search

open-mmlab / mmdetection

OpenMMLab Detection Toolbox and Benchmark
https://mmdetection.readthedocs.io
Apache License 2.0
29.56k stars 9.46k forks source link

mask2former head fails during training because of amp (ValueError: cost matrix is infeasible) #10909

Open we11as22 opened 1 year ago

we11as22 commented 1 year ago

Thanks for your error report and we appreciate it a lot.

Checklist

  1. I have searched related issues but cannot get the expected help.
  2. I have read the FAQ documentation but cannot get the expected help.
  3. The bug has not been fixed in the latest version.

Describe the bug A clear and concise description of what the bug is.

Reproduction

  1. What command or script did you run?
A placeholder for the command.
  1. Did you make any modifications on the code or config? Did you understand what you have modified?
  2. What dataset did you use?

Environment

  1. Please run python mmdet/utils/collect_env.py to collect necessary environment information and paste it here.
  2. You may add addition that may be helpful for locating the problem, such as
    • How you installed PyTorch [e.g., pip, conda, source]
    • Other environment variables that may be related (such as $PATH, $LD_LIBRARY_PATH, $PYTHONPATH, etc.)

Error traceback If applicable, paste the error trackback here.

A placeholder for trackback.

Bug fix If you have already identified the reason, you can provide the information here. If you are willing to create a PR to fix it, please also leave a comment here and that would be much appreciated!

hhaAndroid commented 1 year ago

@AlexeySudakovB01-109 Thank you for the feedback. We will work on fixing it as soon as possible.

liuxingbin commented 10 months ago

any update?

leo-q8 commented 7 months ago

Looking forward to further work！

leo-q8 commented 7 months ago

@we11as22 @hhaAndroid I am facing the same issue, when training Mask2Former with --amp. And I have found a solution here huggingface/transformers#21644. As it proposed, this is caused by how scipy.optimize.linear_sum_assignment handles infinite values. Replacing these with very large numbers seems to fix the issue. And I added the following code before line 125.

cost = torch.minimum(cost, torch.tensor(1e10))
cost = torch.maximum(cost, torch.tensor(-1e10))

https://github.com/open-mmlab/mmdetection/blob/cfd5d3a985b0249de009b67d04f37263e11cdf3d/mmdet/models/task_modules/assigners/hungarian_assigner.py#L123-L131

But when trying to train with amp again, I encountered the NaN mask loss problem as follows: image

Further debugging revealed issues in the calculation of loss_mask in this line. In amp mode, multiplying the FP16 variable num_total_masks by a large integer variable self.num_points (it is 12544 in my configs)results in an overflowed FP16 variable, further leading to the occurrence of NaN mask loss. image By casting num_total_masks to the fp32 type with avg_factor=num_total_masks.float() * self.num_points, I managed to eliminate this issue. The model can now be successfully trained in amp mode, but the accuracy is still under validation 😂 . Hope this is helpful to you！