-
-
Notifications
You must be signed in to change notification settings - Fork 16.2k
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.
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
models/export.py setting model.train() may be changed to false in onnx #3346
Comments
👋 Hello @ChaofWang, thank you for your interest in 🚀 YOLOv5! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution. If this is a 🐛 Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you. If this is a custom training ❓ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available. For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com. RequirementsPython 3.8 or later with all requirements.txt dependencies installed, including $ pip install -r requirements.txt EnvironmentsYOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):
StatusIf this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit. |
@ChaofWang hi, thanks for the bug report! You are correct, ONNX export is not operating in train mode for some reason. It seems the ONNX exporter is forcing it back into eval mode. |
@ChaofWang I've implemented a solution per your recommendation by adding two arguments to the onnx export function in export.py L100: training=torch.onnx.TrainingMode.TRAINING if opt.train else torch.onnx.TrainingMode.EVAL,
do_constant_folding=not opt.train, If this solution works for you please submit a PR with this update, thank you! |
TODO: ONNX export in .train() mode fix |
@glenn-jocher hi, this solution works for me. I have submitted PR for this update |
@ChaofWang good news 😃! Your original issue may now be fixed ✅ in PR #3362. To receive this update:
Thank you for spotting this issue and informing us of the problem. Please let us know if this update resolves the issue for you, and feel free to inform us of any other issues you discover or feature requests that come to mind. Happy trainings with YOLOv5 🚀! |
🐛 Bug
I think you may want to have no grid construction in Detect layer by setting model.train() when export model to onnx with --train. But in "torch.onnx.export", model.training seems to be reset to mode.training=False by default.
In fact, the param 'training' in the "torch.onnx.export" can be set to "training=torch.onnx.TrainingMode.TRAINING", so that the model can be set to train for conversion instead of using model.train(). But it doesn't seem to be a good way to be recommended. And I found that this part was correct before version 4.0, but after version 5.0, the "export" was deleted in Detect.
To Reproduce (REQUIRED)
Input:
Output:
Expected behavior
This part should be skipped.
yolov5/models/yolo.py
Lines 50 to 62 in aad99b6
Environment
If applicable, add screenshots to help explain your problem.
The text was updated successfully, but these errors were encountered: