Only when I change the pytorch version to 1.7.0 (default version is 1.8.1) in requirments.txt then model for android is sucessfully exported.

[duwangthefirst]

Basically I'm following the official colab notebook.
the training procedure is all fine.

According to https://github.com/pytorch/android-demo-app/tree/master/ObjectDetection, I updated models/export.py like this:

• Change from model.model[-1].export = not opt.grid to model.model[-1].export = False
• Add the following two lines of model optimization code, between ts = torch.jit.trace(model, img) and ts.save(f):

from torch.utils.mobile_optimizer import optimize_for_mobile
ts = optimize_for_mobile(ts)

and then I run:

python models/export.py --weights runs/train/exp/weights/last.pt --img-size 640 --batch-size 1

then it failed to generate last.torchscript.pt.

Only after I change the requirements.txt from:

torch>=1.7.0
torchvision>=0.8.1

to:

torch==1.7.0
torchvision==0.8.1

and the installed pytorch will be of version 1.7.0 instead of 1.8.1 and finally the last.torchscript.pt is sucessfully generated.

My question is: When will this MODEL EXPORT function be compatible with the latest version of pytorch ?

[github-actions]

👋 Hello @duwangthefirst, 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 and Kaggle notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. 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]

@duwangthefirst I just tested export.py in Colab with PyTorch 1.8.1 and torchscript export works correctly, both with and with --grid:

If you believe you have a reproducible issue, we suggest you close this issue and raise a new one using the 🐛 Bug Report template, providing screenshots and a minimum reproducible example to help us better understand and diagnose your problem. Thank you!

[glenn-jocher]

@duwangthefirst I've added default support for the PyTorch tutorial recommendation of running optimize_for_mobile() during torchscript export in PR #2908.