metadata

license: apache-2.0
datasets:
  - TPM-28/alpaca-cleaned-fr
language:
  - fr
library_name: transformers
pipeline_tag: text-generation
tags:
  - Qlora
  - mistral

TPM-28/Mistral-7B-Inst-v0.2-Alpaca-FR

Model

Mistral-7B-Inst-v0.2-Alpaca-FR is a French language model, trained using the QLORA (Quantization Learning for Learned Optimized Representation Access) approach on the TPM-28/alpaca-cleaned-fr dataset.

The alpaca-cleaned-fr dataset is a French translation of the English dataset pinzhenchen/alpaca-cleaned-fr, performed using the LibreTranslate translator.

Usage

To query the model, the request must be formatted according to the following template:

Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
{System}

### Input:
{User}

### Response:
{Assistant}

The model will then generate an appropriate response to complete the request.

Limitations

The model may be biased by the original (translated) data used for training.