Weiming Hu

icon indicating a website link huweim.github.io

icon company Shanghai Jiao Tong University icon location Shanghai, China icon location Learning Computer Architecture; Ph.D. Student @ SJTU CS

Results 5 comments of Weiming Hu

stack during import tensorrt_llm

Same issue, how do you solve it?

Fail to reproduce the perplexity of Llama-2 7B on wikitext

For me, I use the codebase of GPTQ or AWQ to run the wikitext evaluation.

Fail to reproduce the perplexity of Llama-2 7B on wikitext

> I think the baseline in GPTQ or AWQ for wikitext2 is correct. However, lm-eval contains most of the datasets so I want to use it for wikitext2. But the...

Fail to reproduce the perplexity of Llama-2 7B on wikitext

> > > I think the baseline in GPTQ or AWQ for wikitext2 is correct. However, lm-eval contains most of the datasets so I want to use it for wikitext2....

Index in triton

> [#5262](https://github.com/triton-lang/triton/pull/5262) adds support for this as > > output = tl.gather(x, idx, axis=0) Is it a register-level operation?