Skip to content

Conversation

@littletomatodonkey
Copy link
Collaborator

@littletomatodonkey littletomatodonkey commented Apr 14, 2021

att, four models top1 accs on ImageNet1k are as follows.

base_patch4_window12_384: {'top1_acc': 0.84390}
base_patch4_window7_224: {'top1_acc': 0.8299600000762939}
small_patch4_window7_224: {'top1_acc': 0.82746}
tiny_patch4_window7_224: {'top1_acc': 0.806879999961853}
  • todo

    • support training
  • Current problem

    • If you train the transformer using GPU and encountered the following problem, you can set arch=sm_61, more detailed info can be seen in link
ERROR - (Fatal) Operator uniform_random raises an thrust::system::system_error exception.
The exception content is
:parallel_for failed: cudaErrorInvalidConfiguration: invalid configuration argument. (at /paddle/paddle/fluid/imperative/tracer.cc:172)

@PaddlePaddle PaddlePaddle locked and limited conversation to collaborators Apr 21, 2021
@PaddlePaddle PaddlePaddle unlocked this conversation Apr 21, 2021

class SwinTransformer(nn.Layer):
""" Swin Transformer
A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` -
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

PyTorch -> PaddlePaddle ?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done,thx

@PaddlePaddle PaddlePaddle locked and limited conversation to collaborators Apr 24, 2021
@PaddlePaddle PaddlePaddle unlocked this conversation Apr 24, 2021
@zhangwenwen
Copy link

need to recompile paddle-paddle library manually?

@Darki-luo
Copy link

Darki-luo commented May 27, 2021

att, four models top1 accs on ImageNet1k are as follows.

base_patch4_window12_384: {'top1_acc': 0.84390}
base_patch4_window7_224: {'top1_acc': 0.8299600000762939}
small_patch4_window7_224: {'top1_acc': 0.82746}
tiny_patch4_window7_224: {'top1_acc': 0.806879999961853}
  • todo

    • support training
  • Current problem

    • If you train the transformer using GPU and encountered the following problem, you can set arch=sm_61, more detailed info can be seen in link
ERROR - (Fatal) Operator uniform_random raises an thrust::system::system_error exception.
The exception content is
:parallel_for failed: cudaErrorInvalidConfiguration: invalid configuration argument. (at /paddle/paddle/fluid/imperative/tracer.cc:172)

请问parallel_for failed的问题有跟进修复吗?

@littletomatodonkey littletomatodonkey deleted the dev/add_swin_transformer branch June 10, 2021 15:10
@littletomatodonkey littletomatodonkey restored the dev/add_swin_transformer branch June 22, 2021 04:07
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants