7 @@ Computes NPU get float status operator function. tensor([0., 0., 0., 0., 0., 0., 0., 0.], device='npu:0') ``` -> npu_clear_float_status(self) -> Tensor +> torch_npu.npu_clear_float_status(self) -> Tensor Set the value of address 0x40000 to 0 in each core. @@ -2844,7 +2844,7 @@ Set the value of address 0x40000 to 0 in each core. tensor([0., 0., 0., 0., 0., 0., 0., 0.], device='npu:0') ``` -> npu_confusion_transpose(self, perm, shape, transpose_first) -> Tensor +> torch_npu.npu_confusion_transpose(self, perm, shape, transpose_first) -> Tensor Confuse reshape and transpose. @@ -2873,7 +2873,7 @@ Confuse reshape and transpose. torch.Size([2, 6, 12]) ``` -> npu_bmmV2(self, mat2, output_sizes) -> Tensor +> torch_npu.npu_bmmV2(self, mat2, output_sizes) -> Tensor Multiplies matrix "a" by matrix "b", producing "a * b" . @@ -2896,7 +2896,7 @@ Multiplies matrix "a" by matrix "b", producing "a * b" . torch.Size([10, 3, 5]) ``` -> fast_gelu(self) -> Tensor +> torch_npu.fast_gelu(self) -> Tensor Computes the gradient for the fast_gelu of "x" . @@ -2918,7 +2918,7 @@ Computes the gradient for the fast_gelu of "x" . tensor([0.4403, 0.2733], device='npu:0') ``` -> npu_sub_sample(self, per_images, positive_fraction) -> Tensor +> torch_npu.npu_sub_sample(self, per_images, positive_fraction) -> Tensor Randomly sample a subset of positive and negative examples,and overwrite the label vector to the ignore value (-1) for all elements that are not included in the sample. @@ -2944,7 +2944,7 @@ Randomly sample a subset of positive and negative examples,and overwrite the lab dtype=torch.int32) ``` -> npu_deformable_conv2d(input, weight, offset, bias, kernel_size, stride, padding, dilation=[1,1,1,1], groups=1, deformable_groups=1, modulated=True) -> (Tensor, Tensor) +> torch_npu.npu_deformable_conv2d(input, weight, offset, bias, kernel_size, stride, padding, dilation=[1,1,1,1], groups=1, deformable_groups=1, modulated=True) -> (Tensor, Tensor) Computes the deformed convolution output with the expected input. @@ -2977,7 +2977,7 @@ Computes the deformed convolution output with the expected input. torch.Size([16, 32, 32, 32]) ``` -> npu_mish(self) -> Tensor +> torch_npu.npu_mish(self) -> Tensor Computes hyperbolic tangent of "x" element-wise. @@ -2998,7 +2998,7 @@ Computes hyperbolic tangent of "x" element-wise. torch.Size([10, 30, 10]) ``` -> npu_anchor_response_flags(self, featmap_size, stride, num_base_anchors) -> Tensor +> torch_npu.npu_anchor_response_flags(self, featmap_size, stride, num_base_anchors) -> Tensor Generate the responsible flags of anchor in a single feature map. @@ -3021,7 +3021,7 @@ Generate the responsible flags of anchor in a single feature map. torch.Size([32400]) ``` -> npu_yolo_boxes_encode(self, gt_bboxes, stride, performance_mode=False) -> Tensor +> torch_npu.npu_yolo_boxes_encode(self, gt_bboxes, stride, performance_mode=False) -> Tensor Generates bounding boxes based on yolo's "anchor" and "ground-truth" boxes. It is a customized mmdetection operator. @@ -3046,7 +3046,7 @@ Generates bounding boxes based on yolo's "anchor" and "ground-truth" boxes. It i torch.Size([2, 4]) ``` -> npu_grid_assign_positive(self, overlaps, box_responsible_flags, max_overlaps, argmax_overlaps, gt_max_overlaps, gt_argmax_overlaps, num_gts, pos_iou_thr, min_pos_iou, gt_max_assign_all) -> Tensor +> torch_npu.npu_grid_assign_positive(self, overlaps, box_responsible_flags, max_overlaps, argmax_overlaps, gt_max_overlaps, gt_argmax_overlaps, num_gts, pos_iou_thr, min_pos_iou, gt_max_assign_all) -> Tensor Performs Position Sensitive PS ROI Pooling Grad. @@ -3082,7 +3082,7 @@ Performs Position Sensitive PS ROI Pooling Grad. torch.Size([4]) ``` -> npu_normalize_batch(self, seq_len, normalize_type=0) -> Tensor +> torch_npu.npu_normalize_batch(self, seq_len, normalize_type=0) -> Tensor Performs batch normalization . @@ -3114,7 +3114,7 @@ Performs Position Sensitive PS ROI Pooling Grad. device='npu:0') ``` -> npu_masked_fill_range(self, start, end, value, axis=-1) -> Tensor +> torch_npu.npu_masked_fill_range(self, start, end, value, axis=-1) -> Tensor masked fill tensor along with one axis by range.boxes. It is a customized masked fill range operator . @@ -3149,7 +3149,7 @@ masked fill tensor along with one axis by range.boxes. It is a customized masked [0.0366, 0.9738, 0.4689, 0.0979]], device='npu:0') ``` -> npu_linear(input, weight, bias=None) -> Tensor +> torch_npu.npu_linear(input, weight, bias=None) -> Tensor Multiplies matrix "a" by matrix "b", producing "a * b" . @@ -3174,7 +3174,7 @@ masked fill tensor along with one axis by range.boxes. It is a customized masked [5.3273, 6.3089, 3.9601, 3.2410]], device='npu:0') ``` -> npu_bert_apply_adam.old(Tensor(a!) var, Tensor(b!) m, Tensor(c!) v, lr, beta1, beta2, epsilon, grad, max_grad_norm, global_grad_norm, weight_decay, step_size=None, adam_mode=0) -> (Tensor(a!), Tensor(b!), Tensor(c!)) +> torch_npu.npu_bert_apply_adam.old(Tensor(a!) var, Tensor(b!) m, Tensor(c!) v, lr, beta1, beta2, epsilon, grad, max_grad_norm, global_grad_norm, weight_decay, step_size=None, adam_mode=0) -> (Tensor(a!), Tensor(b!), Tensor(c!)) count adam result. @@ -3215,7 +3215,7 @@ masked fill tensor along with one axis by range.boxes. It is a customized masked device='npu:0') ``` -> npu_bert_apply_adam(lr, beta1, beta2, epsilon, grad, max_grad_norm, global_grad_norm, weight_decay, step_size=None, adam_mode=0, *, out=（var,m,v）) +> torch_npu.npu_bert_apply_adam(lr, beta1, beta2, epsilon, grad, max_grad_norm, global_grad_norm, weight_decay, step_size=None, adam_mode=0, *, out=（var,m,v）) count adam result. @@ -3260,7 +3260,7 @@ masked fill tensor along with one axis by range.boxes. It is a customized masked device='npu:0') ``` -> npu_giou(self, gtboxes, trans=False, is_cross=False, mode=0) -> Tensor +> torch_npu.npu_giou(self, gtboxes, trans=False, is_cross=False, mode=0) -> Tensor First calculate the minimum closure area of the two boxes, IoU, the proportion of the closed area that does not belong to the two boxes in the closure area, and finally subtract this proportion from IoU to get GIoU . @@ -3296,7 +3296,7 @@ First calculate the minimum closure area of the two boxes, IoU, the proportion o [1.]], device='npu:0', dtype=torch.float16) ``` -> npu_silu(self) -> Tensor +> torch_npu.npu_silu(self) -> Tensor Computes the for the Swish of "x" . @@ -3318,7 +3318,7 @@ tensor([[0.4397, 0.7178, 0.5190, 0.2654, 0.2230, 0.2674, 0.6051, 0.3522], device='npu:0') ``` -> npu_reshape(self, shape, bool can_refresh=False) -> Tensor +> torch_npu.npu_reshape(self, shape, bool can_refresh=False) -> Tensor Reshapes a tensor. Only the tensor shape is changed, without changing the data. @@ -3343,7 +3343,7 @@ Reshapes a tensor. Only the tensor shape is changed, without changing the data. [0.1009, 0.7133, 0.8118, 0.6193]], device='npu:0') ``` -> npu_rotated_overlaps(self, query_boxes, trans=False) -> Tensor +> torch_npu.npu_rotated_overlaps(self, query_boxes, trans=False) -> Tensor Calculate the overlapping area of the rotated box. @@ -3370,7 +3370,7 @@ Calculate the overlapping area of the rotated box. [0.0000, 0.0611, 0.0000]]], device='npu:0', dtype=torch.float16) ``` -> npu_rotated_iou(self, query_boxes, trans=False, mode=0, is_cross=True) -> Tensor +> torch_npu.npu_rotated_iou(self, query_boxes, trans=False, mode=0, is_cross=True) -> Tensor Calculate the IOU of the rotated box. @@ -3401,7 +3401,7 @@ Calculate the IOU of the rotated box. [0.0000e+00, 5.9605e-08]]], device='npu:0', dtype=torch.float16) ``` -> npu_rotated_box_encode(anchor_box, gt_bboxes, weight) -> Tensor +> torch_npu.npu_rotated_box_encode(anchor_box, gt_bboxes, weight) -> Tensor Rotate Bounding Box Encoding. @@ -3430,7 +3430,7 @@ Rotate Bounding Box Encoding. [ 1.1328]]], device='npu:0', dtype=torch.float16) ``` - > npu_rotated_box_decode(anchor_boxes, deltas, weight) -> Tensor + > torch_npu.npu_rotated_box_decode(anchor_boxes, deltas, weight) -> Tensor Rotate Bounding Box Encoding -- Gitee From 8aac7af89967d17ed688ab118fbd1b12cebdc1ed Mon Sep 17 00:00:00 2001 From: pengzirong Date: Wed, 16 Mar 2022 15:15:42 +0800 Subject: [PATCH 2/2] =?UTF-8?q?1.8.1API=E4=B8=8D=E6=94=AF=E6=8C=81?= =?UTF-8?q?=E5=86=85=E5=AE=B9=E8=B0=83=E6=95=B4=EF=BC=8CAMP=E6=94=AF?= =?UTF-8?q?=E6=8C=81=E5=BA=A6=E8=B0=83=E6=95=B4,=E4=BF=AE=E6=94=B9?= =?UTF-8?q?=E3=80=8APytorch=E9=80=82=E9=85=8D=E7=AE=97=E5=AD=90=E6=B8=85?= =?UTF-8?q?=E5=8D=95=E3=80=8B-->=E3=80=8APytorch=20API=20=E6=94=AF?= =?UTF-8?q?=E6=8C=81=E6=B8=85=E5=8D=95=E3=80=8B=20=E4=BF=AE=E6=94=B9?= =?UTF-8?q?=E6=A8=A1=E5=9E=8B=E8=BF=81=E7=A7=BB=E9=83=A8=E5=88=86=E6=A8=A1?= =?UTF-8?q?=E5=9E=8B=E6=94=AF=E6=8C=81=E5=88=97=E8=A1=A8=E4=B8=BAModelzoo?= =?UTF-8?q?=E5=AF=B9=E5=BA=94=E9=93=BE=E6=8E=A5=20=E6=8F=90=E4=BE=9B?= =?UTF-8?q?=E3=80=8AMindStudio=20=E7=94=A8=E6=88=B7=E6=8C=87=E5=8D=97?= =?UTF-8?q?=E3=80=8B=E3=80=81=E3=80=8ACANN=E8=BD=AF=E4=BB=B6=E5=AE=89?= =?UTF-8?q?=E8=A3=85=20=E6=8C=87=E5=8D=97=E3=80=8B=E5=AF=B9=E5=BA=94?= =?UTF-8?q?=E8=B7=B3=E8=BD=AC=E9=93=BE=E6=8E=A5?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 修改模型迁移部分模型支持列表为Modelzoo对应链接 提供《MindStudio 用户指南》、《CANN软件安装指南》对应跳转链接 修改《网络模型移植&训练指南》采集训练过程章节，解析性能数据文件描述;删除环境变量 TRI_COMBINED_ENABLE=1 delete 环境变量 TRI_COMBINED_ENABLE=1 update --- ...57\346\214\201\346\270\205\345\215\225.md" | 1606 +++++++++-------- ...11\350\243\205\346\214\207\345\215\227.md" | 9 +- ...55\347\273\203\346\214\207\345\215\227.md" | 298 +-- 3 files changed, 815 insertions(+), 1098 deletions(-) diff --git "a/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" "b/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" index 9fd2577488..76590e1b60 100644 --- "a/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" +++ "b/docs/zh/PyTorch API\346\224\257\346\214\201\346\270\205\345\215\225.md" @@ -6,7 +6,7 @@ | 2 | torch.is_storage | 是 | | 3 | torch.is_complex | 是，支持判断，但当前硬件限制不支持复数 | | 4 | torch.is_floating_point | 是 | -| 5 | torch.is_nonzero | 是 | +| 5 | torch.is_nonzero | 否 | | 6 | torch.set_default_dtype | 是 | | 7 | torch.get_default_dtype | 是 | | 8 | torch.set_default_tensor_type | 是 | @@ -35,85 +35,76 @@ | 31 | torch.quantize_per_tensor | 是 | | 32 | torch.quantize_per_channel | 是 | | 33 | torch.dequantize | 是 | -| 34 | torch.complex | 是 | -| 35 | torch.polar | 是 | -| 36 | torch.heaviside | 是 | +| 34 | torch.complex | 否 | +| 35 | torch.polar | 否 | +| 36 | torch.heaviside | 否 | | 37 | torch.cat | 是 | | 38 | torch.chunk | 是 | -| 39 | torch.column_stack | 是 | -| 40 | torch.dstack | 是 | -| 41 | torch.hstack | 是 | +| 39 | torch.column_stack | 否 | +| 40 | torch.dstack | 否 | +| 41 | torch.hstack | 否 | | 42 | torch.gather | 是 | | 43 | torch.index_select | 是 | | 44 | torch.masked_select | 是 | -| 45 | torch.movedim | 是 | -| 46 | torch.moveaxis | 是 | +| 45 | torch.movedim | 否 | +| 46 | torch.moveaxis | 否 | | 47 | torch.narrow | 是 | | 48 | torch.nonzero | 是 | | 49 | torch.reshape | 是 | -| 50 | torch.row_stack | 是 | +| 50 | torch.row_stack | 否 | | 51 | torch.scatter | 是 | | 52 | torch.scatter_add | 是 | | 53 | torch.split | 是 | | 54 | torch.squeeze | 是 | | 55 | torch.stack | 是 | -| 56 | torch.swapaxes | 是 | -| 57 | torch.swapdims | 是 | +| 56 | torch.swapaxes | 否 | +| 57 | torch.swapdims | 否 | | 58 | torch.t | 是 | | 59 | torch.take | 是 | -| 60 | torch.tensor_split | 是 | -| 61 | torch.tile | 是 | +| 60 | torch.tensor_split | 否 | +| 61 | torch.tile | 否 | | 62 | torch.transpose | 是 | | 63 | torch.unbind | 是 | | 64 | torch.unsqueeze | 是 | -| 65 | torch.vstack | 是 | +| 65 | torch.vstack | 否 | | 