Kishore Chandra Sahoo
Kishore Chandra Sahoo
I converted parseq `torch` model to `onnx` to `trt` with `tensorrt version 8.4.1.5` How do I load that `trt` model for inferencing?
Looks fine. We can review and close minor PRs.
Definitely. I was just telling to reduce the PR count and to narrow down the focus on real issues and PRs. In case it's no quality, you can reject it...
@paritosh-96 Were you able to resolve the GSSAPI issue?
@freakboy3742 Thanks for your reply and elaborate explanation. I understood it and definitely it requires significant amount of time & effort for this development. Thank You.
Seems your model's output dimension is not as expected. You need to check how you converted Torch model to Onnx. You can use ultralytics yolo to export to onnx.
command : `yolo export model= format=onnx batch=16|32 dynamic=True device=0` Use `batch`, `dynamic` as per your requirement.
`visibility` and `presence` are not available in latest code. @PaulTR :Any plan to include in new future? Those parameters are helpful to identify if full body is on screen or...
Few other things to ponder: - Android reads images as bitmap with ARGB format. Whereas Python OpenCV reads with RGB. - While converting to MPImage in Android, We cannot pass...
I converted to RGB in Opencv and generated embedding. Now I get below results: ## Python: ``` [38, 236, 0, 208, 208, 208, 52, 243, 47, 212, 127, 230, 214,...