Title Yi-6B-200K can't be used

[EasonFeng5870]

Reminder

• I have searched the Github Discussion and issues and have not found anything similar to this.

Environment

- OS: 
- Python:
- PyTorch:
- CUDA:

we are setting up a Yi-6B-200K model, it can't answer any questions, even common 1+1=

This is a bad model. 

Am I wrong? btw, this model doesn't have a useful chat template in tokenizer_config.json, so we try a lot of time to find the use of this model:

### user: 1+1= \n ### assistant:

Current Behavior

which is cant used for any questions.

Expected Behavior

No response

Steps to Reproduce

just query

Anything Else?

none

[Yimi81]

The 6B-200K we provided is a base model, not a chat model, so there is no chat template available. It cannot be used for chat, it only works after fine-tuning
In addition, if there is a problem of continuous repetition when using 9B, you can refer to the text_generation.py in the demo folder to give some repetition penalties and other parameters during generation.