thatname

/

Ziya-LLaMA-13B-v1-ggml

like 2

modify
added_tokens.json
{
"\n🤖：": 39409,
"🧑：": 39408
}
execute shell：
python convert.py model/Ziya-LLaMA-13B
output error：
Vocab size mismatch (model has 39424，but added_tokens.json has 39410)

Can you convert to a gptq 4_0 model? use autogptq https://github.com/PanQiWei/AutoGPTQ?

modify
added_tokens.json
{
"\n🤖：": 39409,
"🧑：": 39408
}
execute shell：
python convert.py model/Ziya-LLaMA-13B
output error：
Vocab size mismatch (model has 39424，but added_tokens.json has 39410)

I've added the add_tokens.json to this repo, you can use it with convert.py to convert the original Ziya models.

thank you very march!