vignesh

Results 8 comments of vignesh

'Upsample' object has no attribute 'recompute_scale_factor' it's possible newer versions of PyTorch have additional argument requirements on certain modules like the Upsample one you show is causing issues. The current...

I think you should consider a minimum for the clipper loss function

Can anybody help me with this

Ohh I see @chrisociepa Thank you

in LlaMA2 source code, they obtain 'X_V' with the origin 'x', instead of 'swish' its the same but in here the 'swish' function with beta =1 is nothing but SiLU...

> @vigneshyaadav27 Feel free to send a pull request with the fixes. Sure

This is due to the HNSW algorithm used for retrieval, which only supports a maximum of 2000 dimensions.

Oh boy !! I thought I was the one missing something I penned down the entire forward pass derivation and I did not understand why the Inverse over O is...