Hao Zhang

icon indicating a website link https://cseweb.ucsd.edu/~haozhang/

icon location Asst. prof. at UCSD working on LLMs.

Results 174 comments of Hao Zhang

[WIP] Fixe FSDP saving error

Closing this PR, I am going to start a new PR with the fix.

when will add the stream feature in api？

what do you mean by stream feature? aren't our current CLI and web interface both streaming?

Not all parameters were saved completely.

I am not sure we can directly use the method `trainer.deepspeed`? since our train.py script does not use any deepspeed functionality

Fine Tuning on Low Memory GPU

#543

AssertionError: No inf checks were recorded for this optimizer.

@samarthsarin Please provide more information (full stack trace); it is hard to help by only seeing an assertion error.

execute fastchat.serve.cli error

you are running out of memory.

safe_save_model_for_hf_trainer function save mode is very small

@lw3259111 Just try to get a clarification -- are you training your own llama 33B using deepspeed?

Does the Vicuna API currently support multi-round conversations?

Do you mean the openAI-like API we added recently or something else?

Does the Vicuna API currently support multi-round conversations?

Btw, we do not hold a public API service.

Does the Vicuna API currently support multi-round conversations?

@wujunjiesd I think so, see this PR: https://github.com/lm-sys/FastChat/pull/663