Glenn Jocher

Custom model loading has been simplified now with PyTorch Hub in PR https://github.com/ultralytics/yolov5/pull/1677 🚀 ## Custom Models This example loads a custom 20-class [VOC](https://github.com/ultralytics/yolov5/blob/master/data/voc.yaml)-trained YOLOv5s model `'yolov5s_voc_best.pt'` with PyTorch Hub....

@EconML `results` is a Detections() instance, defined in models/common.py: https://github.com/ultralytics/yolov5/blob/c0ffcdf998aa0e00f1f39e79ee8124b74b53d23f/models/common.py#L190-L191

@Lifeng1129 I've heard this request before, so I've created and merged a new PR #1897 to add this capability. To receive this update you'll need to `force_reload` your pytorch hub...

@jmanuelnavarro hub does not need network connectivity. Load the model with your first method as that's working.

@dan0nchik that's a good question. There's no capability for this currently. It would be nice to have something like `results.save('path/to/dir')` right?

@dan0nchik I've implemented your feature idea in PR #2179. You can now pass a directory to save results to: ```python results.save() # save to current directory results.save('path/to/dir') # save to...

@zqstdy hello! Thanks for reaching out and detailing the issue you encountered with the single-channel input setup. From what you have described and the screenshots, it seems like there might...

Hello! Yes, you can train a YOLOv8 model with single-channel grayscale images. Make sure your dataset images are properly converted to grayscale and that your model configuration is set to...

@zqstdy hello! Thank you for the additional information and clarification. Indeed, training YOLOv8 with single-channel images involves some specific setup. The `nc: 1` property you've referenced is for adjusting the...

@zqstdy hello! Thanks for reaching out and sharing the link to your blog post. It looks like there might be a small discrepancy in your configuration or setup compared to...