Glenn Jocher
Glenn Jocher
> @soumik12345 do you know how we can resolve the wandb>=0.12.2 issue below? > > https://github.com/ultralytics/yolov5/blob/2794483e091d50416289614a1a35f158fd25bee2/utils/loggers/__init__.py#L25-L29 > > Every training with W&B now displays a warning to the users now...
@soumik12345 thanks for looking into this! The version constraint is not enforced, i.e. there's no hard constraint, it's just that this warning message appears when `wandb>0.12.10` is installed.
AssertionError: Invalid CUDA '--device 0' requested, use '--device cpu' or pass valid CUDA device(s)
@shubhambagwari it appears you may have environment problems. Please ensure you meet all dependency requirements if you are attempting to run YOLOv5 locally. If in doubt, create a new virtual...
@ZiDuNet your code is out of date. To update: - **[Git](https://github.com/ultralytics/yolov5)** โ `git pull` from within your `yolov5/` directory or `git clone https://github.com/ultralytics/yolov5` again - **[PyTorch Hub](https://pytorch.org/hub/ultralytics_yolov5/)** โ Force-reload `model...
@youngjae-avikus ๐ Hello! Thanks for asking about **training speed issues**. YOLOv5 ๐ can be trained on CPU (slowest), [single-GPU](https://github.com/ultralytics/yolov5/wiki/Train-Custom-Data), or [multi-GPU](https://github.com/ultralytics/yolov5/issues/475) (fastest). If you would like to increase your training...
@Chenzkun this could be caused by simply running out of CUDA memory, I'd try reducing the batch size to 6 or 4. If the error still occurs you might try...
@Chenzkun thanks for the bug report! It looks like AutoBatch pushed the ceiling too close in your case, and `--batch-size 2` rather than `--batch-size 3` would be the best option...
@Chenzkun ๐ Hello! Thanks for asking about **CUDA memory issues**. YOLOv5 ๐ can be trained on CPU, [single-GPU](https://docs.ultralytics.com/yolov5/tutorials/train_custom_data), or [multi-GPU](https://docs.ultralytics.com/yolov5/tutorials/multi_gpu_training). When training on GPU it is important to keep your...
@tanzerlana ๐ Hello! Thanks for asking about benchmarks. YOLOv5 ๐ inference is officially supported in 11 formats: ๐ก ProTip: Export to ONNX or OpenVINO for up to 3x CPU speedup....
@sarpx I don't understand your question. You can view CSV or TensorBoard logging results instead of W&B. See Colab notebook for examples in Visualize section. https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb?hl=en