train.py: error: argument --weights: expected one argument

[LokedSher]

Search before asking

• I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

I want to train a yolov5s on my own dataset and randomly initialize the weights. I have modified coco128.yaml and run the sh
python .\train.py --img 640 --batch 16 --data .\data\coco128.yaml --cfg .\models\yolov5s.yaml --weights '' --device 0 as the tutorial said,then the error occured:
train.py: error: argument --weights: expected one argument
I don't know how to deal with it

Additional

No response

[glenn-jocher]

@LokedSher some environments might need "" instead of ''

[tianty]

@LokedSher some environments might need "" instead of ''

maybe he wants to kown how to train the yolov5 model from the randomly initialization without pre-trained model. But the default parameters in train.py is 'default=ROOT / 'yolov5s.pt'. How to set the '--weights' in train.py to get the randomly initialization.

[glenn-jocher]

@tianty for a randomly initialized yaml: --cfg yolov5s.yaml --weights ''

[github-actions]

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 🚀 resources:

• Wiki – https://github.com/ultralytics/yolov5/wiki
• Tutorials – https://docs.ultralytics.com/yolov5
• Docs – https://docs.ultralytics.com

Access additional Ultralytics ⚡ resources:

• Ultralytics HUB – https://ultralytics.com/hub
• Vision API – https://ultralytics.com/yolov5
• About Us – https://ultralytics.com/about
• Join Our Team – https://ultralytics.com/work
• Contact Us – https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐!

[wilkinsjle]

@tianty for a randomly initialized yaml: --cfg yolov5s.yaml --weights ''

This works for my Mac but not on my Windows machine ):

On Windows the error occurs: train.py: error: argument --weights: expected one argument. This is also what the original poster has said.

Using "" also hasn't worked.

[wilkinsjle]

Ok just found the fix:

If you are using Powershell, use [string]::Empty instead of "" or '' to specify an empty string.

[glenn-jocher]

@wilkinsjle great find! Using [string]::Empty in Powershell to specify an empty string instead of "" or '' should resolve the issue. Let me know if you need further assistance.

[tomhoq]

[string]::Empty causes a different error for me:

Traceback (most recent call last):
  File "E:\Tomaz\Yolov5\yolov5\train.py", line 848, in <module>
    main(opt)
  File "E:\Tomaz\Yolov5\yolov5\train.py", line 623, in main
    train(opt.hyp, opt, device, callbacks)
  File "E:\Tomaz\Yolov5\yolov5\train.py", line 196, in train
    model = Model(cfg, ch=3, nc=nc, anchors=hyp.get("anchors")).to(device)  # create
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "E:\Tomaz\Yolov5\yolov5\models\yolo.py", line 227, in __init__
    with open(cfg, encoding="ascii", errors="ignore") as f:
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: ''

while running this:
python train.py --img 640 --batch 16 --epochs 10 --data custom_data.yaml --weights [string]::Empty

[glenn-jocher]

Hi @tomhoq,

Thank you for your detailed feedback. It looks like using [string]::Empty is causing a FileNotFoundError because the script is interpreting it as a literal string rather than an empty value.

To train YOLOv5 with randomly initialized weights on a Windows machine using Powershell, you can try the following workaround:

1. Use a non-existent file path: This will force the script to initialize weights randomly.

   python train.py --img 640 --batch 16 --epochs 10 --data custom_data.yaml --weights "path/to/nonexistent/file.pt"

2. Modify the script: Alternatively, you can modify the train.py script to handle an empty string correctly. Locate the section where the weights are loaded and add a condition to check for an empty string:

   if opt.weights == '[string]::Empty' or opt.weights == '':
       opt.weights = None

3. Use a placeholder: Another approach is to use a placeholder that the script can interpret correctly:

   python train.py --img 640 --batch 16 --epochs 10 --data custom_data.yaml --weights "None"

If you continue to experience issues, please ensure you are using the latest version of YOLOv5 by pulling the latest changes from the repository. This can help ensure that any recent bug fixes or improvements are included in your environment.

Feel free to reach out if you have any further questions or need additional assistance. Happy training! 🚀