Gokul NC

Unfortunately, we have not calculated that. If you get a chance to calculate those stats, kindly do let us know too so that we can include it in our README....

Yes, all the videos are under CC license, which allow reuse (with attribution). We had sampled a good number of lecture videos (with subtitles) from the playlist to verify if...

Hi @AswiniKumar1995 Thank you for pointing out the issue, it has been fixed now. Please refresh the page and try again.

Hi @AswiniKumar1995 , apologies for late reply. No, this is not supported as of now. We only support ISLR (isolated sign language recognition) as of now, which means we provide...

Hi @AswiniKumar1995 , apologies for late reply. Yes, you can get token-level output from BERT, but you will have to make changes in the source-code. This is not available by...

Hello @leekezar ! This is due to the CosineAnnealing LR Scheduler: https://github.com/AI4Bharat/OpenHands/blob/7fb2c0a0f0581026ac1a9a522bac838526e62f16/examples/configs/wlasl/decoupled_gcn.yaml#L109-L113 The oscillation usually becomes negligible after a good no. of epochs. But you can also remove this from...

We should do this for videos as well as poses.

Hi Amit, as of now (for a few months), we do not have any plans. But yes, we will be expanding into CSLR and do have future plans. :)

Hi @AmitMY Thanks for sharing this. Interesting, we did not know about this `.pose` format. Before starting our project, we looked if any there was any standard format for storing...

Thanks Mathias & Amit. Makes sense. Regarding this: > normalization or similar processing should never change the underlying data > if you change the underlying data, it is not the...