66 | torch.where | 是 | ## Generators -| 序号 | API名称 | 是否支持 | -| ---- | ------------------------------- | -------- | -| 1 | torch._C.Generator | 是 | -| 2 | torch._C.Generator.device | 是 | -| 3 | torch._C.Generator.get_state | 否 | -| 4 | torch._C.Generator.initial_seed | 是 | -| 5 | torch._C.Generator.manual_seed | 是 | -| 6 | torch._C.Generator.seed | 是 | -| 7 | torch._C.Generator.set_state | 否 | +| 序号 | API名称 | 是否支持 | +| ---- | -------------------------------- | -------- | +| 1 | torch._C.Generator | 是 | +| 2 | torch._C.torch.default_generator | 是 | ## Random sampling -| 序号 | API名称 | 是否支持 | -| ---- | ------------------------------------------ | -------- | -| 1 | torch.seed | 是 | -| 2 | torch.manual_seed | 是 | -| 3 | torch.initial_seed | 是 | -| 4 | torch.get_rng_state | 是 | -| 5 | torch.set_rng_state | 是 | -| 6 | torch.torch.default_generator | 是 | -| 7 | torch.bernoulli | 是 | -| 8 | torch.multinomial | 是 | -| 9 | torch.normal | 是 | -| 10 | torch.poisson | 否 | -| 11 | torch.rand | 是 | -| 12 | torch.rand_like | 是 | -| 13 | torch.randint | 是 | -| 14 | torch.randint_like | 是 | -| 15 | torch.randn | 是 | -| 16 | torch.randn_like | 是 | -| 17 | torch.randperm | 是 | -| 18 | torch.Tensor.bernoulli_() | 是 | -| 19 | torch.Tensor.cauchy_() | 是 | -| 20 | torch.Tensor.exponential_() | 否 | -| 21 | torch.Tensor.geometric_() | 否 | -| 22 | torch.Tensor.log_normal_() | 否 | -| 23 | torch.Tensor.normal_() | 是 | -| 24 | torch.Tensor.random_() | 是 | -| 25 | torch.Tensor.uniform_() | 是 | -| 26 | torch.quasirandom.SobolEngine | 是 | -| 27 | torch.quasirandom.SobolEngine.draw | 是 | -| 28 | torch.quasirandom.SobolEngine.fast_forward | 是 | -| 29 | torch.quasirandom.SobolEngine.reset | 是 | +| 序号 | API名称 | 是否支持 | +| ---- | ----------------------------- | -------- | +| 1 | torch.seed | 是 | +| 2 | torch.manual_seed | 是 | +| 3 | torch.initial_seed | 是 | +| 4 | torch.get_rng_state | 是 | +| 5 | torch.set_rng_state | 是 | +| 7 | torch.bernoulli | 是 | +| 8 | torch.multinomial | 是 | +| 9 | torch.normal | 是 | +| 10 | torch.poisson | 否 | +| 11 | torch.rand | 是 | +| 12 | torch.rand_like | 是 | +| 13 | torch.randint | 是 | +| 14 | torch.randint_like | 是 | +| 15 | torch.randn | 是 | +| 16 | torch.randn_like | 是 | +| 17 | torch.randperm | 是 | +| 18 | torch.Tensor.bernoulli_() | 是 | +| 19 | torch.Tensor.cauchy_() | 是 | +| 20 | torch.Tensor.exponential_() | 否 | +| 21 | torch.Tensor.geometric_() | 否 | +| 22 | torch.Tensor.log_normal_() | 否 | +| 23 | torch.Tensor.normal_() | 是 | +| 24 | torch.Tensor.random_() | 是 | +| 25 | torch.Tensor.uniform_() | 是 | +| 26 | torch.quasirandom.SobolEngine | 是 | ## Serialization @@ -127,259 +118,262 @@ | 序号 | API名称 | 是否支持 | | ---- | -------------------------------------- | -------- | | 1 | torch.abs | 是 | -| 2 | torch.absolute | 是 | +| 2 | torch.absolute | 否 | | 3 | torch.acos | 是 | -| 4 | torch.arccos | 是 | -| 5 | torch.acosh | 是 | -| 6 | torch.arccosh | 是 | +| 4 | torch.arccos | 否 | +| 5 | torch.acosh | 否 | +| 6 | torch.arccosh | 否 | | 7 | torch.add | 是 | | 8 | torch.addcdiv | 是 | | 9 | torch.addcmul | 是 | | 10 | torch.angle | 否 | | 11 | torch.asin | 是 | -| 12 | torch.arcsin | 是 | -| 13 | torch.sinh | 是 | -| 14 | torch.arcsinh | 是 | +| 12 | torch.arcsin | 否 | +| 13 | torch.asinh | 否 | +| 14 | torch.arcsinh | 否 | | 15 | torch.atan | 是 | -| 16 | torch.atanh | 是 | -| 17 | torch.arctanh | 是 | -| 18 | torch.atan2 | 是 | -| 19 | torch.bitwise_not | 是 | -| 20 | torch.bitwise_and | 是 | -| 21 | torch.bitwise_or | 是 | -| 22 | torch.bitwise_xor | 是 | -| 23 | torch.ceil | 是 | -| 24 | torch.clamp | 是 | -| 25 | torch.clip | 是 | -| 26 | torch.conj | 否 | -| 27 | torch.copysign | 是 | -| 28 | torch.cos | 是 | -| 29 | torch.cosh | 是 | -| 30 | torch.deg2rad | 是 | -| 31 | torch.div | 是 | -| 32 | torch.divide | 是 | -| 33 | torch.digamma | 否 | -| 34 | torch.erf | 是 | -| 35 | torch.erfc | 是 | -| 36 | torch.erfinv | 是 | -| 37 | torch.exp | 是 | -| 38 | torch.exp2 | | -| 39 | torch.expm1 | 是 | -| 40 | torch.fake_quantize_per_channel_affine | 是 | -| 41 | torch.fake_quantize_per_tensor_affine | 是 | -| 42 | torch.fix | 是 | -| 43 | torch.float_power | 是 | -| 44 | torch.floor | 是 | -| 45 | torch.floor_divide | 是 | -| 46 | torch.fmod | 是 | -| 47 | torch.frac | 是 | -| 48 | torch.imag | 否 | -| 49 | torch.ldexp | 是 | -| 50 | torch.lerp | 是 | -| 51 | torch.lgamma | 否 | -| 52 | torch.log | 是 | -| 53 | torch.log10 | 是 | -| 54 | torch.log1p | 是 | -| 55 | torch.log2 | 是 | -| 56 | torch.logaddexp | 是 | -| 57 | torch.logaddexp2 | 是 | -| 58 | torch.logical_and | 是 | -| 59 | torch.logical_not | 是 | -| 60 | torch.logical_or | 是 | -| 61 | torch.logical_xor | 是 | -| 62 | torch.logit | 是 | -| 63 | torch.hypot | 是 | -| 64 | torch.i0 | 是 | -| 65 | torch.igamma | 是 | -| 66 | torch.igammac | 是 | -| 67 | torch.mul | 是 | -| 68 | torch.multiply | 是 | -| 69 | torch.mvlgamma | 否 | -| 70 | torch.nan_to_num | 是 | -| 71 | torch.neg | 是 | -| 72 | torch.negative | 是 | -| 73 | torch.nextafter | 是 | -| 74 | torch.polygamma | 否 | -| 75 | torch.pow | 是 | -| 76 | torch.rad2deg | 是 | -| 77 | torch.real | 是 | -| 78 | torch.reciprocal | 是 | -| 79 | torch.remainder | 是 | -| 80 | torch.round | 是 | -| 81 | torch.rsqrt | 是 | -| 82 | torch.sigmoid | 是 | -| 83 | torch.sign | 是 | -| 84 | torch.sgn | 是 | -| 85 | torch.signbit | 是 | -| 86 | torch.sin | 是 | -| 87 | torch.sinc | 是 | -| 88 | torch.sinh | 是 | -| 89 | torch.sqrt | 是 | -| 90 | torch.square | 是 | -| 91 | torch.sub | 是 | -| 92 | torch.subtract | 是 | -| 93 | torch.tan | 是 | -| 94 | torch.tanh | 是 | -| 95 | torch.true_divide | 是 | -| 96 | torch.trunc | 是 | -| 97 | torch.xlogy | 是 | -| 98 | torch.argmax | 是 | -| 99 | torch.argmin | 是 | -| 100 | torch.amax | 是 | -| 101 | torch.amin | 是 | -| 102 | torch.all | 是 | -| 103 | torch.any | 是 | -| 104 | torch.max | 是 | -| 105 | torch.min | 是 | -| 106 | torch.dist | 是 | -| 107 | torch.logsumexp | 是 | -| 108 | torch.mean | 是 | -| 109 | torch.median | 是 | -| 110 | torch.namedian | 是 | -| 111 | torch.mode | 否 | -| 112 | torch.norm | 是 | -| 113 | torch.nansum | 是 | -| 114 | torch.prod | 是 | -| 115 | torch.quantile | 是 | -| 116 | torch.nanquantile | 是 | -| 117 | torch.std | 是 | -| 118 | torch.std_mean | 是 | -| 119 | torch.sum | 是 | -| 120 | torch.unique | 是 | -| 121 | torch.unique_consecutive | 否 | -| 122 | torch.var | 否 | -| 123 | torch.var_mean | 否 | -| 124 | torch.count_nonzero | 是 | -| 125 | torch.allclose | 是 | -| 126 | torch.argsort | 是 | -| 127 | torch.eq | 是 | -| 128 | torch.equal | 是 | -| 129 | torch.ge | 是 | -| 130 | torch.greater_qual | 是 | -| 131 | torch.gt | 是 | -| 132 | torch.greater | 是 | -| 133 | torch.isclose | 是 | -| 134 | torch.isfinite | 是 | -| 135 | torch.isinf | 是 | -| 136 | torch.isposinf | 是 | -| 137 | torch.isneginf | 是 | -| 138 | torch.isnan | 是 | -| 139 | torch.isreal | 是 | -| 140 | torch.kthvalue | 是 | -| 141 | torch.le | 是 | -| 142 | torch.less_qual | 是 | -| 143 | torch.lt | 是 | -| 144 | torch.less | 是 | -| 145 | torch.maximum | 是 | -| 146 | torch.minimum | 是 | -| 147 | torch.fmax | 是 | -| 148 | torch.fmin | 是 | -| 149 | torch.ne | 是 | -| 150 | torch.not_equal | 是 | -| 151 | torch.sort | 是 | -| 152 | torch.topk | 是 | -| 153 | torch.msort | 是 | -| 154 | torch.fft | 否 | -| 155 | torch.ifft | 否 | -| 156 | torch.rfft | 否 | -| 157 | torch.irfft | 否 | -| 158 | torch.stft | 否 | -| 159 | torch.istft | 是 | -| 160 | torch.bartlett_window | 是 | -| 161 | torch.blackman_window | 是 | -| 162 | torch.hamming_window | 是 | -| 163 | torch.hann_window | 是 | -| 164 | torch.atleast_1d | 是 | -| 165 | torch.atleast_2d | 是 | -| 166 | torch.atleast_3d | 是 | -| 167 | torch.bincount | 是 | -| 168 | torch.block_diag | 是 | -| 169 | torch.broadcast_tensors | 是 | -| 170 | torch.broadcast_to | 是 | -| 171 | torch.broadcast_shapes | 是 | -| 172 | torch.bucketize | 是 | -| 173 | torch.cartesian_prod | 是 | -| 174 | torch.cdist | 是 | -| 175 | torch.clone | 是 | -| 176 | torch.combinations | 否 | -| 177 | torch.cross | 是 | -| 178 | torch.cummax | 是 | -| 179 | torch.cummin | 是 | -| 180 | torch.cumprod | 是 | -| 181 | torch.cumsum | 是 | -| 182 | torch.diag | 是 | -| 183 | torch.diag_embed | 是 | -| 184 | torch.diagflat | 是 | -| 185 | torch.diagonal | 是 | -| 186 | torch.diff | 是 | -| 187 | torch.einsum | 是 | -| 188 | torch.flatten | 是 | -| 189 | torch.flip | 是 | -| 190 | torch.fliplr | 是 | -| 191 | torch.flipud | 是 | -| 192 | torch.kron | 是 | -| 193 | torch.rot90 | 是 | -| 194 | torch.gcd | 是 | -| 195 | torch.histc | 否 | -| 196 | torch.meshgrid | 是 | -| 197 | torch.lcm | 是 | -| 198 | torhc.logcumsumexp | 是 | -| 199 | torch.ravel | 是 | -| 200 | torch.renorm | 是 | -| 201 | torch.repeat_interleave | 是 | -| 202 | torch.roll | 是 | -| 203 | torch.searchsorted | 是 | -| 204 | torch.tensordot | 是 | -| 205 | torch.trace | 否 | -| 206 | torch.tril | 是 | -| 207 | torch.tril_indices | 是 | -| 208 | torch.triu | 是 | -| 209 | torch.triu_indices | 是 | -| 210 | torch.view_as_real | 是 | -| 211 | torch.view_as_complex | 是 | -| 212 | torch.addbmm | 是 | -| 213 | torch.addmm | 是 | -| 214 | torch.addmv | 是 | -| 215 | torch.addr | 是 | -| 216 | torch.baddbmm | 是 | -| 217 | torch.bmm | 是 | -| 218 | torch.chain_matmul | 是 | -| 219 | torch.cholesky | 否 | -| 220 | torch.cholesky_inverse | 否 | -| 221 | torch.cholesky_solve | 否 | -| 222 | torch.dot | 是 | -| 223 | torch.eig | 否 | -| 224 | torch.geqrf | 否 | -| 225 | torch.ger | 是 | -| 226 | torch.inner | 是 | -| 227 | torch.inverse | 是 | -| 228 | torch.det | 否 | -| 229 | torch.logdet | 否 | -| 230 | torch.slogdet | 是 | -| 231 | torch.lstsq | 否 | -| 232 | torch.lu | 否 | -| 233 | torch.lu_solve | 否 | -| 234 | torch.lu_unpack | 否 | -| 235 | torch.matmul | 是 | -| 236 | torch.matrix_power | 是 | -| 237 | torch.matrix_rank | 是 | -| 238 | torch.matrix_exp | 是 | -| 239 | torch.mm | 是 | -| 240 | torch.mv | 是 | -| 241 | torch.orgqr | 否 | -| 242 | torch.ormqr | 否 | -| 243 | torch.outer | 是 | -| 244 | torch.pinverse | 是 | -| 245 | torch.qr | 是 | -| 246 | torch.solve | 否 | -| 247 | torch.svd | 是 | -| 248 | torch.svd_lowrank | 是 | -| 249 | torch.pca_lowrank | 是 | -| 250 | torch.symeig | 是 | -| 251 | torch.lobpcg | 否 | -| 252 | torch.trapz | 是 | -| 253 | torch.triangular_solve | 是 | -| 254 | torch.vdot | 是 | +| 16 | torch.arctan | 否 | +| 17 | torch.atanh | 否 | +| 18 | torch.arctanh | 否 | +| 19 | torch.atan2 | 是 | +| 20 | torch.bitwise_not | 是 | +| 21 | torch.bitwise_and | 是 | +| 22 | torch.bitwise_or | 是 | +| 23 | torch.bitwise_xor | 是 | +| 24 | torch.ceil | 是 | +| 25 | torch.clamp | 是 | +| 26 | torch.clip | 否 | +| 27 | torch.conj | 否 | +| 28 | torch.copysign | 否 | +| 29 | torch.cos | 是 | +| 30 | torch.cosh | 是 | +| 31 | torch.deg2rad | 否 | +| 32 | torch.div | 是 | +| 33 | torch.divide | 否 | +| 34 | torch.digamma | 否 | +| 35 | torch.erf | 是 | +| 36 | torch.erfc | 是 | +| 37 | torch.erfinv | 是 | +| 38 | torch.exp | 是 | +| 39 | torch.exp2 | 否 | +| 40 | torch.expm1 | 是 | +| 41 | torch.fake_quantize_per_channel_affine | 否 | +| 42 | torch.fake_quantize_per_tensor_affine | 否 | +| 43 | torch.fix | 否 | +| 44 | torch.float_power | 否 | +| 45 | torch.floor | 是 | +| 46 | torch.floor_divide | 是 | +| 47 | torch.fmod | 是 | +| 48 | torch.frac | 是 | +| 49 | torch.imag | 否 | +| 50 | torch.ldexp | 否 | +| 51 | torch.lerp | 是 | +| 52 | torch.lgamma | 否 | +| 53 | torch.log | 是 | +| 54 | torch.log10 | 是 | +| 55 | torch.log1p | 是 | +| 56 | torch.log2 | 是 | +| 57 | torch.logaddexp | 否 | +| 58 | torch.logaddexp2 | 否 | +| 59 | torch.logical_and | 是 | +| 60 | torch.logical_not | 是 | +| 61 | torch.logical_or | 是 | +| 62 | torch.logical_xor | 是 | +| 63 | torch.logit | 否 | +| 64 | torch.hypot | 否 | +| 65 | torch.i0 | 否 | +| 66 | torch.igamma | 否 | +| 67 | torch.igammac | 否 | +| 68 | torch.mul | 是 | +| 69 | torch.multiply | 否 | +| 70 | torch.mvlgamma | 否 | +| 71 | torch.nan_to_num | 否 | +| 72 | torch.neg | 是 | +| 73 | torch.negative | 否 | +| 74 | torch.nextafter | 否 | +| 75 | torch.polygamma | 否 | +| 76 | torch.pow | 是 | +| 77 | torch.rad2deg | 否 | +| 78 | torch.real | 是 | +| 79 | torch.reciprocal | 是 | +| 80 | torch.remainder | 是 | +| 81 | torch.round | 是 | +| 82 | torch.rsqrt | 是 | +| 83 | torch.sigmoid | 是 | +| 84 | torch.sign | 是 | +| 85 | torch.sgn | 否 | +| 86 | torch.signbit | 否 | +| 87 | torch.sin | 是 | +| 88 | torch.sinc | 否 | +| 89 | torch.sinh | 是 | +| 90 | torch.sqrt | 是 | +| 91 | torch.square | 是 | +| 92 | torch.sub | 是 | +| 93 | torch.subtract | 否 | +| 94 | torch.tan | 是 | +| 95 | torch.tanh | 是 | +| 96 | torch.true_divide | 是 | +| 97 | torch.trunc | 是 | +| 98 | torch.xlogy | 否 | +| 99 | torch.argmax | 是 | +| 100 | torch.argmin | 是 | +| 101 | torch.amax | 否 | +| 102 | torch.amin | 否 | +| 103 | torch.all | 是 | +| 104 | torch.any | 是 | +| 105 | torch.max | 是 | +| 106 | torch.min | 是 | +| 107 | torch.dist | 是 | +| 108 | torch.logsumexp | 是 | +| 109 | torch.mean | 是 | +| 110 | torch.median | 是 | +| 111 | torch.namedian | 否 | +| 112 | torch.mode | 否 | +| 113 | torch.norm | 是 | +| 114 | torch.nansum | 否 | +| 115 | torch.prod | 是 | +| 116 | torch.quantile | 否 | +| 117 | torch.nanquantile | 否 | +| 118 | torch.std | 是 | +| 119 | torch.std_mean | 是 | +| 120 | torch.sum | 是 | +| 121 | torch.unique | 是 | +| 122 | torch.unique_consecutive | 否 | +| 123 | torch.var | 否 | +| 124 | torch.var_mean | 否 | +| 125 | torch.count_nonzero | 否 | +| 126 | torch.allclose | 是 | +| 127 | torch.argsort | 是 | +| 128 | torch.eq | 是 | +| 129 | torch.equal | 是 | +| 130 | torch.ge | 是 | +| 131 | torch.greater_qual | 否 | +| 132 | torch.gt | 是 | +| 133 | torch.greater | 否 | +| 134 | torch.isclose | 否 | +| 135 | torch.isfinite | 是 | +| 136 | torch.isinf | 是 | +| 137 | torch.isposinf | 否 | +| 138 | torch.isneginf | 否 | +| 139 | torch.isnan | 是 | +| 140 | torch.isreal | 否 | +| 141 | torch.kthvalue | 是 | +| 142 | torch.le | 是 | +| 143 | torch.less_qual | 否 | +| 144 | torch.lt | 是 | +| 145 | torch.less | 否 | +| 146 | torch.maximum | 否 | +| 147 | torch.minimum | 否 | +| 148 | torch.fmax | 否 | +| 149 | torch.fmin | 否 | +| 150 | torch.ne | 是 | +| 151 | torch.not_equal | 否 | +| 152 | torch.sort | 是 | +| 153 | torch.topk | 是 | +| 154 | torch.msort | 否 | +| 155 | torch.fft | 否 | +| 156 | torch.ifft | 否 | +| 157 | torch.rfft | 否 | +| 158 | torch.irfft | 否 | +| 159 | torch.stft | 否 | +| 160 | torch.istft | 否 | +| 161 | torch.bartlett_window | 是 | +| 162 | torch.blackman_window | 是 | +| 163 | torch.hamming_window | 是 | +| 164 | torch.hann_window | 是 | +| 165 | torch.kasier_window | 否 | +| 166 | torch.atleast_1d | 否 | +| 167 | torch.atleast_2d | 否 | +| 168 | torch.atleast_3d | 否 | +| 169 | torch.bincount | 是 | +| 170 | torch.block_diag | 否 | +| 171 | torch.broadcast_tensors | 是 | +| 172 | torch.broadcast_to | 否 | +| 173 | torch.broadcast_shapes | 否 | +| 174 | torch.bucketize | 否 | +| 175 | torch.cartesian_prod | 是 | +| 176 | torch.cdist | 是 | +| 177 | torch.clone | 是 | +| 178 | torch.combinations | 否 | +| 179 | torch.cross | 是 | +| 180 | torch.cummax | 是 | +| 181 | torch.cummin | 是 | +| 182 | torch.cumprod | 是 | +| 183 | torch.cumsum | 是 | +| 184 | torch.diag | 是 | +| 185 | torch.diag_embed | 是 | +| 186 | torch.diagflat | 是 | +| 187 | torch.diagonal | 是 | +| 188 | torch.diff | 否 | +| 189 | torch.einsum | 是 | +| 190 | torch.flatten | 是 | +| 191 | torch.flip | 是 | +| 192 | torch.fliplr | 否 | +| 193 | torch.flipud | 否 | +| 194 | torch.kron | 否 | +| 195 | torch.rot90 | 是 | +| 196 | torch.gcd | 否 | +| 197 | torch.histc | 否 | +| 198 | torch.meshgrid | 是 | +| 199 | torch.lcm | 否 | +| 200 | torhc.logcumsumexp | 否 | +| 201 | torch.ravel | 否 | +| 202 | torch.renorm | 是 | +| 203 | torch.repeat_interleave | 是 | +| 204 | torch.roll | 是 | +| 205 | torch.searchsorted | 否 | +| 206 | torch.tensordot | 是 | +| 207 | torch.trace | 否 | +| 208 | torch.tril | 是 | +| 209 | torch.tril_indices | 是 | +| 210 | torch.triu | 是 | +| 211 | torch.triu_indices | 是 | +| 212 | torch.vander | 否 | +| 213 | torch.view_as_real | 否 | +| 214 | torch.view_as_complex | 否 | +| 215 | torch.addbmm | 是 | +| 216 | torch.addmm | 是 | +| 217 | torch.addmv | 是 | +| 218 | torch.addr | 是 | +| 219 | torch.baddbmm | 是 | +| 220 | torch.bmm | 是 | +| 221 | torch.chain_matmul | 是 | +| 222 | torch.cholesky | 否 | +| 223 | torch.cholesky_inverse | 否 | +| 224 | torch.cholesky_solve | 否 | +| 225 | torch.dot | 是 | +| 226 | torch.eig | 否 | +| 227 | torch.geqrf | 否 | +| 228 | torch.ger | 是 | +| 229 | torch.inner | 否 | +| 230 | torch.inverse | 是 | +| 231 | torch.det | 否 | +| 232 | torch.logdet | 否 | +| 233 | torch.slogdet | 是 | +| 234 | torch.lstsq | 否 | +| 235 | torch.lu | 否 | +| 236 | torch.lu_solve | 否 | +| 237 | torch.lu_unpack | 否 | +| 238 | torch.matmul | 是 | +| 239 | torch.matrix_power | 是 | +| 240 | torch.matrix_rank | 是 | +| 241 | torch.matrix_exp | 否 | +| 242 | torch.mm | 是 | +| 243 | torch.mv | 是 | +| 244 | torch.orgqr | 否 | +| 245 | torch.ormqr | 否 | +| 246 | torch.outer | 否 | +| 247 | torch.pinverse | 是 | +| 248 | torch.qr | 是 | +| 249 | torch.solve | 否 | +| 250 | torch.svd | 是 | +| 251 | torch.svd_lowrank | 是 | +| 252 | torch.pca_lowrank | 是 | +| 253 | torch.symeig | 是 | +| 254 | torch.lobpcg | 否 | +| 255 | torch.trapz | 是 | +| 256 | torch.triangular_solve | 是 | +| 257 | torch.vdot | 否 | ## Utilities @@ -389,9 +383,9 @@ | 2 | torch.result_type | 是 | | 3 | torch.can_cast | 是 | | 4 | torch.promote_types | 是 | -| 6 | torch.use_deterministic_algorithms | 是 | -| 7 | torch.are_deterministic_algorithms_enabled | 是 | -| 8 | torch._assert | 是 | +| 6 | torch.use_deterministic_algorithms | 否 | +| 7 | torch.are_deterministic_algorithms_enabled | 否 | +| 8 | torch._assert | 否 | ## Other @@ -782,341 +776,342 @@ | 序号 | API名称 | 是否支持 | | ---- | -------------------------------------------------------- | ---------------------------- | | 1 | torch.nn.Parameter | 是 | -| | torch.nn.UninitializedParameter | 是 | -| 2 | torch.nn.Module | 是 | -| 3 | torch.nn.Module.add_module | 是 | -| 4 | torch.nn.Module.apply | 是 | -| 5 | torch.nn.Module.bfloat16 | 否 | -| 6 | torch.nn.Module.buffers | 是 | -| 7 | torch.nn.Module.children | 是 | -| 8 | torch.nn.Module.cpu | 是 | -| 9 | torch.nn.Module.cuda | 否 | -| 10 | torch.nn.Module.double | 否 | -| 11 | torch.nn.Module.dump_patches | 是 | -| 12 | torch.nn.Module.eval | 是 | -| 13 | torch.nn.Module.extra_repr | 是 | -| 14 | torch.nn.Module.float | 是 | -| 15 | torch.nn.Module.forward | 是 | -| 16 | torch.nn.Module.half | 是 | -| 17 | torch.nn.Module.load_state_dict | 是 | -| 18 | torch.nn.Module.modules | 是 | -| 19 | torch.nn.Module.named_buffers | 是 | -| 20 | torch.nn.Module.named_children | 是 | -| 21 | torch.nn.Module.named_modules | 是 | -| 22 | torch.nn.Module.named_parameters | 是 | -| 23 | torch.nn.Module.parameters | 是 | -| 24 | torch.nn.Module.register_backward_hook | 是 | -| 25 | torch.nn.Module.register_buffer | 是 | -| 26 | torch.nn.Module.register_forward_hook | 是 | -| 27 | torch.nn.Module.register_forward_pre_hook | 是 | -| 28 | torch.nn.Module.register_parameter | 是 | -| | torch.nn.register_module_forward_pre_hook | 是 | -| | torch.nn.register_module_forward_hook | 是 | -| | torch.nn.register_module_backward_hook | 是 | -| 29 | torch.nn.Module.requires_grad_ | 是 | -| 30 | torch.nn.Module.state_dict | 是 | -| 31 | torch.nn.Module.to | 是 | -| 32 | torch.nn.Module.train | 是 | -| 33 | torch.nn.Module.type | 是 | -| 34 | torch.nn.Module.zero_grad | 是 | -| 35 | torch.nn.Sequential | 是 | -| 36 | torch.nn.ModuleList | 是 | -| 37 | torch.nn.ModuleList.append | 是 | -| 38 | torch.nn.ModuleList.extend | 是 | -| 39 | torch.nn.ModuleList.insert | 是 | -| 40 | torch.nn.ModuleDict | 是 | -| 41 | torch.nn.ModuleDict.clear | 是 | -| 42 | torch.nn.ModuleDict.items | 是 | -| 43 | torch.nn.ModuleDict.keys | 是 | -| 44 | torch.nn.ModuleDict.pop | 是 | -| 45 | torch.nn.ModuleDict.update | 是 | -| 46 | torch.nn.ModuleDict.values | 是 | -| 47 | torch.nn.ParameterList | 是 | -| 48 | torch.nn.ParameterList.append | 是 | -| 49 | torch.nn.ParameterList.extend | 是 | -| 50 | torch.nn.ParameterDict | 是 | -| 51 | torch.nn.ParameterDict.clear | 是 | -| 52 | torch.nn.ParameterDict.items | 是 | -| 53 | torch.nn.ParameterDict.keys | 是 | -| 54 | torch.nn.ParameterDict.pop | 是 | -| 55 | torch.nn.ParameterDict.update | 是 | -| 56 | torch.nn.ParameterDict.values | 是 | -| 57 | torch.nn.Conv1d | 是 | -| 58 | torch.nn.Conv2d | 是 | -| 59 | torch.nn.Conv3d | 是 | -| 60 | torch.nn.ConvTranspose1d | 是 | -| 61 | torch.nn.ConvTranspose2d | 是 | -| 62 | torch.nn.ConvTranspose3d | 是 | -| | torch.nn.LazyConv1d | 是 | -| | torch.nn.LazyConv2d | 是 | -| | torch.nn.LazyConv3d | 是 | -| | torch.nn.LazyConvTranspose1d | 是 | -| | torch.nn.LazyConvTranspose2d | 是 | -| | torch.nn.LazyConvTranspose3d | 是 | -| 63 | torch.nn.Unfold | 是 | -| 64 | torch.nn.Fold | 是 | -| 65 | torch.nn.MaxPool1d | 是 | -| 66 | torch.nn.MaxPool2d | 是 | -| 67 | torch.nn.MaxPool3d | 是 | -| 68 | torch.nn.MaxUnpool1d | 是 | -| 69 | torch.nn.MaxUnpool2d | 是 | -| 70 | torch.nn.MaxUnpool3d | 是 | -| 71 | torch.nn.AvgPool1d | 是 | -| 72 | torch.nn.AvgPool2d | 是 | -| 73 | torch.nn.AvgPool3d | 是 | -| 74 | torch.nn.FractionalMaxPool2d | 否 | -| 75 | torch.nn.LPPool1d | 是 | -| 76 | torch.nn.LPPool2d | 是 | -| 77 | torch.nn.AdaptiveMaxPool1d | 是 | -| 78 | torch.nn.AdaptiveMaxPool2d | 是 | -| 79 | torch.nn.AdaptiveMaxPool3d | 否 | -| 80 | torch.nn.AdaptiveAvgPool1d | 是 | -| 81 | torch.nn.AdaptiveAvgPool2d | 是 | -| 82 | torch.nn.AdaptiveAvgPool3d | 是，仅支持D=1，H=1，W=1场景 | -| 83 | torch.nn.ReflectionPad1d | 否 | -| 84 | torch.nn.ReflectionPad2d | 是 | -| 85 | torch.nn.ReplicationPad1d | 否 | -| 86 | torch.nn.ReplicationPad2d | 是 | -| 87 | torch.nn.ReplicationPad3d | 否 | -| 88 | torch.nn.ZeroPad2d | 是 | -| 89 | torch.nn.ConstantPad1d | 是 | -| 90 | torch.nn.ConstantPad2d | 是 | -| 91 | torch.nn.ConstantPad3d | 是 | -| 92 | torch.nn.ELU | 是 | -| 93 | torch.nn.Hardshrink | 是 | -| | torch.nn.Hardsigmoid | 是 | -| 94 | torch.nn.Hardtanh | 是 | -| | torch.nn.Hardswish | 是 | -| 95 | torch.nn.LeakyReLU | 是 | -| 96 | torch.nn.LogSigmoid | 是 | -| 97 | torch.nn.MultiheadAttention | 是 | -| 98 | torch.nn.PReLU | 是 | -| 99 | torch.nn.ReLU | 是 | -| 100 | torch.nn.ReLU6 | 是 | -| 101 | torch.nn.RReLU | 是 | -| 102 | torch.nn.SELU | 是 | -| 103 | torch.nn.CELU | 是 | -| 104 | torch.nn.GELU | 是 | -| 105 | torch.nn.Sigmoid | 是 | -| | torch.nn.SiLU | 是 | -| 106 | torch.nn.Softplus | 是 | -| 107 | torch.nn.Softshrink | 是，SoftShrink场景暂不支持 | -| 108 | torch.nn.Softsign | 是 | -| 109 | torch.nn.Tanh | 是 | -| 110 | torch.nn.Tanhshrink | 是 | -| 111 | torch.nn.Threshold | 是 | -| 112 | torch.nn.Softmin | 是 | -| 113 | torch.nn.Softmax | 是 | -| 114 | torch.nn.Softmax2d | 是 | -| 115 | torch.nn.LogSoftmax | 是 | -| 116 | torch.nn.AdaptiveLogSoftmaxWithLoss | 否 | -| 117 | torch.nn.AdaptiveLogSoftmaxWithLoss.log_prob | 否 | -| 118 | torch.nn.AdaptiveLogSoftmaxWithLoss.predict | 否 | -| 119 | torch.nn.BatchNorm1d | 是 | -| 120 | torch.nn.BatchNorm2d | 是 | -| 121 | torch.nn.BatchNorm3d | 是 | -| 122 | torch.nn.GroupNorm | 是 | -| 123 | torch.nn.SyncBatchNorm | 是 | -| 124 | torch.nn.SyncBatchNorm.convert_sync_batchnorm | 是 | -| 125 | torch.nn.InstanceNorm1d | 是 | -| 126 | torch.nn.InstanceNorm2d | 是 | -| 127 | torch.nn.InstanceNorm3d | 是 | -| 128 | torch.nn.LayerNorm | 是 | -| 129 | torch.nn.LocalResponseNorm | 是 | -| 130 | torch.nn.RNNBase | 是 | -| 131 | torch.nn.RNNBase.flatten_parameters | 是 | -| 132 | torch.nn.RNN | 是 | -| 133 | torch.nn.LSTM | 是 | -| 134 | torch.nn.GRU | 是，DynamicGRUV2场景暂不支持 | -| 135 | torch.nn.RNNCell | 是 | -| 136 | torch.nn.LSTMCell | 是 | -| 137 | torch.nn.GRUCell | 是 | -| 138 | torch.nn.Transformer | 是 | -| 139 | torch.nn.Transformer.forward | 是 | -| 140 | torch.nn.Transformer.generate_square_subsequent_mask | 是 | -| 141 | torch.nn.TransformerEncoder | 是 | -| 142 | torch.nn.TransformerEncoder.forward | 是 | -| 143 | torch.nn.TransformerDecoder | 是 | -| 144 | torch.nn.TransformerDecoder.forward | 是 | -| 145 | torch.nn.TransformerEncoderLayer | 是 | -| 146 | torch.nn.TransformerEncoderLayer.forward | 是 | -| 147 | torch.nn.TransformerDecoderLayer | 是 | -| 148 | torch.nn.TransformerDecoderLayer.forward | 是 | -| 149 | torch.nn.Identity | 是 | -| 150 | torch.nn.Linear | 是 | -| 151 | torch.nn.Bilinear | 是 | -| 152 | torch.nn.Dropout | 是 | -| 153 | torch.nn.Dropout2d | 是 | -| 154 | torch.nn.Dropout3d | 是 | -| 155 | torch.nn.AlphaDropout | 是 | -| 156 | torch.nn.Embedding | 是 | -| 157 | torch.nn.Embedding.from_pretrained | 是 | -| 158 | torch.nn.EmbeddingBag | 是 | -| 159 | torch.nn.EmbeddingBag.from_pretrained | 是 | -| 160 | torch.nn.CosineSimilarity | 是 | -| 161 | torch.nn.PairwiseDistance | 是 | -| 162 | torch.nn.L1Loss | 是 | -| 163 | torch.nn.MSELoss | 是 | -| 164 | torch.nn.CrossEntropyLoss | 是 | -| 165 | torch.nn.CTCLoss | 是 | -| 166 | torch.nn.NLLLoss | 是 | -| 167 | torch.nn.PoissonNLLLoss | 是 | -| | torch.nn.GaussianNLLLoss | 是 | -| 168 | torch.nn.KLDivLoss | 是 | -| 169 | torch.nn.BCELoss | 是 | -| 170 | torch.nn.BCEWithLogitsLoss | 是 | -| 171 | torch.nn.MarginRankingLoss | 是 | -| 172 | torch.nn.HingeEmbeddingLoss | 是 | -| 173 | torch.nn.MultiLabelMarginLoss | 是 | -| 174 | torch.nn.SmoothL1Loss | 是 | -| 175 | torch.nn.SoftMarginLoss | 是 | -| 176 | torch.nn.MultiLabelSoftMarginLoss | 是 | -| 177 | torch.nn.CosineEmbeddingLoss | 是 | -| 178 | torch.nn.MultiMarginLoss | 否 | -| 179 | torch.nn.TripletMarginLoss | 是 | -| | torch.nn.TripletMarginLossWithDistanceLoss | 是 | -| 180 | torch.nn.PixelShuffle | 是 | -| | torch.nn.PixelUnshuffle | 是 | -| 181 | torch.nn.Upsample | 是 | -| 182 | torch.nn.UpsamplingNearest2d | 是 | -| 183 | torch.nn.UpsamplingBilinear2d | 是 | -| | torch.nn.ChannelShuffle | 是 | -| 184 | torch.nn.DataParallel | 否 | -| 185 | torch.nn.parallel.DistributedDataParallel | 是 | -| 186 | torch.nn.parallel.DistributedDataParallel.no_sync | 是 | -| 187 | torch.nn.utils.clip_grad_norm_ | 是 | -| 188 | torch.nn.utils.clip_grad_value_ | 是 | -| 189 | torch.nn.utils.parameters_to_vector | 是 | -| 190 | torch.nn.utils.vector_to_parameters | 是 | -| | torch.nn.utils.Prune.BasePruningMethod | 是 | -| 197 | torch.nn.utils.prune.PruningContainer | 是 | -| 198 | torch.nn.utils.prune.PruningContainer.add_pruning_method | 是 | -| 199 | torch.nn.utils.prune.PruningContainer.apply | 是 | -| 200 | torch.nn.utils.prune.PruningContainer.apply_mask | 是 | -| 201 | torch.nn.utils.prune.PruningContainer.compute_mask | 是 | -| 202 | torch.nn.utils.prune.PruningContainer.prune | 是 | -| 203 | torch.nn.utils.prune.PruningContainer.remove | 是 | -| 204 | torch.nn.utils.prune.Identity | 是 | -| 205 | torch.nn.utils.prune.Identity.apply | 是 | -| 206 | torch.nn.utils.prune.Identity.apply_mask | 是 | -| 207 | torch.nn.utils.prune.Identity.prune | 是 | -| 208 | torch.nn.utils.prune.Identity.remove | 是 | -| 209 | torch.nn.utils.prune.RandomUnstructured | 是 | -| 210 | torch.nn.utils.prune.RandomUnstructured.apply | 是 | -| 211 | torch.nn.utils.prune.RandomUnstructured.apply_mask | 是 | -| 212 | torch.nn.utils.prune.RandomUnstructured.prune | 是 | -| 213 | torch.nn.utils.prune.RandomUnstructured.remove | 是 | -| 214 | torch.nn.utils.prune.L1Unstructured | 是 | -| 215 | torch.nn.utils.prune.L1Unstructured.apply | 是 | -| 216 | torch.nn.utils.prune.L1Unstructured.apply_mask | 是 | -| 217 | torch.nn.utils.prune.L1Unstructured.prune | 是 | -| 218 | torch.nn.utils.prune.L1Unstructured.remove | 是 | -| 219 | torch.nn.utils.prune.RandomStructured | 是 | -| 220 | torch.nn.utils.prune.RandomStructured.apply | 是 | -| 221 | torch.nn.utils.prune.RandomStructured.apply_mask | 是 | -| 222 | torch.nn.utils.prune.RandomStructured.compute_mask | 是 | -| 223 | torch.nn.utils.prune.RandomStructured.prune | 是 | -| 224 | torch.nn.utils.prune.RandomStructured.remove | 是 | -| 225 | torch.nn.utils.prune.LnStructured | 是 | -| 226 | torch.nn.utils.prune.LnStructured.apply | 是 | -| 227 | torch.nn.utils.prune.LnStructured.apply_mask | 是 | -| 228 | torch.nn.utils.prune.LnStructured.compute_mask | 是 | -| 229 | torch.nn.utils.prune.LnStructured.prune | 是 | -| 230 | torch.nn.utils.prune.LnStructured.remove | 是 | -| 231 | torch.nn.utils.prune.CustomFromMask | 是 | -| 232 | torch.nn.utils.prune.CustomFromMask.apply | 是 | -| 233 | torch.nn.utils.prune.CustomFromMask.apply_mask | 是 | -| 234 | torch.nn.utils.prune.CustomFromMask.prune | 是 | -| 235 | torch.nn.utils.prune.CustomFromMask.remove | 是 | -| 236 | torch.nn.utils.prune.identity | 是 | -| 237 | torch.nn.utils.prune.random_unstructured | 是 | -| 238 | torch.nn.utils.prune.l1_unstructured | 是 | -| 239 | torch.nn.utils.prune.random_structured | 是 | -| 240 | torch.nn.utils.prune.ln_structured | 是 | -| 241 | torch.nn.utils.prune.global_unstructured | 是 | -| 242 | torch.nn.utils.prune.custom_from_mask | 是 | -| 243 | torch.nn.utils.prune.remove | 是 | -| 244 | torch.nn.utils.prune.is_pruned | 是 | -| 245 | torch.nn.utils.weight_norm | 是 | -| 246 | torch.nn.utils.remove_weight_norm | 是 | -| 247 | torch.nn.utils.spectral_norm | 是 | -| 248 | torch.nn.utils.remove_spectral_norm | 是 | -| 249 | torch.nn.utils.rnn.PackedSequence | 是 | -| 250 | torch.nn.utils.rnn.pack_padded_sequence | 是 | -| 251 | torch.nn.utils.rnn.pad_packed_sequence | 否 | -| 252 | torch.nn.utils.rnn.pad_sequence | 是 | -| 253 | torch.nn.utils.rnn.pack_sequence | 否 | -| 254 | torch.nn.Flatten | 是 | -| | torch.nn.Unflatten | 是 | -| | torch.nn.modules.lazy.LazyModuleMixin | 是 | -| 255 | torch.quantization.quantize | 否 | -| 256 | torch.quantization.quantize_dynamic | 否 | -| 257 | torch.quantization.quantize_qat | 否 | -| 258 | torch.quantization.prepare | 是 | -| 259 | torch.quantization.prepare_qat | 否 | -| 260 | torch.quantization.convert | 否 | -| 261 | torch.quantization.QConfig | 是 | -| 262 | torch.quantization.QConfigDynamic | 是 | -| 263 | torch.quantization.fuse_modules | 是 | -| 264 | torch.quantization.QuantStub | 是 | -| 265 | torch.quantization.DeQuantStub | 是 | -| 266 | torch.quantization.QuantWrapper | 是 | -| 267 | torch.quantization.add_quant_dequant | 是 | -| 268 | torch.quantization.add_observer_ | 是 | -| 269 | torch.quantization.swap_module | 是 | -| 270 | torch.quantization.propagate_qconfig_ | 是 | -| 271 | torch.quantization.default_eval_fn | 是 | -| 272 | torch.quantization.MinMaxObserver | 是 | -| 273 | torch.quantization.MovingAverageMinMaxObserver | 是 | -| 274 | torch.quantization.PerChannelMinMaxObserver | 是 | -| 275 | torch.quantization.MovingAveragePerChannelMinMaxObserver | 是 | -| 276 | torch.quantization.HistogramObserver | 否 | -| 277 | torch.quantization.FakeQuantize | 否 | -| 278 | torch.quantization.NoopObserver | 是 | -| 279 | torch.quantization.get_observer_dict | 是 | -| 280 | torch.quantization.RecordingObserver | 是 | -| 281 | torch.nn.intrinsic.ConvBn2d | 是 | -| 282 | torch.nn.intrinsic.ConvBnReLU2d | 是 | -| 283 | torch.nn.intrinsic.ConvReLU2d | 是 | -| 284 | torch.nn.intrinsic.ConvReLU3d | 是 | -| 285 | torch.nn.intrinsic.LinearReLU | 是 | -| 286 | torch.nn.intrinsic.qat.ConvBn2d | 否 | -| 287 | torch.nn.intrinsic.qat.ConvBnReLU2d | 否 | -| 288 | torch.nn.intrinsic.qat.ConvReLU2d | 否 | -| 289 | torch.nn.intrinsic.qat.LinearReLU | 否 | -| 290 | torch.nn.intrinsic.quantized.ConvReLU2d | 否 | -| 291 | torch.nn.intrinsic.quantized.ConvReLU3d | 否 | -| 292 | torch.nn.intrinsic.quantized.LinearReLU | 否 | -| 293 | torch.nn.qat.Conv2d | 否 | -| 294 | torch.nn.qat.Conv2d.from_float | 否 | -| 295 | torch.nn.qat.Linear | 否 | -| 296 | torch.nn.qat.Linear.from_float | 否 | -| 297 | torch.nn.quantized.functional.relu | 否 | -| 298 | torch.nn.quantized.functional.linear | 否 | -| 299 | torch.nn.quantized.functional.conv2d | 否 | -| 300 | torch.nn.quantized.functional.conv3d | 否 | -| 301 | torch.nn.quantized.functional.max_pool2d | 否 | -| 302 | torch.nn.quantized.functional.adaptive_avg_pool2d | 否 | -| 303 | torch.nn.quantized.functional.avg_pool2d | 否 | -| 304 | torch.nn.quantized.functional.interpolate | 否 | -| 305 | torch.nn.quantized.functional.upsample | 否 | -| 306 | torch.nn.quantized.functional.upsample_bilinear | 否 | -| 307 | torch.nn.quantized.functional.upsample_nearest | 否 | -| 308 | torch.nn.quantized.ReLU | 否 | -| 309 | torch.nn.quantized.ReLU6 | 否 | -| 310 | torch.nn.quantized.Conv2d | 否 | -| 311 | torch.nn.quantized.Conv2d.from_float | 否 | -| 312 | torch.nn.quantized.Conv3d | 否 | -| 313 | torch.nn.quantized.Conv3d.from_float | 否 | -| 314 | torch.nn.quantized.FloatFunctional | 是 | -| 315 | torch.nn.quantized.QFunctional | 否 | -| 316 | torch.nn.quantized.Quantize | 是 | -| 317 | torch.nn.quantized.DeQuantize | 否 | -| 318 | torch.nn.quantized.Linear | 否 | -| 319 | torch.nn.quantized.Linear.from_float | 否 | -| 320 | torch.nn.quantized.dynamic.Linear | 否 | -| 321 | torch.nn.quantized.dynamic.Linear.from_float | 否 | -| 322 | torch.nn.quantized.dynamic.LSTM | 否 | +| 2 | torch.nn.UninitializedParameter | 否 | +| 3 | torch.nn.Module | 是 | +| 4 | torch.nn.Module.add_module | 是 | +| 5 | torch.nn.Module.apply | 是 | +| 6 | torch.nn.Module.bfloat16 | 否 | +| 7 | torch.nn.Module.buffers | 是 | +| 8 | torch.nn.Module.children | 是 | +| 9 | torch.nn.Module.cpu | 是 | +| 10 | torch.nn.Module.cuda | 否 | +| 11 | torch.nn.Module.double | 否 | +| 12 | torch.nn.Module.dump_patches | 是 | +| 13 | torch.nn.Module.eval | 是 | +| 14 | torch.nn.Module.extra_repr | 是 | +| 15 | torch.nn.Module.float | 是 | +| 16 | torch.nn.Module.forward | 是 | +| 17 | torch.nn.Module.half | 是 | +| 18 | torch.nn.Module.load_state_dict | 是 | +| 19 | torch.nn.Module.modules | 是 | +| 20 | torch.nn.Module.named_buffers | 是 | +| 21 | torch.nn.Module.named_children | 是 | +| 22 | torch.nn.Module.named_modules | 是 | +| 23 | torch.nn.Module.named_parameters | 是 | +| 24 | torch.nn.Module.parameters | 是 | +| 25 | torch.nn.Module.register_backward_hook | 是 | +| 26 | torch.nn.Module.register_buffer | 是 | +| 27 | torch.nn.Module.register_forward_hook | 是 | +| 28 | torch.nn.Module.register_forward_pre_hook | 是 | +| 29 | torch.nn.Module.register_parameter | 是 | +| 30 | torch.nn.register_module_forward_pre_hook | 否 | +| 31 | torch.nn.register_module_forward_hook | 否 | +| 32 | torch.nn.register_module_backward_hook | 否 | +| 33 | torch.nn.Module.requires_grad_ | 是 | +| 34 | torch.nn.Module.state_dict | 是 | +| 35 | torch.nn.Module.to | 是 | +| 36 | torch.nn.Module.train | 是 | +| 37 | torch.nn.Module.type | 是 | +| 38 | torch.nn.Module.zero_grad | 是 | +| 39 | torch.nn.Sequential | 是 | +| 40 | torch.nn.ModuleList | 是 | +| 41 | torch.nn.ModuleList.append | 是 | +| 42 | torch.nn.ModuleList.extend | 是 | +| 43 | torch.nn.ModuleList.insert | 是 | +| 44 | torch.nn.ModuleDict | 是 | +| 45 | torch.nn.ModuleDict.clear | 是 | +| 46 | torch.nn.ModuleDict.items | 是 | +| 47 | torch.nn.ModuleDict.keys | 是 | +| 48 | torch.nn.ModuleDict.pop | 是 | +| 49 | torch.nn.ModuleDict.update | 是 | +| 50 | torch.nn.ModuleDict.values | 是 | +| 51 | torch.nn.ParameterList | 是 | +| 52 | torch.nn.ParameterList.append | 是 | +| 53 | torch.nn.ParameterList.extend | 是 | +| 54 | torch.nn.ParameterDict | 是 | +| 55 | torch.nn.ParameterDict.clear | 是 | +| 56 | torch.nn.ParameterDict.items | 是 | +| 57 | torch.nn.ParameterDict.keys | 是 | +| 58 | torch.nn.ParameterDict.pop | 是 | +| 59 | torch.nn.ParameterDict.update | 是 | +| 60 | torch.nn.ParameterDict.values | 是 | +| 61 | torch.nn.Conv1d | 是 | +| 62 | torch.nn.Conv2d | 是 | +| 63 | torch.nn.Conv3d | 是 | +| 64 | torch.nn.ConvTranspose1d | 是 | +| 65 | torch.nn.ConvTranspose2d | 是 | +| 66 | torch.nn.ConvTranspose3d | 是 | +| 67 | torch.nn.LazyConv1d | 否 | +| 68 | torch.nn.LazyConv2d | 否 | +| 69 | torch.nn.LazyConv3d | 否 | +| 70 | torch.nn.LazyConvTranspose1d | 否 | +| 71 | torch.nn.LazyConvTranspose2d | 否 | +| 72 | torch.nn.LazyConvTranspose3d | 否 | +| 73 | torch.nn.Unfold | 是 | +| 74 | torch.nn.Fold | 是 | +| 75 | torch.nn.MaxPool1d | 是 | +| 76 | torch.nn.MaxPool2d | 是 | +| 77 | torch.nn.MaxPool3d | 是 | +| 78 | torch.nn.MaxUnpool1d | 是 | +| 79 | torch.nn.MaxUnpool2d | 是 | +| 80 | torch.nn.MaxUnpool3d | 是 | +| 81 | torch.nn.AvgPool1d | 是 | +| 82 | torch.nn.AvgPool2d | 是 | +| 83 | torch.nn.AvgPool3d | 是 | +| 84 | torch.nn.FractionalMaxPool2d | 否 | +| 85 | torch.nn.LPPool1d | 是 | +| 86 | torch.nn.LPPool2d | 是 | +| 87 | torch.nn.AdaptiveMaxPool1d | 是 | +| 88 | torch.nn.AdaptiveMaxPool2d | 是 | +| 89 | torch.nn.AdaptiveMaxPool3d | 否 | +| 90 | torch.nn.AdaptiveAvgPool1d | 是 | +| 91 | torch.nn.AdaptiveAvgPool2d | 是 | +| 92 | torch.nn.AdaptiveAvgPool3d | 是，仅支持D=1，H=1，W=1场景 | +| 93 | torch.nn.ReflectionPad1d | 否 | +| 94 | torch.nn.ReflectionPad2d | 是 | +| 95 | torch.nn.ReplicationPad1d | 否 | +| 96 | torch.nn.ReplicationPad2d | 是 | +| 97 | torch.nn.ReplicationPad3d | 否 | +| 98 | torch.nn.ZeroPad2d | 是 | +| 99 | torch.nn.ConstantPad1d | 是 | +| 100 | torch.nn.ConstantPad2d | 是 | +| 101 | torch.nn.ConstantPad3d | 是 | +| 102 | torch.nn.ELU | 是 | +| 103 | torch.nn.Hardshrink | 是 | +| 104 | torch.nn.Hardsigmoid | 否 | +| 105 | torch.nn.Hardtanh | 是 | +| 106 | torch.nn.Hardswish | 否 | +| 107 | torch.nn.LeakyReLU | 是 | +| 108 | torch.nn.LogSigmoid | 是 | +| 109 | torch.nn.MultiheadAttention | 是 | +| 110 | torch.nn.PReLU | 是 | +| 111 | torch.nn.ReLU | 是 | +| 112 | torch.nn.ReLU6 | 是 | +| 113 | torch.nn.RReLU | 是 | +| 114 | torch.nn.SELU | 是 | +| 115 | torch.nn.CELU | 是 | +| 116 | torch.nn.GELU | 是 | +| 117 | torch.nn.Sigmoid | 是 | +| 118 | torch.nn.SiLU | 否 | +| 119 | torch.nn.Softplus | 是 | +| 120 | torch.nn.Softshrink | 是，SoftShrink场景暂不支持 | +| 121 | torch.nn.Softsign | 是 | +| 122 | torch.nn.Tanh | 是 | +| 123 | torch.nn.Tanhshrink | 是 | +| 124 | torch.nn.Threshold | 是 | +| 125 | torch.nn.Softmin | 是 | +| 126 | torch.nn.Softmax | 是 | +| 127 | torch.nn.Softmax2d | 是 | +| 128 | torch.nn.LogSoftmax | 是 | +| 129 | torch.nn.AdaptiveLogSoftmaxWithLoss | 否 | +| 130 | torch.nn.AdaptiveLogSoftmaxWithLoss.log_prob | 否 | +| 131 | torch.nn.AdaptiveLogSoftmaxWithLoss.predict | 否 | +| 132 | torch.nn.BatchNorm1d | 是 | +| 133 | torch.nn.BatchNorm2d | 是 | +| 134 | torch.nn.BatchNorm3d | 是 | +| 135 | torch.nn.GroupNorm | 是 | +| 136 | torch.nn.SyncBatchNorm | 是 | +| 137 | torch.nn.SyncBatchNorm.convert_sync_batchnorm | 是 | +| 138 | torch.nn.InstanceNorm1d | 是 | +| 139 | torch.nn.InstanceNorm2d | 是 | +| 140 | torch.nn.InstanceNorm3d | 是 | +| 141 | torch.nn.LayerNorm | 是 | +| 142 | torch.nn.LocalResponseNorm | 是 | +| 143 | torch.nn.RNNBase | 是 | +| 144 | torch.nn.RNNBase.flatten_parameters | 是 | +| 145 | torch.nn.RNN | 是 | +| 146 | torch.nn.LSTM | 是 | +| 147 | torch.nn.GRU | 是，DynamicGRUV2场景暂不支持 | +| 148 | torch.nn.RNNCell | 是 | +| 149 | torch.nn.LSTMCell | 是 | +| 150 | torch.nn.GRUCell | 是 | +| 151 | torch.nn.Transformer | 是 | +| 152 | torch.nn.Transformer.forward | 是 | +| 153 | torch.nn.Transformer.generate_square_subsequent_mask | 是 | +| 154 | torch.nn.TransformerEncoder | 是 | +| 155 | torch.nn.TransformerEncoder.forward | 是 | +| 156 | torch.nn.TransformerDecoder | 是 | +| 157 | torch.nn.TransformerDecoder.forward | 是 | +| 158 | torch.nn.TransformerEncoderLayer | 是 | +| 159 | torch.nn.TransformerEncoderLayer.forward | 是 | +| 160 | torch.nn.TransformerDecoderLayer | 是 | +| 161 | torch.nn.TransformerDecoderLayer.forward | 是 | +| 162 | torch.nn.Identity | 是 | +| 163 | torch.nn.Linear | 是 | +| 164 | torch.nn.Bilinear | 是 | +| 165 | torch.nn.LazyLinear | 否 | +| 166 | torch.nn.Dropout | 是 | +| 167 | torch.nn.Dropout2d | 是 | +| 168 | torch.nn.Dropout3d | 是 | +| 169 | torch.nn.AlphaDropout | 是 | +| 170 | torch.nn.Embedding | 是 | +| 171 | torch.nn.Embedding.from_pretrained | 是 | +| 172 | torch.nn.EmbeddingBag | 是 | +| 173 | torch.nn.EmbeddingBag.from_pretrained | 是 | +| 174 | torch.nn.CosineSimilarity | 是 | +| 175 | torch.nn.PairwiseDistance | 是 | +| 176 | torch.nn.L1Loss | 是 | +| 177 | torch.nn.MSELoss | 是 | +| 178 | torch.nn.CrossEntropyLoss | 是 | +| 179 | torch.nn.CTCLoss | 是 | +| 180 | torch.nn.NLLLoss | 是 | +| 181 | torch.nn.PoissonNLLLoss | 是 | +| 182 | torch.nn.GaussianNLLLoss | 否 | +| 183 | torch.nn.KLDivLoss | 是 | +| 184 | torch.nn.BCELoss | 是 | +| 185 | torch.nn.BCEWithLogitsLoss | 是 | +| 186 | torch.nn.MarginRankingLoss | 是 | +| 187 | torch.nn.HingeEmbeddingLoss | 是 | +| 188 | torch.nn.MultiLabelMarginLoss | 是 | +| 189 | torch.nn.SmoothL1Loss | 是 | +| 190 | torch.nn.SoftMarginLoss | 是 | +| 191 | torch.nn.MultiLabelSoftMarginLoss | 是 | +| 192 | torch.nn.CosineEmbeddingLoss | 是 | +| 193 | torch.nn.MultiMarginLoss | 否 | +| 194 | torch.nn.TripletMarginLoss | 是 | +| 195 | torch.nn.TripletMarginLossWithDistanceLoss | 否 | +| 196 | torch.nn.PixelShuffle | 是 | +| 197 | torch.nn.PixelUnshuffle | 否 | +| 198 | torch.nn.Upsample | 是 | +| 199 | torch.nn.UpsamplingNearest2d | 是 | +| 200 | torch.nn.UpsamplingBilinear2d | 是 | +| 201 | torch.nn.ChannelShuffle | 否 | +| 202 | torch.nn.DataParallel | 否 | +| 203 | torch.nn.parallel.DistributedDataParallel | 是 | +| 204 | torch.nn.parallel.DistributedDataParallel.no_sync | 是 | +| 205 | torch.nn.utils.clip_grad_norm_ | 是 | +| 206 | torch.nn.utils.clip_grad_value_ | 是 | +| 207 | torch.nn.utils.parameters_to_vector | 是 | +| 208 | torch.nn.utils.vector_to_parameters | 是 | +| 209 | torch.nn.utils.Prune.BasePruningMethod | 否 | +| 210 | torch.nn.utils.prune.PruningContainer | 是 | +| 211 | torch.nn.utils.prune.PruningContainer.add_pruning_method | 是 | +| 212 | torch.nn.utils.prune.PruningContainer.apply | 是 | +| 213 | torch.nn.utils.prune.PruningContainer.apply_mask | 是 | +| 214 | torch.nn.utils.prune.PruningContainer.compute_mask | 是 | +| 215 | torch.nn.utils.prune.PruningContainer.prune | 是 | +| 216 | torch.nn.utils.prune.PruningContainer.remove | 是 | +| 217 | torch.nn.utils.prune.Identity | 是 | +| 218 | torch.nn.utils.prune.Identity.apply | 是 | +| 219 | torch.nn.utils.prune.Identity.apply_mask | 是 | +| 220 | torch.nn.utils.prune.Identity.prune | 是 | +| 221 | torch.nn.utils.prune.Identity.remove | 是 | +| 222 | torch.nn.utils.prune.RandomUnstructured | 是 | +| 223 | torch.nn.utils.prune.RandomUnstructured.apply | 是 | +| 224 | torch.nn.utils.prune.RandomUnstructured.apply_mask | 是 | +| 225 | torch.nn.utils.prune.RandomUnstructured.prune | 是 | +| 226 | torch.nn.utils.prune.RandomUnstructured.remove | 是 | +| 227 | torch.nn.utils.prune.L1Unstructured | 是 | +| 228 | torch.nn.utils.prune.L1Unstructured.apply | 是 | +| 229 | torch.nn.utils.prune.L1Unstructured.apply_mask | 是 | +| 230 | torch.nn.utils.prune.L1Unstructured.prune | 是 | +| 231 | torch.nn.utils.prune.L1Unstructured.remove | 是 | +| 232 | torch.nn.utils.prune.RandomStructured | 是 | +| 233 | torch.nn.utils.prune.RandomStructured.apply | 是 | +| 234 | torch.nn.utils.prune.RandomStructured.apply_mask | 是 | +| 235 | torch.nn.utils.prune.RandomStructured.compute_mask | 是 | +| 236 | torch.nn.utils.prune.RandomStructured.prune | 是 | +| 237 | torch.nn.utils.prune.RandomStructured.remove | 是 | +| 238 | torch.nn.utils.prune.LnStructured | 是 | +| 239 | torch.nn.utils.prune.LnStructured.apply | 是 | +| 240 | torch.nn.utils.prune.LnStructured.apply_mask | 是 | +| 241 | torch.nn.utils.prune.LnStructured.compute_mask | 是 | +| 242 | torch.nn.utils.prune.LnStructured.prune | 是 | +| 243 | torch.nn.utils.prune.LnStructured.remove | 是 | +| 244 | torch.nn.utils.prune.CustomFromMask | 是 | +| 245 | torch.nn.utils.prune.CustomFromMask.apply | 是 | +| 246 | torch.nn.utils.prune.CustomFromMask.apply_mask | 是 | +| 247 | torch.nn.utils.prune.CustomFromMask.prune | 是 | +| 248 | torch.nn.utils.prune.CustomFromMask.remove | 是 | +| 249 | torch.nn.utils.prune.identity | 是 | +| 250 | torch.nn.utils.prune.random_unstructured | 是 | +| 251 | torch.nn.utils.prune.l1_unstructured | 是 | +| 252 | torch.nn.utils.prune.random_structured | 是 | +| 253 | torch.nn.utils.prune.ln_structured | 是 | +| 254 | torch.nn.utils.prune.global_unstructured | 是 | +| 255 | torch.nn.utils.prune.custom_from_mask | 是 | +| 256 | torch.nn.utils.prune.remove | 是 | +| 257 | torch.nn.utils.prune.is_pruned | 是 | +| 258 | torch.nn.utils.weight_norm | 是 | +| 259 | torch.nn.utils.remove_weight_norm | 是 | +| 260 | torch.nn.utils.spectral_norm | 是 | +| 261 | torch.nn.utils.remove_spectral_norm | 是 | +| 262 | torch.nn.utils.rnn.PackedSequence | 是 | +| 263 | torch.nn.utils.rnn.pack_padded_sequence | 是 | +| 264 | torch.nn.utils.rnn.pad_packed_sequence | 否 | +| 265 | torch.nn.utils.rnn.pad_sequence | 是 | +| 266 | torch.nn.utils.rnn.pack_sequence | 否 | +| 267 | torch.nn.Flatten | 是 | +| 268 | torch.nn.Unflatten | 否 | +| 269 | torch.nn.modules.lazy.LazyModuleMixin | 否 | +| 270 | torch.quantization.quantize | 否 | +| 271 | torch.quantization.quantize_dynamic | 否 | +| 272 | torch.quantization.quantize_qat | 否 | +| 273 | torch.quantization.prepare | 是 | +| 274 | torch.quantization.prepare_qat | 否 | +| 275 | torch.quantization.convert | 否 | +| 276 | torch.quantization.QConfig | 是 | +| 277 | torch.quantization.QConfigDynamic | 是 | +| 278 | torch.quantization.fuse_modules | 是 | +| 279 | torch.quantization.QuantStub | 是 | +| 280 | torch.quantization.DeQuantStub | 是 | +| 281 | torch.quantization.QuantWrapper | 是 | +| 282 | torch.quantization.add_quant_dequant | 是 | +| 283 | torch.quantization.add_observer_ | 是 | +| 284 | torch.quantization.swap_module | 是 | +| 285 | torch.quantization.propagate_qconfig_ | 是 | +| 286 | torch.quantization.default_eval_fn | 是 | +| 287 | torch.quantization.MinMaxObserver | 是 | +| 288 | torch.quantization.MovingAverageMinMaxObserver | 是 | +| 289 | torch.quantization.PerChannelMinMaxObserver | 是 | +| 290 | torch.quantization.MovingAveragePerChannelMinMaxObserver | 是 | +| 291 | torch.quantization.HistogramObserver | 否 | +| 292 | torch.quantization.FakeQuantize | 否 | +| 293 | torch.quantization.NoopObserver | 是 | +| 294 | torch.quantization.get_observer_dict | 是 | +| 295 | torch.quantization.RecordingObserver | 是 | +| 296 | torch.nn.intrinsic.ConvBn2d | 是 | +| 297 | torch.nn.intrinsic.ConvBnReLU2d | 是 | +| 298 | torch.nn.intrinsic.ConvReLU2d | 是 | +| 299 | torch.nn.intrinsic.ConvReLU3d | 是 | +| 300 | torch.nn.intrinsic.LinearReLU | 是 | +| 301 | torch.nn.intrinsic.qat.ConvBn2d | 否 | +| 302 | torch.nn.intrinsic.qat.ConvBnReLU2d | 否 | +| 303 | torch.nn.intrinsic.qat.ConvReLU2d | 否 | +| 304 | torch.nn.intrinsic.qat.LinearReLU | 否 | +| 305 | torch.nn.intrinsic.quantized.ConvReLU2d | 否 | +| 306 | torch.nn.intrinsic.quantized.ConvReLU3d | 否 | +| 307 | torch.nn.intrinsic.quantized.LinearReLU | 否 | +| 308 | torch.nn.qat.Conv2d | 否 | +| 309 | torch.nn.qat.Conv2d.from_float | 否 | +| 310 | torch.nn.qat.Linear | 否 | +| 311 | torch.nn.qat.Linear.from_float | 否 | +| 312 | torch.nn.quantized.functional.relu | 否 | +| 313 | torch.nn.quantized.functional.linear | 否 | +| 314 | torch.nn.quantized.functional.conv2d | 否 | +| 