Alexey

Try to set `paste_in: 0.00` in `hyp.scratch.yaml` file Or repace `compute_loss_ota` to `compute_loss` in this line https://github.com/WongKinYiu/yolov7/blob/main/train.py#L362 Does it help?

> Besides, compute_loss required 3 parameters, so we have to remove "imgs" params in the input, does it affect the training? This is correct. I added a fix, now you...

Also do you have any test, does it work faster and still correctly?

Paperswithcode provides a comparison to other networks with references to values in papers: https://paperswithcode.com/sota/real-time-object-detection-on-coco?dimension=FPS%20%28V100%2C%20b%3D1%29


This is just a warning, just ignore it: https://colab.research.google.com/github/WongKinYiu/yolov7/blob/main/tools/YOLOv7trt.ipynb

@triple-Mu Is there an issue with batch-processing for ONNX models?

Use `--conf-thres 0.001` during export https://github.com/WongKinYiu/yolov7/blob/main/test.py#L25 ``` python export.py --weights yolov7-tiny.pt --grid --end2end --simplify \ --topk-all 100 --iou-thres 0.65 --conf-thres 0.001 --img-size 640 640 --max-wh 640 ```

When tested in an identical environment on a nVidia T4 GPU: **YOLOv7** (51.2% AP, 12.7ms) is `1.5x` times faster and `+6.3%` AP more accurate than **YOLOv5s6** (44.9% AP, 18.7ms) https://colab.research.google.com/gist/AlexeyAB/56912451a33981d977ff9ea61025ae40/yolov7trtlinaom.ipynb#scrollTo=-tMYe8f27US9...

It is strange that you get 56 FPS (18ms) for `yolov7.pt` on Titan RTX (130 TFlops-TC), while there is higher **79 FPS (12.6ms)** on GPU T4 (65 TFLops-TC) while Titan...