Files

NanoCode012 09959fac70 Feat: add Magistral Small 2509 and native mistral3 tokenizer support (#3165 )

* feat: update mistral common

* feat: add mistral3processor

* fix: loading

* fix: cast pixel_values to fp32

* fix: image tensor conversion

* feat: add FA2 support for pixtral based models

* fix: update mistral small 3.1 to use native tokenizer

* fix: install tips

* fix: improve info on sample dataset files

* chore: move mistral configs into subfolders

* fix: remove unneeded patch

* fix: indent

* feat: add integration tests

* chore: move

* feat: add magistral 2509 docs and example

* fix: convert tensor to bool

* feat: expand tests

* chore: move tests

2025-09-18 15:42:20 +07:00

2.9 KiB

Raw Blame History

Finetune Voxtral with Axolotl

Voxtral is a 3B/24B parameter opensource model from MistralAI found on HuggingFace. This guide shows how to fine-tune it with Axolotl.

Thanks to the team at MistralAI for giving us early access to prepare for this release.

Getting started

Install Axolotl following the installation guide.

Here is an example of how to install from pip:

# Ensure you have Pytorch installed (Pytorch 2.6.0 min)
pip3 install packaging==23.2 setuptools==75.8.0 wheel ninja
pip3 install --no-build-isolation 'axolotl[flash-attn]>=0.12.0'

Please install the below.

# audio
pip3 install librosa==0.11.0
pip3 install 'mistral_common[audio]==1.8.3'

# Install CCE https://docs.axolotl.ai/docs/custom_integrations.html#cut-cross-entropy
python scripts/cutcrossentropy_install.py | sh

Download sample dataset files

# for text + audio only
wget https://huggingface.co/datasets/Nanobit/text-audio-2k-test/resolve/main/En-us-African_elephant.oga

Run the finetuning example:

# text only
axolotl train examples/voxtral/voxtral-mini-qlora.yml

# text + audio
axolotl train examples/voxtral/voxtral-mini-audio-qlora.yml

These configs use about 4.8 GB VRAM.

Let us know how it goes. Happy finetuning! 🚀

TIPS

For inference, the official MistralAI team recommends temperature: 0.2 and top_p: 0.95 for audio understanding and temperature: 0.0 for transcription.
You can run a full finetuning by removing the adapter: qlora and load_in_4bit: true from the config.
Read more on how to load your own dataset at docs.
The text dataset format follows the OpenAI Messages format as seen here.
The multimodal dataset format follows the OpenAI multi-content Messages format as seen here.

Optimization Guides

Limitations

We only support the mistral-common tokenizer for Supervised Fine-tuning at the moment and for type: chat_template only.

In addition, we do not support overriding tokens yet.

Future Work

Add parity to Preference Tuning, RL, etc.
Add parity to other tokenizer configs like overriding tokens.

2.9 KiB Raw Blame History