qlora instruct

This repo contains a qlora adapter for Mistral-7B-Instruct-v0.1, fine-tuned on a private instructions dataset, exclusively in Polish.

The fine-tuning took 1 hour on a single RTX 4090 with the following hyperparameters:

During the training of the adapter, the instruction format was maintained. Each prompt should be surrounded by [INST] and [\INST] tokens to leverage instruction fine-tuning.

This adapter allows the model to speak Polish more accurately than vanilla Mistral-7B-Instruct-v0.1.

<p align="center"> <img src="https://huggingface.co/Azurro/Mistral-7B-Instruct-v0.1-qlora-polish/raw/main/mistral-7b-instruct-qlora-pl.jpg"> </p>