315 | torch.nn.quantized.functional.conv3d | 否 | +| 316 | torch.nn.quantized.functional.max_pool2d | 否 | +| 317 | torch.nn.quantized.functional.adaptive_avg_pool2d | 否 | +| 318 | torch.nn.quantized.functional.avg_pool2d | 否 | +| 319 | torch.nn.quantized.functional.interpolate | 否 | +| 320 | torch.nn.quantized.functional.upsample | 否 | +| 321 | torch.nn.quantized.functional.upsample_bilinear | 否 | +| 322 | torch.nn.quantized.functional.upsample_nearest | 否 | +| 323 | torch.nn.quantized.ReLU | 否 | +| 324 | torch.nn.quantized.ReLU6 | 否 | +| 325 | torch.nn.quantized.Conv2d | 否 | +| 326 | torch.nn.quantized.Conv2d.from_float | 否 | +| 327 | torch.nn.quantized.Conv3d | 否 | +| 328 | torch.nn.quantized.Conv3d.from_float | 否 | +| 329 | torch.nn.quantized.FloatFunctional | 是 | +| 330 | torch.nn.quantized.QFunctional | 否 | +| 331 | torch.nn.quantized.Quantize | 是 | +| 332 | torch.nn.quantized.DeQuantize | 否 | +| 333 | torch.nn.quantized.Linear | 否 | +| 334 | torch.nn.quantized.Linear.from_float | 否 | +| 335 | torch.nn.quantized.dynamic.Linear | 否 | +| 336 | torch.nn.quantized.dynamic.Linear.from_float | 否 | +| 337 | torch.nn.quantized.dynamic.LSTM | 否 | ## Functions(torch.nn.functional) @@ -1178,111 +1173,116 @@ | 54 | torch.nn.functional.log_softmax | 是 | | 55 | torch.nn.functional.tanh | 是 | | 56 | torch.nn.functional.sigmoid | 是 | -| 57 | torch.nn.functional.hardsigmoid | 是 | -| 58 | torch.nn.functional.silu | 是 | -| 59 | torch.nn.functional.batch_norm | 是 | -| 60 | torch.nn.functional.instance_norm | 是 | -| 61 | torch.nn.functional.layer_norm | 是 | -| 62 | torch.nn.functional.local_response_norm | 是 | -| 63 | torch.nn.functional.normalize | 是 | -| 64 | torch.nn.functional.linear | 是 | -| 65 | torch.nn.functional.bilinear | 是 | -| 66 | torch.nn.functional.dropout | 是 | -| 67 | torch.nn.functional.alpha_dropout | 是 | -| 68 | torch.nn.functional.dropout2d | 是 | -| 69 | torch.nn.functional.dropout3d | 是 | -| 70 | torch.nn.functional.embedding | 是 | -| 71 | torch.nn.functional.embedding_bag | 是 | -| 72 | torch.nn.functional.one_hot | 是 | -| 73 | torch.nn.functional.pairwise_distance | 是 | -| 74 | torch.nn.functional.cosine_similarity | 是 | -| 75 | torch.nn.functional.pdist | 是 | -| 76 | torch.nn.functional.binary_cross_entropy | 是 | -| 77 | torch.nn.functional.binary_cross_entropy_with_logits | 是 | -| 78 | torch.nn.functional.poisson_nll_loss | 是 | -| 79 | torch.nn.functional.cosine_embedding_loss | 是 | -| 80 | torch.nn.functional.cross_entropy | 是 | -| 81 | torch.nn.functional.ctc_loss | 是（仅支持2维输入） | -| 82 | torch.nn.functional.hinge_embedding_loss | 是 | -| 83 | torch.nn.functional.kl_div | 是 | -| 84 | torch.nn.functional.l1_loss | 是 | -| 85 | torch.nn.functional.mse_loss | 是 | -| 86 | torch.nn.functional.margin_ranking_loss | 是 | -| 87 | torch.nn.functional.multilabel_margin_loss | 是 | -| 88 | torch.nn.functional.multilabel_soft_margin_loss | 是 | -| 89 | torch.nn.functional.multi_margin_loss | 否 | -| 90 | torch.nn.functional.nll_loss | 是 | -| 91 | torch.nn.functional.smooth_l1_loss | 是 | -| 92 | torch.nn.functional.soft_margin_loss | 是 | -| 93 | torch.nn.functional.triplet_margin_loss | 是 | -| 94 | torch.nn.functional.triplet_margin_with_distance_loss | 是 | -| 95 | torch.nn.functional.pixel_shuffle | 是 | -| 96 | torch.nn.functional.pad | 是 | -| 97 | torch.nn.functional.interpolate | 是 | -| 98 | torch.nn.functional.upsample | 是 | -| 99 | torch.nn.functional.upsample_nearest | 是 | -| 100 | torch.nn.functional.upsample_bilinear | 是 | -| 101 | torch.nn.functional.grid_sample | 是 | -| 102 | torch.nn.functional.affine_grid | 是 | -| 103 | torch.nn.parallel.data_parallel | 否 | +| 57 | torch.nn.functional.hardsigmoid | 否 | +| 58 | torch.nn.functional.hardswish | 否 | +| 59 | torch.nn.functional.feature_alpha_dropout | 否 | +| 60 | torch.nn.functional.silu | 否 | +| 61 | torch.nn.functional.batch_norm | 是 | +| 62 | torch.nn.functional.instance_norm | 是 | +| 63 | torch.nn.functional.layer_norm | 是 | +| 64 | torch.nn.functional.local_response_norm | 是 | +| 65 | torch.nn.functional.normalize | 是 | +| 66 | torch.nn.functional.linear | 是 | +| 67 | torch.nn.functional.bilinear | 是 | +| 68 | torch.nn.functional.dropout | 是 | +| 69 | torch.nn.functional.alpha_dropout | 是 | +| 70 | torch.nn.functional.dropout2d | 是 | +| 71 | torch.nn.functional.dropout3d | 是 | +| 72 | torch.nn.functional.embedding | 是 | +| 73 | torch.nn.functional.embedding_bag | 是 | +| 74 | torch.nn.functional.one_hot | 是 | +| 75 | torch.nn.functional.pairwise_distance | 是 | +| 76 | torch.nn.functional.cosine_similarity | 是 | +| 77 | torch.nn.functional.pdist | 是 | +| 78 | torch.nn.functional.binary_cross_entropy | 是 | +| 79 | torch.nn.functional.binary_cross_entropy_with_logits | 是 | +| 80 | torch.nn.functional.poisson_nll_loss | 是 | +| 81 | torch.nn.functional.cosine_embedding_loss | 是 | +| 82 | torch.nn.functional.cross_entropy | 是 | +| 83 | torch.nn.functional.ctc_loss | 是（仅支持2维输入） | +| 84 | torch.nn.functional.hinge_embedding_loss | 是 | +| 85 | torch.nn.functional.kl_div | 是 | +| 86 | torch.nn.functional.l1_loss | 是 | +| 87 | torch.nn.functional.mse_loss | 是 | +| 88 | torch.nn.functional.margin_ranking_loss | 是 | +| 89 | torch.nn.functional.multilabel_margin_loss | 是 | +| 90 | torch.nn.functional.multilabel_soft_margin_loss | 是 | +| 91 | torch.nn.functional.multi_margin_loss | 否 | +| 92 | torch.nn.functional.nll_loss | 是 | +| 93 | torch.nn.functional.smooth_l1_loss | 是 | +| 94 | torch.nn.functional.soft_margin_loss | 是 | +| 95 | torch.nn.functional.triplet_margin_loss | 是 | +| 96 | torch.nn.functional.triplet_margin_with_distance_loss | 否 | +| 97 | torch.nn.functional.pixel_shuffle | 否 | +| 98 | torch.nn.functional.pad | 是 | +| 99 | torch.nn.functional.interpolate | 是 | +| 100 | torch.nn.functional.upsample | 是 | +| 101 | torch.nn.functional.upsample_nearest | 是 | +| 102 | torch.nn.functional.upsample_bilinear | 是 | +| 103 | torch.nn.functional.grid_sample | 是 | +| 104 | torch.nn.functional.affine_grid | 是 | +| 105 | torch.nn.parallel.data_parallel | 否 | ## torch.distributed -| 序号 | API名称 | 是否支持 | -| ---- | --------------------------------------- | -------- | -| 1 | torch.distributed.is_available | 是 | -| 2 | torch.distributed.init_process_group | 是 | -| 3 | torch.distributed.Backend | 是 | -| 4 | torch.distributed.get_backend | 是 | -| 5 | torch.distributed.get_rank | 是 | -| 6 | torch.distributed.get_world_size | 是 | -| 7 | torch.distributed.is_initialized | 是 | -| 8 | torch.distributed.is_mpi_available | 是 | -| 9 | torch.distributed.is_nccl_available | 是 | -| 10 | torch.distributed.new_group | 是 | -| 11 | torch.distributed.Store | 是 | -| 12 | torch.distributed.TCPStore | 是 | -| 13 | torch.distributed.HashStore | 是 | -| 14 | torch.distributed.FileStore | 是 | -| 15 | torch.distributed.PrefixStore | 是 | -| 16 | torch.distributed.Store.set | 是 | -| 17 | torch.distributed.Store.get | 是 | -| 18 | torch.distributed.Store.add | 是 | -| 19 | torch.distributed.Store.wait | 是 | -| 20 | torch.distributed.Store.num_keys | 是 | -| 21 | torch.distributed.Store.set_timeout | 是 | -| 22 | torch.distributed.send | 否 | -| 23 | torch.distributed.recv | 否 | -| 24 | torch.distributed.isend | 否 | -| 25 | torch.distributed.irecv | 否 | -| 26 | is_completed | 是 | -| 27 | wait | 是 | -| 28 | torch.distributed.broadcast | 是 | -| 29 | torch.distributed.broadcast_object_list | 是 | -| 30 | torch.distributed.all_reduce | 是 | -| 31 | torch.distributed.reduce | 否 | -| 32 | torch.distributed.all_gather | 是 | -| 33 | torch.distributed.all_gather_object | 是 | -| 34 | torch.distributed.gather | 否 | -| 35 | torch.distributed.scatter | 否 | -| 36 | torch.distributed.scatter_object_list | 是 | -| 37 | torch.distributed.reduce_scatter | 是 | -| 38 | torch.distributed.all_to_all | 是 | -| 39 | torch.distributed.barrier | 是 | -| 40 | torch.distributed.ReduceOp | 是 | -| 41 | torch.distributed.reduce_op | 是 | -| 42 | torch.distributed.broadcast_multigpu | 否 | -| 43 | torch.distributed.all_reduce_multigpu | 否 | -| 44 | torch.distributed.reduce_multigpu | 否 | -| 45 | torch.distributed.all_gather_multigpu | 否 | -| 46 | torch.distributed.launch | 是 | -| 47 | torch.multiprocessing.spawn | 是 | +| 序号 | API名称 | 是否支持 | +| ---- | ----------------------------------------- | -------- | +| 1 | torch.distributed.is_available | 否 | +| 2 | torch.distributed.init_process_group | 是 | +| 3 | torch.distributed.Backend | 是 | +| 4 | torch.distributed.get_backend | 是 | +| 5 | torch.distributed.get_rank | 是 | +| 6 | torch.distributed.get_world_size | 是 | +| 7 | torch.distributed.is_initialized | 是 | +| 8 | torch.distributed.is_mpi_available | 是 | +| 9 | torch.distributed.is_nccl_available | 是 | +| 10 | torch.distributed.new_group | 是 | +| 11 | torch.distributed.Store | 否 | +| 12 | torch.distributed.TCPStore | 否 | +| 13 | torch.distributed.HashStore | 否 | +| 14 | torch.distributed.FileStore | 否 | +| 15 | torch.distributed.PrefixStore | 否 | +| 16 | torch.distributed.Store.set | 否 | +| 17 | torch.distributed.Store.get | 否 | +| 18 | torch.distributed.Store.add | 否 | +| 19 | torch.distributed.Store.wait | 否 | +| 20 | torch.distributed.Store.num_keys | 否 | +| 21 | torch.distributed.Store.delete_keys | 否 | +| 22 | torch.distributed.Store.set_timeout | 否 | +| 23 | torch.distributed.send | 否 | +| 24 | torch.distributed.recv | 否 | +| 25 | torch.distributed.isend | 否 | +| 26 | torch.distributed.irecv | 否 | +| 27 | is_completed | 是 | +| 28 | wait | 是 | +| 29 | torch.distributed.broadcast | 是 | +| 30 | torch.distributed.broadcast_object_list | 否 | +| 31 | torch.distributed.all_reduce | 是 | +| 32 | torch.distributed.reduce | 否 | +| 33 | torch.distributed.all_gather | 是 | +| 34 | torch.distributed.all_gather_object | 否 | +| 35 | torch.distributed.gather_object | 否 | +| 36 | torch.distributed.gather | 否 | +| 37 | torch.distributed.scatter | 否 | +| 38 | torch.distributed.scatter_object_list | 否 | +| 39 | torch.distributed.reduce_scatter | 否 | +| 40 | torch.distributed.reduce_scatter_multigpu | 否 | +| 41 | torch.distributed.all_to_all | 否 | +| 42 | torch.distributed.barrier | 是 | +| 43 | torch.distributed.ReduceOp | 是 | +| 44 | torch.distributed.reduce_op | 是 | +| 45 | torch.distributed.broadcast_multigpu | 否 | +| 46 | torch.distributed.all_reduce_multigpu | 否 | +| 47 | torch.distributed.reduce_multigpu | 否 | +| 48 | torch.distributed.all_gather_multigpu | 否 | +| 49 | torch.distributed.launch | 是 | +| 50 | torch.multiprocessing.spawn | 是 | ## torch.npu | 序号 | API名称 | npu对应API名称 | 是否支持 | | ---- | --------------------------------------------- | ------------------------------------------------ | -------- | -| 1 | torch.cuda.can_device_access_peer | torch_npu.pnu.can_device_access_peer | 是 | +| 1 | torch.cuda.can_device_access_peer | torch_npu.pnu.can_device_access_peer | 否 | | 2 | torch.cuda.current_blas_handle | torch_npu.npu.current_blas_handle | 否 | | 3 | torch.cuda.current_device | torch_npu.npu.current_device | 是 | | 4 | torch.cuda.current_stream | torch_npu.npu.current_stream | 是 | @@ -1293,7 +1293,7 @@ | 9 | torch.cuda.get_device_capability | torch_npu.npu.get_device_capability | 否 | | 10 | torch.cuda.get_device_name | torch_npu.npu.get_device_name | 否 | | 11 | torch.cuda.get_device_properties | torch_npu.npu.get_device_properties | 是 | -| 12 | torch.cuda.get_gencode_flags | torch_npu.npu.get_gencode_flags | 是 | +| 12 | torch.cuda.get_gencode_flags | torch_npu.npu.get_gencode_flags | 否 | | 13 | torch.cuda.init | torch_npu.npu.init | 是 | | 14 | torch.cuda.ipc_collect | torch_npu.npu.ipc_collect | 否 | | 15 | torch.cuda.is_available | torch_npu.npu.is_available | 是 | @@ -1301,71 +1301,73 @@ | 17 | torch.cuda.set_device | torch_npu.npu.set_device | 部分支持 | | 18 | torch.cuda.stream | torch_npu.npu.stream | 是 | | 19 | torch.cuda.synchronize | torch_npu.npu.synchronize | 是 | -| 20 | torch.cuda.get_rng_state | torch_npu.npu.get_rng_state | 否 | -| 21 | torch.cuda.get_rng_state_all | torch_npu.npu.get_rng_state_all | 否 | -| 22 | torch.cuda.set_rng_state | torch_npu.npu.set_rng_state | 否 | -| 23 | torch.cuda.set_rng_state_all | torch_npu.npu.set_rng_state_all | 否 | -| 24 | torch.cuda.manual_seed | torch_npu.npu.manual_seed | 否 | -| 25 | torch.cuda.manual_seed_all | torch_npu.npu.manual_seed_all | 否 | -| 26 | torch.cuda.seed | torch_npu.npu.seed | 否 | -| 27 | torch.cuda.seed_all | torch_npu.npu.seed_all | 否 | -| 28 | torch.cuda.initial_seed | torch_npu.npu.initial_seed | 否 | -| 29 | torch.cuda.comm.broadcast | torch_npu.npu.comm.broadcast | 否 | -| 30 | torch.cuda.comm.broadcast_coalesced | torch_npu.npu.comm.broadcast_coalesced | 否 | -| 31 | torch.cuda.comm.reduce_add | torch_npu.npu.comm.reduce_add | 否 | -| 32 | torch.cuda.comm.scatter | torch_npu.npu.comm.scatter | 否 | -| 33 | torch.cuda.comm.gather | torch_npu.npu.comm.gather | 否 | -| 34 | torch.cuda.Stream | torch_npu.npu.Stream | 是 | -| 35 | torch.cuda.Stream.query | torch_npu.npu.Stream.query | 否 | -| 36 | torch.cuda.Stream.record_event | torch_npu.npu.Stream.record_event | 是 | -| 37 | torch.cuda.Stream.synchronize | torch_npu.npu.Stream.synchronize | 是 | -| 38 | torch.cuda.Stream.wait_event | torch_npu.npu.Stream.wait_event | 是 | -| 39 | torch.cuda.Stream.wait_stream | torch_npu.npu.Stream.wait_stream | 是 | -| 40 | torch.cuda.Event | torch_npu.npu.Event | 是 | -| 41 | torch.cuda.Event.elapsed_time | torch_npu.npu.Event.elapsed_time | 是 | -| 42 | torch.cuda.Event.from_ipc_handle | torch_npu.npu.Event.from_ipc_handle | 否 | -| 43 | torch.cuda.Event.ipc_handle | torch_npu.npu.Event.ipc_handle | 否 | -| 44 | torch.cuda.Event.query | torch_npu.npu.Event.query | 是 | -| 45 | torch.cuda.Event.record | torch_npu.npu.Event.record | 是 | -| 46 | torch.cuda.Event.synchronize | torch_npu.npu.Event.synchronize | 是 | -| 47 | torch.cuda.Event.wait | torch_npu.npu.Event.wait | 是 | -| 48 | torch.cuda.empty_cache | torch_npu.npu.empty_cache | 是 | -| 49 | torch.cuda.memory_stats | torch_npu.npu.memory_stats | 是 | -| 50 | torch.cuda.memory_summary | torch_npu.npu.memory_summary | 是 | -| 51 | torch.cuda.memory_snapshot | torch_npu.npu.memory_snapshot | 是 | -| 52 | torch.cuda.memory_allocated | torch_npu.npu.memory_allocated | 是 | -| 53 | torch.cuda.max_memory_allocated | torch_npu.npu.max_memory_allocated | 是 | -| 54 | torch.cuda.reset_max_memory_allocated | torch_npu.npu.reset_max_memory_allocated | 是 | -| 55 | torch.cuda.memory_reserved | torch_npu.npu.memory_reserved | 是 | -| 56 | torch.cuda.max_memory_reserved | torch_npu.npu.max_memory_reserved | 是 | -| 57 | torch.cuda.set_per_process_memory_fraction | torch_npu.npu.set_per_process_memory_fraction | 是 | -| 58 | torch.cuda.memory_cached | torch_npu.npu.memory_cached | 是 | -| 59 | torch.cuda.max_memory_cached | torch_npu.npu.max_memory_cached | 是 | -| 60 | torch.cuda.reset_max_memory_cached | torch_npu.npu.reset_max_memory_cached | 是 | -| 61 | torch.cuda.nvtx.mark | torch_npu.npu.nvtx.mark | 否 | -| 62 | torch.cuda.nvtx.range_push | torch_npu.npu.nvtx.range_push | 否 | -| 63 | torch.cuda.nvtx.range_pop | torch_npu.npu.nvtx.range_pop | 否 | -| 64 | torch.cuda.amp.autocast | torch_npu.npu.amp.autocast | 是 | -| 65 | torch.cuda.amp.custom_fwd | torch_npu.npu.amp.custom_fwd | 是 | -| 66 | torch.cuda.amp.custom_bwd | torch_npu.npu.amp.custom_bwd | 是 | -| 67 | torch.cuda._sleep | torch_npu.npu._sleep | 否 | -| 68 | torch.cuda.Stream.priority_range | torch_npu.npu.Stream.priority_range | 否 | -| 69 | torch.cuda.get_device_properties | torch_npu.npu.get_device_properties | 否 | -| 70 | torch.cuda.amp.GradScaler | torch_npu.npu.amp.GradScaler | 否 | -| 71 | torch.cuda.amp.GradScaler.get_backoff_factor | torch_npu.npu.amp.GradScaler.get_backoff_factor | 是 | -| 72 | torch.cuda.amp.GradScaler.get_growth_factor | torch_npu.npu.amp.GradScaler.get_growth_factor | 是 | -| 73 | torch.cuda.amp.GradScaler.get_growth_interval | torch_npu.npu.amp.GradScaler.get_growth_interval | 是 | -| 74 | torch.cuda.amp.GradScaler.get_scale | torch_npu.npu.amp.GradScaler.get_scale | 是 | -| 75 | torch.cuda.amp.GradScaler.is_enabled | torch_npu.npu.amp.GradScaler.is_enabled | 是 | -| 76 | torch.cuda.amp.GradScaler.load_state_dict | torch_npu.npu.amp.GradScaler.load_state_dict | 是 | -| 77 | torch.cuda.amp.GradScaler.scale | torch_npu.npu.amp.GradScaler.scale | 是 | -| 78 | torch.cuda.amp.GradScaler.set_backoff_factor | torch_npu.npu.amp.GradScaler.set_backoff_factor | 是 | -| 79 | torch.cuda.amp.GradScaler.set_growth_factor | torch_npu.npu.amp.GradScaler.set_growth_factor | 是 | -| 80 | torch.cuda.amp.GradScaler.set_growth_interval | torch_npu.npu.amp.GradScaler.set_growth_interval | 是 | -| 81 | torch.cuda.amp.GradScaler.state_dict | torch_npu.npu.amp.GradScaler.state_dict | 是 | -| 82 | torch.cuda.amp.GradScaler.step | torch_npu.npu.amp.GradScaler.step | 是 | -| 83 | torch.cuda.amp.GradScaler.unscale_ | torch_npu.npu.amp.GradScaler.unscale_ | 是 | -| 84 | torch.cuda.amp.GradScaler.update | torch_npu.npu.amp.GradScaler.update | 是 | +| 20 | torch.cuda.get_arch_list | torch_npu.npu.get_arch_list | 否 | +| 21 | torch.cuda.get_rng_state | torch_npu.npu.get_rng_state | 否 | +| 22 | torch.cuda.get_rng_state_all | torch_npu.npu.get_rng_state_all | 否 | +| 23 | torch.cuda.set_rng_state | torch_npu.npu.set_rng_state | 否 | +| 24 | torch.cuda.set_rng_state_all | torch_npu.npu.set_rng_state_all | 否 | +| 25 | torch.cuda.manual_seed | torch_npu.npu.manual_seed | 否 | +| 26 | torch.cuda.manual_seed_all | torch_npu.npu.manual_seed_all | 否 | +| 27 | torch.cuda.seed | torch_npu.npu.seed | 否 | +| 28 | torch.cuda.seed_all | torch_npu.npu.seed_all | 否 | +| 29 | torch.cuda.initial_seed | torch_npu.npu.initial_seed | 否 | +| 30 | torch.cuda.comm.broadcast | torch_npu.npu.comm.broadcast | 否 | +| 31 | torch.cuda.comm.broadcast_coalesced | torch_npu.npu.comm.broadcast_coalesced | 否 | +| 32 | torch.cuda.comm.reduce_add | torch_npu.npu.comm.reduce_add | 否 | +| 33 | torch.cuda.comm.scatter | torch_npu.npu.comm.scatter | 否 | +| 34 | torch.cuda.comm.gather | torch_npu.npu.comm.gather | 否 | +| 35 | torch.cuda.Stream | torch_npu.npu.Stream | 是 | +| 36 | torch.cuda.Stream.query | torch_npu.npu.Stream.query | 否 | +| 37 | torch.cuda.Stream.record_event | torch_npu.npu.Stream.record_event | 是 | +| 38 | torch.cuda.Stream.synchronize | torch_npu.npu.Stream.synchronize | 是 | +| 39 | torch.cuda.Stream.wait_event | torch_npu.npu.Stream.wait_event | 是 | +| 40 | torch.cuda.Stream.wait_stream | torch_npu.npu.Stream.wait_stream | 是 | +| 41 | torch.cuda.Event | torch_npu.npu.Event | 是 | +| 42 | torch.cuda.Event.elapsed_time | torch_npu.npu.Event.elapsed_time | 是 | +| 43 | torch.cuda.Event.from_ipc_handle | torch_npu.npu.Event.from_ipc_handle | 否 | +| 44 | torch.cuda.Event.ipc_handle | torch_npu.npu.Event.ipc_handle | 否 | +| 45 | torch.cuda.Event.query | torch_npu.npu.Event.query | 是 | +| 46 | torch.cuda.Event.record | torch_npu.npu.Event.record | 是 | +| 47 | torch.cuda.Event.synchronize | torch_npu.npu.Event.synchronize | 是 | +| 48 | torch.cuda.Event.wait | torch_npu.npu.Event.wait | 是 | +| 49 | torch.cuda.empty_cache | torch_npu.npu.empty_cache | 是 | +| 50 | torch.cuda.list_gpu_processes | torch_npu.npu.list_gpu_processes | 否 | +| 51 | torch.cuda.memory_stats | torch_npu.npu.memory_stats | 是 | +| 52 | torch.cuda.memory_summary | torch_npu.npu.memory_summary | 是 | +| 53 | torch.cuda.memory_snapshot | torch_npu.npu.memory_snapshot | 是 | +| 54 | torch.cuda.memory_allocated | torch_npu.npu.memory_allocated | 是 | +| 55 | torch.cuda.max_memory_allocated | torch_npu.npu.max_memory_allocated | 是 | +| 56 | torch.cuda.reset_max_memory_allocated | torch_npu.npu.reset_max_memory_allocated | 是 | +| 57 | torch.cuda.memory_reserved | torch_npu.npu.memory_reserved | 是 | +| 58 | torch.cuda.max_memory_reserved | torch_npu.npu.max_memory_reserved | 是 | +| 59 | torch.cuda.set_per_process_memory_fraction | torch_npu.npu.set_per_process_memory_fraction | 否 | +| 60 | torch.cuda.memory_cached | torch_npu.npu.memory_cached | 是 | +| 61 | torch.cuda.max_memory_cached | torch_npu.npu.max_memory_cached | 是 | +| 62 | torch.cuda.reset_max_memory_cached | torch_npu.npu.reset_max_memory_cached | 是 | +| 63 | torch.cuda.nvtx.mark | torch_npu.npu.nvtx.mark | 否 | +| 64 | torch.cuda.nvtx.range_push | torch_npu.npu.nvtx.range_push | 否 | +| 65 | torch.cuda.nvtx.range_pop | torch_npu.npu.nvtx.range_pop | 否 | +| 66 | torch.cuda.amp.autocast | torch_npu.npu.amp.autocast | 否 | +| 67 | torch.cuda.amp.custom_fwd | torch_npu.npu.amp.custom_fwd | 否 | +| 68 | torch.cuda.amp.custom_bwd | torch_npu.npu.amp.custom_bwd | 否 | +| 69 | torch.cuda._sleep | torch_npu.npu._sleep | 否 | +| 70 | torch.cuda.Stream.priority_range | torch_npu.npu.Stream.priority_range | 否 | +| 71 | torch.cuda.get_device_properties | torch_npu.npu.get_device_properties | 否 | +| 72 | torch.cuda.amp.GradScaler | torch_npu.npu.amp.GradScaler | 否 | +| 73 | torch.cuda.amp.GradScaler.get_backoff_factor | torch_npu.npu.amp.GradScaler.get_backoff_factor | 否 | +| 74 | torch.cuda.amp.GradScaler.get_growth_factor | torch_npu.npu.amp.GradScaler.get_growth_factor | 否 | +| 75 | torch.cuda.amp.GradScaler.get_growth_interval | torch_npu.npu.amp.GradScaler.get_growth_interval | 否 | +| 76 | torch.cuda.amp.GradScaler.get_scale | torch_npu.npu.amp.GradScaler.get_scale | 否 | +| 77 | torch.cuda.amp.GradScaler.is_enabled | torch_npu.npu.amp.GradScaler.is_enabled | 否 | +| 78 | torch.cuda.amp.GradScaler.load_state_dict | torch_npu.npu.amp.GradScaler.load_state_dict | 否 | +| 79 | torch.cuda.amp.GradScaler.scale | torch_npu.npu.amp.GradScaler.scale | 否 | +| 80 | torch.cuda.amp.GradScaler.set_backoff_factor | torch_npu.npu.amp.GradScaler.set_backoff_factor | 否 | +| 81 | torch.cuda.amp.GradScaler.set_growth_factor | torch_npu.npu.amp.GradScaler.set_growth_factor | 否 | +| 82 | torch.cuda.amp.GradScaler.set_growth_interval | torch_npu.npu.amp.GradScaler.set_growth_interval | 否 | +| 83 | torch.cuda.amp.GradScaler.state_dict | torch_npu.npu.amp.GradScaler.state_dict | 否 | +| 84 | torch.cuda.amp.GradScaler.step | torch_npu.npu.amp.GradScaler.step | 否 | +| 85 | torch.cuda.amp.GradScaler.unscale_ | torch_npu.npu.amp.GradScaler.unscale_ | 否 | +| 86 | torch.cuda.amp.GradScaler.update | torch_npu.npu.amp.GradScaler.update | 否 | torch_npu.npu.set_device()接口只支持在程序开始的位置通过set_device进行指定，不支持多次指定和with torch_npu.npu.device(id)方式的device切换 diff --git "a/docs/zh/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227.md" "b/docs/zh/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227.md" index e95466c3e1..4282b54b45 100644 --- "a/docs/zh/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227.md" +++ "b/docs/zh/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227/PyTorch\345\256\211\350\243\205\346\214\207\345\215\227.md" @@ -197,11 +197,10 @@ ``` export DYNAMIC_COMPILE_ENABLE=1 # 动态shape特性功能，针对shape变化场景，可选，开启设置为1（PyTorch1.8.1不支持该环境变量） export COMBINED_ENABLE=1 # 非连续两个算子组合类场景优化，可选，开启设置为1 - export TRI_COMBINED_ENABLE=1 # 非连续三个算子组合类场景优化，可选，开启设置为1 export ACL_DUMP_DATA=1 # 算子数据dump功能，调试时使用，可选，开启设置为1 export DYNAMIC_OP="ADD#MUL" # 算子实现，ADD和MUL算子在不同场景下有不同的性能表现。可选 ``` - + 4. （可选）当系统为openEuler及其继承操作系统时，如UOS，需设置此命令，取消CPU绑核。 ``` @@ -296,11 +295,6 @@

