sshu2017

I know this is closed but I still encountered this issue when loading ggml-v3-13b-hermes-q5_1.bin - tried gpt4all==0.3.4 and gpt4all==0.3.0 but both gave me this error...

So, I've tested the n_threads on AWS EC2, and so far the optimal value is 48. I don't understand why but with 72CPUs and 96CPUs, the response speed slowed down...

> Hi @abhishekrai43 , sorry for the late reply. With more threads, now I can get a response in ~30 seconds. It was ~150 seconds with everything to the default...

> @sshu2017 - thanks for the report. Instead of screenshots, can you edit the OP to have text-based examples? It is extra work for maintainers to try to reproduce behavior...

> The docstring for `na_values` states: > > > Additional strings to recognize as `NA`/`NaN`. > > You are not providing strings, so this is undefined behavior. Perhaps we should...

Hi @rhshadrach , seem like this issue has been discussed and dealt with 11 years ago. Maybe we could just cherry pick the commit #3841? Not sure why the changes...

Hi @rhshadrach, I created a branch and now an ValueError would be raised when user send in a non-string value as na_values. But many tests indicate that non-string values are...

Hi @rhshadrach , sorry but just to confirm - you are suggesting that the c parser should be fixed and it should be able to take in not only ```...

@ReinierKoops , thanks. After I specified an older version of sklearn, the codes run well now. Thanks! Do you think this worth a PR?