cognitivecomputations

/

dolphin-2.5-mixtral-8x7b

like 1.23k

Follow

Cognitive Computations 1.5k

I use Axolotl
The config is here
https://huggingface.co/ehartford/dolphin-2.5-mixtral-8x7b/blob/main/configs/dolphin-mixtral-8x7b.yml

@ehartford
What script do you use to merge peft adapter with base model? I am having issues with special tokens with the one that I am using.

Edit: nvm, i missed axolotl's documentation about it.

I use Axolotl's merge feature for that