Yudi Zhang

Have you solved this problem yet? I have encountered the same problem.

Even when I only pass the q k v arguments and omit the others, the error still occurs ```Python torch.library.define( "mylib::custom_func_flashinfer", "(Tensor q, Tensor k, Tensor v) -> Tensor", )...

Thank you for your response. I have tried using `torch.compile(..., fullgraph=True, mode="max-autotune-no-cudagraphs")`, and the code runs without errors. However, when I aim to implement CUDA graphs for flashinfer and other...

Have you encountered the problem of `segmentation fault (core dumped)` when using Llama-3-8B and running `python3 get_datastore_chat.py --model-path Meta-Llama-3-8B-Instruct`?