File size: 465 Bytes
0d4fc31
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
GPTQ quantized version of Mistral-7B-v0.2-hf model.

---

~~Mistral 7b v0.2 with attention_dropout=0.6, for training purposes~~

Conversion process:

1. Download original weights from https://models.mistralcdn.com/mistral-7b-v0-2/mistral-7B-v0.2.tar
2. Convert with https://github.com/huggingface/transformers/blob/main/src/transformers/models/mistral/convert_mistral_weights_to_hf.py

3. You may need to copy the tokenizer.model from Mistral-7B-Instruct-v0.2 repo.