（可选）非连续两个算子组合类场景优化，开启设置为1。

-

RI_COMBINED_ENABLE

- -

（可选）非连续三个算子组合类场景优化，开启设置为1。

- -

ACL_DUMP_DATA

（可选）算子数据dump功能，调试时使用，开启设置为1。

@@ -320,6 +314,7 @@ +

安装混合精度模块

#### 前提条件 diff --git "a/docs/zh/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227.md" "b/docs/zh/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227.md" index ea1cff16ac..0d105480d3 100644 --- "a/docs/zh/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227.md" +++ "b/docs/zh/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227/PyTorch\347\275\221\347\273\234\346\250\241\345\236\213\347\247\273\346\244\215&\350\256\255\347\273\203\346\214\207\345\215\227.md" @@ -149,7 +149,7 @@ 模型是否可以迁移成功主要取决于模型算子是否支持昇腾AI处理器。故需要对模型算子对昇腾AI处理器的支持性进行评估，一般有两种方式评估算子支持性 -- 模型迁移前，使用dump op方法获取算子信息，与《PyTorch适配算子清单》算子进行比较，确定是否支持。 +- 模型迁移前，使用dump op方法获取算子信息，与《PyTorch API 支持清单》中自定义算子进行比较，确定是否支持。 - 模型迁移后，在昇腾设备上进行运行训练脚本，若存在不支持昇腾AI设备的算子，会提示报错信息。 若存在不支持算子，可以采用修该模型用等价支持的算子替换或者参考《PyTorch算子开发指南》进行算子开发。 @@ -724,7 +724,7 @@ python3 main.py /home/data/resnet50/imagenet --addr='1.1.1.1' \ #

工具迁移

-Ascend平台提供了脚本转换工具使用户能通过命令行方式将训练脚本迁移到昇腾AI处理器上进行训练，命令行方式工具详细使用说明参见下文。除命令行方式外，用户也可通过MindStudio中集成的PyTorch GPU2Ascend功能进行迁移，详情请参见《MindStudio 用户指南》。 +Ascend平台提供了脚本转换工具使用户能通过命令行方式将训练脚本迁移到昇腾AI处理器上进行训练，命令行方式工具详细使用说明参见下文。除命令行方式外，用户也可通过MindStudio中集成的PyTorch GPU2Ascend功能进行迁移，详情请参见[《MindStudio 用户指南》](https://www.hiascend.com/document/detail/zh/mindstudio/304/msug)。

功能介绍

@@ -738,299 +738,17 @@ Ascend平台提供了脚本转换工具使用户能通过命令行方式将训 >- [表1](#zh-cn_topic_0000001133095885_table4705239194613)里的脚本转换后的执行逻辑与转换前保持一致。 >- 此脚本转换工具当前仅支持PyTorch训练脚本转换。 -**表 1** 模型支持列表 +##### 模型支持 + +目前支持模型请参考[《昇腾Modelzoo社区》](https://www.hiascend.com/software/modelzoo) ，筛选类型分类：“训练”、框架分类：“Pytorch”的Pytorch训练模型。 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

序号

-

模型名称

-

1

-

3D AttentionNet

-

2

-

3D Nested_UNet

-

3

-

Advanced East

-

4

-

AlexNet

-

5

-

DeeplabV3+(Xception-JFT)

-

6

-

DeepMar

-

7

-

Densenet121

-

8

-

DenseNet161

-

9

-

DenseNet169

-

10

-

DenseNet201

-

11

-

EAST

-

12

-

FCN

-

13

-

FD-GAN

-

14

-

FOTS

-

15

-

GENet

-

16

-

GoogleNet

-

17

-

GRU

-

18

-

Inception V4

-

19

-

InceptionV2

-

20

-

LPRNet

-

21

-

LSTM

-

22

-

MNASNet0_5

-

23

-

MNASNet0_75

-

24

-

MNASNet1_0

-

25

-

MNASNet1_3

-

26

-

MobileNetV1

-

27

-

MobileNetV2

-

28

-

PNet

-

29

-

PSENet

-

30

-

RAFT

-

31

-

RecVAE

-

32

-

ResNet101

-

33

-

ResNet152

-

34

-

ResNet18

-

35

-

ResNet34

-

36

-

ResNet50

-

37

-

Resnext101_32x8d

-

38

-

Resnext50

-

39

-

RNet

-

40

-

Shufflenetv2

-

41

-

SqueezeNet1_0

-

42

-

SqueezeNet1_1

-

43

-

U-Net

-

44

-

VAE+GAN

-

45

-

VGG11

-

46

-

VGG11_BN

-

47

-

VGG13

-

48

-

VGG13_BN

-

49

-

VGG16

-

50

-

VGG16_BN

-

51

-

VGG19

-

52

-

VGG19_BN

-

53

-

VIT-base

-

54

-

Wide_ResNet101_2

-

55

-

Wide_ResNet50_2

-
**系统要求** 脚本转换工具支持Ubuntu 18.04、CentOS 7.6或EulerOS 2.8。 **环境准备** -详情请参考《CANN 软件安装指南》安装开发环境。 +详情请参考[《CANN 软件安装指南》](https://www.hiascend.com/document/detail/zh/canncommercial/504/envdeployment/instg) 安装开发环境。

操作指南

@@ -1749,7 +1467,7 @@ Pytorch1.8.1版本的AMP，类似于Apex AMP的O1模式（动态 loss scale） 2. 解析性能数据文件。 - 请参见《CANN 开发辅助工具指南》中“Profiling工具使用指南（训练）”章节。 + 请参见《CANN 开发辅助工具指南》中“Profiling工具使用指南>高级功能（所有性能调优方式和采集项）>数据解析与导出”章节。 @@ -2227,6 +1945,8 @@ with torch.npu.profile(profiler_result_path="./results", use_e2e_profiler=True

前提条件

+目前pytorch1.8.1暂不支持。 + 优先在同等语义和超参下，跑一定的epoch（推荐完整epoch数的20%），使精度，loss等对齐GPU相应水平，完成后再对齐最终精度。

调测过程

-- Gitee