FastChat

FastChat copied to clipboard

lovelucymuch

The effect of lora finetune with difference target_modules

3

Hello, I using default target_modules `q_proj, v_proj` the result looks good. Will it more good if I using more target_modules trainable? such as : ``` "q_proj", "v_proj", "down_proj", "gate_proj", "up_proj",...

lucasjinreal

Add Gradio web server to Docker Compose

2

MilesQLi

path: https://github.com/lm-sys/FastChat/tree/main/fastchat/eval issue: missing field "metadata" example: { "reviewer_id":"gpt-4-0328-default", "prompt_id":1, "metadata":{ "temperature":0.2, "max_tokens":8192 }, "description":"GPT-4 for general questions", "category":"general" }

wangxingjun778

FlanT5 training and zero tensors

10

Hi, I'm training a FlanT5 network. The training completes successfully, but when I try to run a simple inference, I have a tensor of zeros, so the prediction is null....

GenVr

junruxiong

[Questions] Where can I find the delta weights automatically download?

1

Hi team, I use the command below without changing the delta-path and it was automatically downloaded. It still takes effective when I rerun the command but I just wonder where...

brucezhu512

Failed to set multiple gpus

1

When using parallel computing with multiple GPUs, if I set "--num-gpus 4," only three GPUs are actually activated. How can I solve this issue?

Halflifefa

How to save lora weights only???

1

Currently the weights saved are all models weights include optimzier, is HUGE, how to set save lora weights only? 

lucasjinreal

kuokay