Unable to export weights into onnx

[usamatariq70]

Before submitting a bug report, please be aware that your issue must be reproducible with all of the following, otherwise it is non-actionable, and we can not help you:

• Current repo: run git fetch && git status -uno to check and git pull to update repo
• Common dataset: coco.yaml or coco128.yaml
• Common environment: Colab, Google Cloud, or Docker image. See https://github.com/ultralytics/yolov5#environments

If this is a custom dataset/training question you must include your train*.jpg, test*.jpg and results.png figures, or we can not help you. You can generate these with utils.plot_results().

🐛 Bug

Whenever I run the export command, I get the error which I shared below.

To Reproduce (REQUIRED)

Input:

!export PYTHONPATH="$PWD" && python models/export.py --weights yolov5s.pt --batch 1

Output:

Traceback (most recent call last):
  File "models/export.py", line 33, in <module>
    model = attempt_load(opt.weights, map_location=torch.device('cpu'))  # load FP32 model
  File "./models/experimental.py", line 118, in attempt_load
    model.append(torch.load(w, map_location=map_location)['model'].float().fuse().eval())  # load FP32 model
  File "/usr/local/lib/python3.6/dist-packages/torch/serialization.py", line 584, in load
    return _load(opened_zipfile, map_location, pickle_module, **pickle_load_args)
  File "/usr/local/lib/python3.6/dist-packages/torch/serialization.py", line 842, in _load
    result = unpickler.load()
AttributeError: Can't get attribute 'SiLU' on <module 'torch.nn.modules.activation' from '/usr/local/lib/python3.6/dist-packages/torch/nn/modules/activation.py'>

Expected behavior

That given weights should be converted into onnx weights. Kindly let me know where I am doing anything wrong, or feel free to ask me the changes I made to run the script.

Environment

If applicable, add screenshots to help explain your problem.

• Google Colab
• torch==1.6.0+cu101
• torchsummary==1.5.1
• torchtext==0.3.1
• torchvision==0.7.0+cu101
• onnx==1.8.0
• onnxruntime==1.6.0
• coremltools==4.0

Additional context

I have fixed the problem by making few changes in the files. Also when I upgrade torch to 1.7.0 I get "CoreML export failure: unexpected number of inputs for node x.2 (_ convolution): 13" error. So some issue is there too.

[github-actions]

👋 Hello @usamatariq70, 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.

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Google Colab Notebook with free GPU:
• Kaggle Notebook with free GPU: https://www.kaggle.com/models/ultralytics/yolov5
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Docker Image https://hub.docker.com/r/ultralytics/yolov5. See Docker Quickstart Guide

Status

If 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.

[glenn-jocher]

@usamatariq70 you are not meeting the requirements, specifically torch>=1.7 as the bot already mentioned above.

[usamatariq70]

@usamatariq70 you are not meeting the requirements, specifically torch>=1.7 as the bot already mentioned above.

When I upgrade to torch>=1.7, I get this error "CoreML export failure: unexpected number of inputs for node x.2 (_convolution): 13" on yolov5s.pt

[glenn-jocher]

@usamatariq70 for CoreML you'll have to wait for a new version of coremltools that is compatible with the more recent pytorch versions. You may want to raise an issue directly on the coremltools repo regarding forward compatibility with torch>=1.7.0

[usamatariq70]

Okay thank you

[xujuntdw111]

I meet the same question as you. Either torch=1.6 or torch=1.7 will report an error.
Has the problem been solved？

@usamatariq70 for CoreML you'll have to wait for a new version of coremltools that is compatible with the more recent pytorch versions. You may want to raise an issue directly on the coremltools repo regarding forward compatibility with torch>=1.7.0

[usamatariq70]

I meet the same question as you. Either torch=1.6 or torch=1.7 will report an error.
Has the problem been solved？

@usamatariq70 for CoreML you'll have to wait for a new version of coremltools that is compatible with the more recent pytorch versions. You may want to raise an issue directly on the coremltools repo regarding forward compatibility with torch>=1.7.0

No I reported the issue on CoreML repo but they didnt responded yet. But if you want to convert yolov5 pretrained weights that I have cracked the way