Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

oneflow适配调研 #1

Closed
ccssu opened this issue Apr 3, 2023 · 4 comments
Closed

oneflow适配调研 #1

ccssu opened this issue Apr 3, 2023 · 4 comments

Comments

@ccssu
Copy link
Collaborator

ccssu commented Apr 3, 2023 •
edited
Loading

本报告调研了github地址为https://github.com/VainF/Torch-Pruning 的仓库。
主要集中在分析使用oneflow重构仓库中的代码所需的工作量和难点，以及可能遇到的问题

仓库简介

主要目的： is a versatile library that enables structural network pruning for a wide range of neural networks, including Vision Transformers, Yolov7, FasterRCNN, SSD, ResNet, DenseNet, ConvNext, RegNet, ResNext, FCN, DeepLab, VGG, etc.
项目目录介绍:
benchmarks
[源仓库在计划开发中]https://github.com/Oneflow-Inc/OneFlow-Pruning#plans

tests 测试
torch_pruning 定义的剪枝工具
tutorials

适配开发暂时集中在 test torch_pruning tutorials中。

仓库功能

https://github.com/Oneflow-Inc/OneFlow-Pruning#features

oneflow适配缺失模块

  1. AttributeError: module 'oneflow.nn.modules.conv' has no attribute '_ConvNd'
  2. AttributeError: module 'oneflow.nn' has no attribute 'MultiheadAttention'
  3. AttributeError: type object 'Tensor' has no attribute 'char'
  4. AttributeError: type object 'Tensor' has no attribute 'short'
  5. AttributeError: type object 'Tensor' has no attribute '_has_compatible_shallow_copy_type'
  6. 缺失LazyModuleMixin 模块。 pytorch代码链接 https://github.com/pytorch/pytorch/blob/master/torch/nn/modules/lazy.py
  7. 缺失UninitializedTensorMixin 模块，https://github.com/pytorch/pytorch/blob/76074dc0a3c822298fdf6b444c372efa93e9932a/torch/nn/parameter.py#L84
  8. class UninitializedParameter(UninitializedTensorMixin, Parameter): https://github.com/pytorch/pytorch/blob/76074dc0a3c822298fdf6b444c372efa93e9932a/torch/nn/parameter.py#L168
  9. class UninitializedBuffer(UninitializedTensorMixin, Tensor):https://github.com/pytorch/pytorch/blob/76074dc0a3c822298fdf6b444c372efa93e9932a/torch/nn/parameter.py#L199

结论

在不考虑正确性修复上面模块，如链接所示 Oneflow-Inc/oneflow@master...adapt_to_torch_pruning

测试了 Torch-Pruning/tutorials/0 - QuickStart.ipynb 可以正常执行

测试时运行输出 
(oneflow-dev-clang10-v2) python quick.py 
After pruning:
ResNet(
  (conv1): Conv2d(3, 61, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU()
  (maxpool): MaxPool2d(kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), dilation=(1, 1))
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(61, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(64, 61, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(61, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(64, 61, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(61, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(61, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AdaptiveAvgPool2d()
  (fc): Linear(in_features=512, out_features=1000, bias=True)
)

--------------------------------
          Pruning Group
--------------------------------
[0] prune_out_channels on conv1 (Conv2d(3, 61, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)) => prune_out_channels on conv1 (Conv2d(3, 61, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)), #idxs=3
[1] prune_out_channels on conv1 (Conv2d(3, 61, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)) => prune_out_channels on bn1 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)), #idxs=3
[2] prune_out_channels on bn1 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[3] prune_out_channels on bn1 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[4] prune_out_channels on bn1 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(relu_backward), #idxs=3
[5] prune_out_channels on _ElementWiseOp(relu_backward) => prune_out_channels on _ElementWiseOp(max_pool_2d_backward), #idxs=3
[6] prune_out_channels on _ElementWiseOp(max_pool_2d_backward) => prune_out_channels on _ElementWiseOp(add_n_backward), #idxs=3
[7] prune_out_channels on _ElementWiseOp(max_pool_2d_backward) => prune_in_channels on layer1.0.conv1 (Conv2d(61, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)), #idxs=3
[8] prune_in_channels on layer1.0.conv1 (Conv2d(61, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[9] prune_out_channels on _ElementWiseOp(add_n_backward) => prune_out_channels on layer1.0.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)), #idxs=3
[10] prune_out_channels on _ElementWiseOp(add_n_backward) => prune_out_channels on _ElementWiseOp(relu_backward), #idxs=3
[11] prune_out_channels on _ElementWiseOp(relu_backward) => prune_out_channels on _ElementWiseOp(add_n_backward), #idxs=3
[12] prune_out_channels on _ElementWiseOp(relu_backward) => prune_in_channels on layer1.1.conv1 (Conv2d(61, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)), #idxs=3
[13] prune_in_channels on layer1.1.conv1 (Conv2d(61, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[14] prune_out_channels on _ElementWiseOp(add_n_backward) => prune_out_channels on layer1.1.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)), #idxs=3
[15] prune_out_channels on _ElementWiseOp(add_n_backward) => prune_out_channels on _ElementWiseOp(relu_backward), #idxs=3
[16] prune_out_channels on _ElementWiseOp(relu_backward) => prune_in_channels on layer2.0.downsample.0 (Conv2d(61, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)), #idxs=3
[17] prune_out_channels on _ElementWiseOp(relu_backward) => prune_in_channels on layer2.0.conv1 (Conv2d(61, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)), #idxs=3
[18] prune_in_channels on layer2.0.conv1 (Conv2d(61, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[19] prune_in_channels on layer2.0.downsample.0 (Conv2d(61, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[20] prune_out_channels on layer1.1.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on layer1.1.conv2 (Conv2d(64, 61, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)), #idxs=3
[21] prune_out_channels on layer1.1.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[22] prune_out_channels on layer1.1.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[23] prune_out_channels on layer1.0.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on layer1.0.conv2 (Conv2d(64, 61, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)), #idxs=3
[24] prune_out_channels on layer1.0.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
[25] prune_out_channels on layer1.0.bn2 (BatchNorm2d(61, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=3
--------------------------------

