Mike

``` I solved the problem. You can try under this file File "/home/ec2-user/.local/lib/python3.9/site-packages/torch/_tensor.py", line 732, in array return self.numpy() modify self.numpy() to self.cpu().numpy ``` Your fix is within torch @Zhengzhiyang0000?...

I achieve the following inference times on my webcam with a modest Quadro P2000. Which is way below a 1660ti in terms of specs @Jimmeimetis. **Yolov5s.pt + mobilenetv2_x1_0_msmt17.pt** ```bash 0:...

Notice that the more detection you have the longer time it will take for StrongSORT to finish the association process. Btw, I don't think 1660ti supports half precision inference...

`python track.py --strong-sort-weights /path/to/your/osnet_x0_25_model.pt`

`pip list` gives me : `clearml 1.6.4` I don't know if it is persistent, have only tried it once. I will report back when I try it again

I run the command again. No issue this time: ```bash Uploading dataset changes (90 files compressed to 95.01 MiB) to gs://icm-data-lake/clearml-test Uploading dataset changes (89 files compressed to 94.66 MiB)...

I ended up doing the following: ```python try: import clearml from clearml import Dataset, Task, Model, OutputModel assert hasattr(clearml, '__version__') # verify package import not local dir except (ImportError, AssertionError):...

Current KF: > HOTA: 52.979 > MOTA: 59.983 > IDF1: 67.239 Proposed KF: > HOTA: 52.642 (-0.337) > MOTA: 59.826 (-0.157) > IDF1: 66.289 (-0.950) IDF1 took quite a hit

One more: `AttributeError: 'KalmanFilter' object has no attribute '_initial_error_scaling'`

Will have a look at it this weekend :smile: