Exclude torch==1.12.0, torchvision==0.13.0 (Fix #8395)

[mjun0812]

Fix #8395

There is a bug in PyTorch 1.12.0 regarding CUDA initialization.
Therefore, the environment variable "CUDA_VISIBLE_DEVICES" set after import torch will not work.

This is a fatal problem because YOLOv5 uses CUDA_VISIBLE_DEVICES to allocate GPUs.

yolov5/utils/torch_utils.py

Lines 52 to 85 in fdc9d91

	def select_device(device='', batch_size=0, newline=True):
	# device = None or 'cpu' or 0 or '0' or '0,1,2,3'
	s = f'YOLOv5 🚀 {git_describe() or file_date()} Python-{platform.python_version()} torch-{torch.__version__} '
	device = str(device).strip().lower().replace('cuda:', '').replace('none', '') # to string, 'cuda:0' to '0'
	cpu = device == 'cpu'
	mps = device == 'mps' # Apple Metal Performance Shaders (MPS)
	if cpu or mps:
	os.environ['CUDA_VISIBLE_DEVICES'] = '-1' # force torch.cuda.is_available() = False
	elif device: # non-cpu device requested
	os.environ['CUDA_VISIBLE_DEVICES'] = device # set environment variable - must be before assert is_available()
	assert torch.cuda.is_available() and torch.cuda.device_count() >= len(device.replace(',', '')), \
	f"Invalid CUDA '--device {device}' requested, use '--device cpu' or pass valid CUDA device(s)"
	if not (cpu or mps) and torch.cuda.is_available(): # prefer GPU if available
	devices = device.split(',') if device else '0' # range(torch.cuda.device_count()) # i.e. 0,1,6,7
	n = len(devices) # device count
	if n > 1 and batch_size > 0: # check batch_size is divisible by device_count
	assert batch_size % n == 0, f'batch-size {batch_size} not multiple of GPU count {n}'
	space = ' ' * (len(s) + 1)
	for i, d in enumerate(devices):
	p = torch.cuda.get_device_properties(i)
	s += f"{'' if i == 0 else space}CUDA:{d} ({p.name}, {p.total_memory / (1 << 20):.0f}MiB)\n" # bytes to MB
	arg = 'cuda:0'
	elif mps and getattr(torch, 'has_mps', False) and torch.backends.mps.is_available(): # prefer MPS if available
	s += 'MPS\n'
	arg = 'mps'
	else: # revert to CPU
	s += 'CPU\n'
	arg = 'cpu'
	if not newline:
	s = s.rstrip()
	LOGGER.info(s.encode().decode('ascii', 'ignore') if platform.system() == 'Windows' else s) # emoji-safe
	return torch.device(arg)

As mentioned in this Issue, the next version of PyTorch 1.12.1 will resolve this issue, so it is better to remove the buggy version 1.12.0 and the associated torchvision version from the requirements.txt file.

If you have any comments, please do not hesitate to let me know.
Thanks.

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Updated PyTorch requirements to exclude buggy versions.

📊 Key Changes

• Excluded PyTorch version 1.12.0 due to identified issues.
• Excluded torchvision version 0.13.0 as it's not compatible with current codebase.

🎯 Purpose & Impact

• 🎯 Purpose: To prevent installation of specific PyTorch and torchvision versions that are known to cause problems with the YOLOv5 code, ensuring stability and reliability for users.
• 💥 Impact: Users will avoid potential bugs by not installing these problematic versions, leading to a smoother experience with the YOLOv5 project.

[github-actions]

👋 Hello @mjun0812, thank you for submitting a YOLOv5 🚀 PR! To allow your work to be integrated as seamlessly as possible, we advise you to:

• ✅ Verify your PR is up-to-date with upstream/master. If your PR is behind upstream/master an automatic GitHub Actions merge may be attempted by writing /rebase in a new comment, or by running the following code, replacing 'feature' with the name of your local branch:

git remote add upstream https://github.com/ultralytics/yolov5.git
git fetch upstream
# git checkout feature  # <--- replace 'feature' with local branch name
git merge upstream/master
git push -u origin -f

• ✅ Verify all Continuous Integration (CI) checks are passing.
• ✅ Reduce changes to the absolute minimum required for your bug fix or feature addition. "It is not daily increase but daily decrease, hack away the unessential. The closer to the source, the less wastage there is." -Bruce Lee

[glenn-jocher]

@mjun0812 thanks for the investigation and the PR. PR is merged!