Number of Groups: 13
The last Group: 
--------------------------------
          Pruning Group
--------------------------------
[0] prune_out_channels on layer4.1.conv1 (Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)) => prune_out_channels on layer4.1.conv1 (Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)), #idxs=512
[1] prune_out_channels on layer4.1.conv1 (Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)) => prune_out_channels on layer4.1.bn1 (BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)), #idxs=512
[2] prune_out_channels on layer4.1.bn1 (BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=512
[3] prune_out_channels on layer4.1.bn1 (BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=512
[4] prune_out_channels on layer4.1.bn1 (BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)) => prune_out_channels on _ElementWiseOp(relu_backward), #idxs=512
[5] prune_out_channels on _ElementWiseOp(relu_backward) => prune_in_channels on layer4.1.conv2 (Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)), #idxs=512
[6] prune_in_channels on layer4.1.conv2 (Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)) => prune_out_channels on _ElementWiseOp(accumulate_grad), #idxs=512
--------------------------------

ResNet(
  (conv1): Conv2d(3, 57, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(57, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU()
  (maxpool): MaxPool2d(kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), dilation=(1, 1))
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(57, 57, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(57, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(57, 57, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(57, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(57, 57, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(57, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(57, 57, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(57, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(57, 115, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(115, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(115, 115, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(115, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(57, 115, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(115, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(115, 115, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(115, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(115, 115, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(115, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(115, 230, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(230, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(230, 230, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(230, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(115, 230, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(230, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(230, 230, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(230, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(230, 230, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(230, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(230, 460, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(460, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(460, 460, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(460, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(230, 460, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(460, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(460, 460, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(460, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(460, 460, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(460, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AdaptiveAvgPool2d()
  (fc): Linear(in_features=460, out_features=1000, bias=True)
)
oneflow.Size([1, 1000])
  Iter 1/5, Params: 11.69 M => 9.48 M
  Iter 1/5, MACs: 1.82 G => 1.47 G
ResNet(
  (conv1): Conv2d(3, 51, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(51, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU()
  (maxpool): MaxPool2d(kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), dilation=(1, 1))
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(51, 51, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(51, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(51, 51, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(51, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(51, 51, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(51, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(51, 51, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(51, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(51, 102, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(102, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(102, 102, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(102, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(51, 102, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(102, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(102, 102, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(102, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(102, 102, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(102, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(102, 204, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(204, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(204, 204, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(204, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(102, 204, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(204, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(204, 204, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(204, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(204, 204, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(204, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(204, 409, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(409, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(409, 409, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(409, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(204, 409, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(409, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(409, 409, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(409, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(409, 409, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(409, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AdaptiveAvgPool2d()
  (fc): Linear(in_features=409, out_features=1000, bias=True)
)
oneflow.Size([1, 1000])
  Iter 2/5, Params: 11.69 M => 7.53 M
  Iter 2/5, MACs: 1.82 G => 1.18 G
ResNet(
  (conv1): Conv2d(3, 44, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(44, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU()
  (maxpool): MaxPool2d(kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), dilation=(1, 1))
