Gabriel-8x7b-Instruct-gguf-v0.1

Quantized version of HelloBible/Gabriel-8x7B-Instruct-v0.1

A selection of quantizations of Gabriel-8x7B-Instruct-v0.1, made using llama.cpp version: 3277 (5fac350b)

Evaluation

original model card

Gabriel-8x7B-Instruct-v0.1 is a fine tuned model using a 36k French/English Christian dataset (self made).

Base model used for training: Mixtral-8x7B-Instruct-v0.1

Evaluation