Hasanli Orkhan

any updates?

I have written Python code for `grid_sample` with zero-padding and can pass align_corners argument as well. ``` https://github.com/OrkhanHI/pytorch_grid_sample_python/blob/main/pytorch_grid_sample_python.md ```

It is your webcam issue, you can try with index 1 (-src 1) as well. ``` python run_webcam.py -src 0 --landmark-model shape_predictor_68_face_landmarks.dat --tf-model frozen_model.pb ```

Hi, Have you managed to get weights file for smoking?

@CVUsers can you please provide baidu weights file link for smoking detection?

Because pytorch grid sample is written in C++ and I think also it was using multi-threading. I have changed their C++ code and build only zero padding for grid sample....

any updates?

I am getting the same error. I have installed `MediaSDK`, then cloned `cva_sample` and built it. Should we reinstall again? Ubuntu 20.04 ``` ./bin/video_e2e_sample -par par_file/rtsp/n16_face_detection_1080p.par ``` ``` Found 1...

It turned out my GPU was not Intel Integrated GPU. I enabled it from BIOS, then under `Settings` - Graphics became Mesa Intel® Graphics (RKL GT1) instead of NVIDIA.

In the paper, you mentioned using in-batch negatives in BLIP2 instead of the momentum queue in BLIP. It would be great if you share modified code for retrieval using BLIP2....