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(44, 44, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(44, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(44, 44, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(44, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(44, 44, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(44, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(44, 44, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(44, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(44, 89, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(89, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(89, 89, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(89, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(44, 89, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(89, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(89, 89, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(89, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(89, 89, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(89, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(89, 179, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(179, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(179, 179, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(179, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(89, 179, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(179, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(179, 179, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(179, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(179, 179, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(179, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(179, 358, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(358, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(358, 358, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(358, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(179, 358, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(358, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(358, 358, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(358, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(358, 358, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(358, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AdaptiveAvgPool2d()
  (fc): Linear(in_features=358, out_features=1000, bias=True)
)
oneflow.Size([1, 1000])
  Iter 3/5, Params: 11.69 M => 5.82 M
  Iter 3/5, MACs: 1.82 G => 0.91 G
ResNet(
  (conv1): Conv2d(3, 38, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(38, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU()
  (maxpool): MaxPool2d(kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), dilation=(1, 1))
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(38, 38, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(38, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(38, 38, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(38, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(38, 38, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(38, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(38, 38, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(38, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(38, 76, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(76, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(76, 76, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(76, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(38, 76, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(76, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(76, 76, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(76, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(76, 76, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(76, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(76, 153, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(153, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(153, 153, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(153, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(76, 153, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(153, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(153, 153, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(153, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(153, 153, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(153, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(153, 307, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(307, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(307, 307, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(307, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(153, 307, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(307, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(307, 307, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(307, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(307, 307, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(307, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AdaptiveAvgPool2d()
  (fc): Linear(in_features=307, out_features=1000, bias=True)
)
oneflow.Size([1, 1000])
  Iter 4/5, Params: 11.69 M => 4.32 M
  Iter 4/5, MACs: 1.82 G => 0.68 G
ResNet(
  (conv1): Conv2d(3, 32, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
  (bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (relu): ReLU()
  (maxpool): MaxPool2d(kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), dilation=(1, 1))
  (layer1): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
    (1): BasicBlock(
      (conv1): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer2): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(32, 64, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer3): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (layer4): Sequential(
    (0): BasicBlock(
      (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (downsample): Sequential(
        (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
        (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      )
    )
    (1): BasicBlock(
      (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
      (relu): ReLU()
      (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
      (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
    )
  )
  (avgpool): AdaptiveAvgPool2d()
  (fc): Linear(in_features=256, out_features=1000, bias=True)
)
oneflow.Size([1, 1000])
  Iter 5/5, Params: 11.69 M => 3.06 M
  Iter 5/5, MACs: 1.82 G => 0.49 G

plan

逐一解决上面发现的问题，进行仓库适配
@ccssu
Copy link
Collaborator Author

ccssu commented Apr 3, 2023

AttributeError: module 'oneflow.nn.modules.conv' has no attribute '_ConvNd'

项目使用处

TORCH_CONV = nn.modules.conv._ConvNd

https://github.com/VainF/Torch-Pruning/blob/5e7ae60349474d931e92ee074a46119878915c0c/torch_pruning/ops.py#L85

Renferce

https://github.com/pytorch/pytorch/blob/2ab18a23e1c5ef9a5a0b2da82fc5eaeda969ea9f/torch/nn/modules/conv.py#L47

@ccssu
Copy link
Collaborator Author

ccssu commented Apr 3, 2023

AttributeError: module 'oneflow.nn' has no attribute 'MultiheadAttention'

项目使用处

nn.MultiheadAttention: multihead_attention_counter_hook

https://github.com/VainF/Torch-Pruning/blob/5e7ae60349474d931e92ee074a46119878915c0c/torch_pruning/utils/op_counter.py#L294

Renferce

https://github.com/pytorch/pytorch/blob/2ab18a23e1c5ef9a5a0b2da82fc5eaeda969ea9f/torch/nn/modules/activation.py#L908

@BBuf
Copy link

BBuf commented Apr 3, 2023

AttributeError: module 'oneflow.nn.modules.conv' has no attribute '_ConvNd'

项目使用处

TORCH_CONV = nn.modules.conv._ConvNd

https://github.com/VainF/Torch-Pruning/blob/5e7ae60349474d931e92ee074a46119878915c0c/torch_pruning/ops.py#L85

Renferce

https://github.com/pytorch/pytorch/blob/2ab18a23e1c5ef9a5a0b2da82fc5eaeda969ea9f/torch/nn/modules/conv.py#L47

感觉我们可以改成TORCH_CONV = nn.modules.Conv1D or nn.modules.Conv2D or nn.modules.Conv1D 这样子？

@ccssu ccssu mentioned this issue Apr 4, 2023
Closed
14 tasks
@BBuf
Copy link

BBuf commented Apr 8, 2023

Oneflow-Inc/oneflow#10086 short和char这里支持了

@ccssu ccssu closed this as completed Apr 10, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@BBuf @ccssu