LongChat
LongChat copied to clipboard
DachengLi1
Official repository for LongChat and LongEval
Hi, I try to run inference with `lmsys/longchat-7b-v1.5-32k` from huggingface with following chat template. ``` [INST] \nYou are a helpful, respectful and honest assistant. Always answer as helpfully as possible,...
Hi, "We fine-tune the 7B and 13B models with 80k and 18k conversations, respectively." Could you provide more details about the training data? How the 80k data are prepared? Are...
Anthropic changed their python sdk - making this code line outdated. https://github.com/DachengLi1/LongChat/blob/a824bda25c0082e60973c35c79b0f35d69c6be2d/longeval/utils.py#L307 --- Would love to know if this might help - https://github.com/BerriAI/litellm ~Simple I/O library, that standardizes all the...
Hello, I test the inference speed of longchat-13b-16k. On the longeval topic task, input 9600 token length, output 12 tokens, it takes 23s. Then on the LongBench, input 7367 token...
By using xformers to train llama2, the loss are explosion, do u know why? On V100 only
ValueError: Asking to pad but the tokenizer does not have a padding token. Please select a token to use as `pad_token` `(tokenizer.pad_token = tokenizer.eos_token e.g.)` or add a new pad...
Hello everyone, I encountered a problem: When calling the API, I have adjusted the token limit for the output content to 5000, but the final generated content is still around...
Thanks for your awesome work so that the community can train LLM on very long context! However, I find that in the `preprocess` function, line https://github.com/DachengLi1/LongChat/blob/a824bda25c0082e60973c35c79b0f35d69c6be2d/longchat/train/fine_tune/train.py#L125 and line: https://github.com/DachengLi1/LongChat/blob/a824bda25c0082e60973c35c79b0f35d69c6be2d/longchat/train/fine_tune/train.py#L137 will...
