Error trying to run the code.

[arthurwolf]

Following the exact instructions, here is what I get:

╭─arthur at aquarelle in ~/dev/ai/yolov5_anime on master✘✘✘ 23-12-02 - 2:56:53
╰─⠠⠵ python detect.py --weights weights/yolov5x_anime.pt --source ../manga/data/source/alita/Battle\ Angel\ Alita\ v02/Volume\ 2\ -\ Tears\ Of\ An\ Angel  --output /ram/test 
Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.4, device='', img_size=640, iou_thres=0.5, output='/ram/test', save_txt=False, source='../manga/data/source/alita/Battle Angel Alita v02/Volume 2 - Tears Of An Angel', update=False, view_img=False, weights=['weights/yolov5x_anime.pt'])
/home/arthur/.pyenv/versions/3.7.17/lib/python3.7/site-packages/torch/cuda/__init__.py:88: UserWarning: CUDA initialization: Unexpected error from cudaGetDeviceCount(). Did you run some cuda functions before calling NumCudaDevices() that might have already set an error? Error 804: forward compatibility was attempted on non supported HW (Triggered internally at ../c10/cuda/CUDAFunctions.cpp:109.)
  return torch._C._cuda_getDeviceCount() > 0
Using CPU

Fusing layers... Model Summary: 284 layers, 8.83906e+07 parameters, 8.45317e+07 gradients
image 1/612 /home/arthur/dev/ai/manga/data/source/alita/Battle Angel Alita v02/Volume 2 - Tears Of An Angel/BA-02-007-rotated.jpg: Traceback (most recent call last):
  File "detect.py", line 171, in <module>
    detect()
  File "detect.py", line 72, in detect
    pred = model(img, augment=opt.augment)[0]
  File "/home/arthur/.pyenv/versions/3.7.17/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/arthur/dev/ai/yolov5_anime/models/yolo.py", line 109, in forward
    return self.forward_once(x, profile)  # single-scale inference, train
  File "/home/arthur/dev/ai/yolov5_anime/models/yolo.py", line 129, in forward_once
    x = m(x)  # run
  File "/home/arthur/.pyenv/versions/3.7.17/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1194, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/arthur/.pyenv/versions/3.7.17/lib/python3.7/site-packages/torch/nn/modules/upsampling.py", line 157, in forward
    recompute_scale_factor=self.recompute_scale_factor)
  File "/home/arthur/.pyenv/versions/3.7.17/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1270, in __getattr__
    type(self).__name__, name))
AttributeError: 'Upsample' object has no attribute 'recompute_scale_factor'
╭─arthur at aquarelle in ~/dev/ai/yolov5_anime on master✘✘✘ 23-12-02 - 2:57:35
╰─⠠⠵     

Am I doing something wrong? What should I do to fix this?

Thanks!

[zymk9]

Hi Arthur. This seems like a compatability issue with newer versions of PyTorch. You may check ultralytics/yolov5#6948 for the versions to use. Switching to PyTorch 1.9 or 1.10 might work.

Hope it helps!

[arthurwolf]

Thanks for the help, moving to python3.7 (using pyenv) and torch 1.9 worked.

Deleted the previous comment as I figured out what was going on: the model/weights file was 0 bytes. This actually comes from the download failing, which seems to come from the google drive link being broken: clicking "download" gives me either a download error or an empty file.

Would you mind checking on the file and/or uploading it somewhere else?

(also, what does the output in the --save-txt files mean/how to interpret it exactly? edit: figured some of it out from the source code, it's xywh, I'm going to presume using 0-1 positions. I'll try to see if interpreting it that way works)

Thanks!