Wenxuan Tan

It's using exactly the evaluation set.

It's mostly a network issue. I tried inside docker, and successfully downloaded the metric outside it.

Not currently, but some collaboration with AMD is underway

> To resolve this issue, you can try converting the Triton data type (tl.dtype) to a native Python data type before passing it to the kernel. Here's how you can...

@Forence1999 could you share how you reproduced it? I only got 32.1 with the original hyperparameters. Thanks! ``` python qlora.py \ --model_name_or_path huggyllama/llama-7b \ --use_auth \ --output_dir /fly/results/qlora \ --logging_steps...

This is typically out of memory error.