JiaMingLin

Thanks On Thu, 9 Dec 2021 at 12:17, nesquiq ***@***.***> wrote: > Hi, it is now available here: > https://developer.qualcomm.com/software/ai-datasets/jester > https://developer.qualcomm.com/software/ai-datasets/something-something > > Thanks a lot for noticing!!! >...

It is probably the number of flow frames is minus one of the rgb frames. You can check the split files and actual number of frames in the directories.

It maybe you packing the *.skeleton files in to npy via 120 classes rather than 60 classes. This is my case.

I think they are probably re-scaling from current layer to next layer. You can refer to the paper from google, "Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference",...

Hi, Thanks for this initial works on Brevitas conversion, and I think there is an issue. I've tested the ```QuantLinear``` layer with 8-bit weight and 8-bit ```input_quant``` and expect (8-bit,...