Description
[THIS IS HIGHLY EXPERIMENTAL]
This repo contains fp16 files of Llama-2-13b-Mistral, it's like a bootleg of Mistral, with updated data from 2022/2023, made by applying a LoRA at weight 1.0 of extracted Mistral data and made to be use for merge/finetune. I found it incredibly hard to work with Mistral at the moment, so I got this temporary idea.
The Model being thinkered with a LoRA, it still can output bad informations, but it contain Mistral one as well, giving me hope that we can make better model with those base.
Model and lora used
Prompt template: Alpaca
Below is an instruction that describes a task. Write a response that appropriately completes the request.
### Instruction:
{prompt}
### Response:
Screenshots
Honorable RP mention:
(He is actually 45 yo)
If you want to support me, you can here.