Mistralai Devstral Small 2505 GGUF

🚀 Llamacpp imatrix Quantizations of Devstral-Small-2505 by mistralai

This project provides Llama.cpp imatrix quantizations of the Devstral-Small-2505 model. It offers various quantization options for different performance and quality requirements, and can be run in multiple environments.

🚀 Quick Start

Running the Model

• You can run the quantized models in LM Studio.
• Or run them directly with llama.cpp, or any other llama.cpp based project.

Prompt Format

<s>[SYSTEM_PROMPT]{system_prompt}[/SYSTEM_PROMPT][INST]{prompt}[/INST]

✨ Features

• Multilingual Support: Supports a wide range of languages including English, French, German, Spanish, Portuguese, Italian, Japanese, Korean, Russian, Chinese, Arabic, Farsi, Indonesian, Malay, Nepali, Polish, Romanian, Serbian, Swedish, Turkish, Ukrainian, Vietnamese, Hindi, and Bengali.
• Multiple Quantization Options: Offers a variety of quantization types with different file sizes and quality levels to meet diverse needs.
• Vision Support: Thanks to ngxson, the original mmproj from Mistral Small 3.1 works seamlessly with Devstral Small for vision use.

📦 Installation

Installing huggingface-cli

First, make sure you have hugginface-cli installed:

pip install -U "huggingface_hub[cli]"

Downloading Specific Files

Then, you can target the specific file you want:

huggingface-cli download bartowski/mistralai_Devstral-Small-2505-GGUF --include "mistralai_Devstral-Small-2505-Q4_K_M.gguf" --local-dir ./

Downloading Split Files

If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run:

huggingface-cli download bartowski/mistralai_Devstral-Small-2505-GGUF --include "mistralai_Devstral-Small-2505-Q8_0/*" --local-dir ./

You can either specify a new local-dir (mistralai_Devstral-Small-2505-Q8_0) or download them all in place (./)

📚 Documentation

Model Information

Downloadable Files

Embed/output weights

Some of these quants (Q3_K_XL, Q4_K_L etc) are the standard quantization method with the embeddings and output weights quantized to Q8_0 instead of what they would normally default to.

🔧 Technical Details

ARM/AVX Information

Previously, you would download Q4_0_4_4/4_8/8_8, and these would have their weights interleaved in memory in order to improve performance on ARM and AVX machines by loading up more data in one pass.

Now, however, there is something called "online repacking" for weights. details in this PR. If you use Q4_0 and your hardware would benefit from repacking weights, it will do it automatically on the fly.

As of llama.cpp build b4282 you will not be able to run the Q4_0_X_X files and will instead need to use Q4_0.

Additionally, if you want to get slightly better quality for , you can use IQ4_NL thanks to this PR which will also repack the weights for ARM, though only the 4_4 for now. The loading time may be slower but it will result in an overall speed incrase.

📄 License

This project is licensed under the Apache-2.0 license.

⚠️ Important Note

If you want to learn more about how we process your personal data, please read our Privacy Policy.