mstJuly

I run benchmark on A100 GPU with meta-llama/Meta-Llama-3-8B, cost 4.15 seconds loading model. When I ran `fio fio-config.ini`, the results were as follows: ```bash Sequential Read Test: (g=0): rw=read, bs=(R)...

Other hardware adaptation

2

If I wish to adapt sllm for the Ascend NPU, what modifications should I undertake?