aaronmaiww

Hi again, I'd love to try out the recurrent memory transformer you use in your paper - do you have any plans of making the code public/upload it to huggingface?...

Problem with DeepSpeed for BigBird Sparse

5

Hi, thank you for this excellent repository! I'm trying to set-up `gena-lm-bigbird-base-sparse-t2t`, but I can't install the requisite libraries in the way suggested on huggingface: ``` pip install triton==1.0.0 DS_BUILD_SPARSE_ATTN=1...

AssertionError: libcuda.so cannot found!

1

Hey, thank you for the repo! I'm trying to fine-tune this model but I get the following error during a forward pass: ``` --------------------------------------------------------------------------- AssertionError Traceback (most recent call last)...