Mellum-4b-dpo-python is the third stage of our pipeline (after pretraining and SFT), trained with direct preference optimization on code-quality preferences to produce more readable, useful code.
Pre-trained on over 4 trillion tokens with a context window of 8192 tokens across multiple programming languages, and then fine-tuned, Mellum-4b-dpo-python is tailored specifically for code completion in Python.
The model follows a LLaMA-style architecture with 4 billion parameters, making it efficient for both cloud inference (e.g., via vLLM) and local deployment (e.g., using llama.cpp or Ollama).
Mellum was trained using Automatic Mixed Precision (AMP) with bf16 precision.
The uploaded version on Hugging Face retains the bf16 format for public use.
Designed for integration into professional developer tooling (e.g., intelligent code suggestions in IDEs), AI-powered coding assistants, and research on code understanding and generation, Mellum is also well-suited for educational applications and fine-tuning experiments.
Here are examples of how to run and sample from the model.
1@misc{Mellum-4b-dpo-python,
2 title = {Mellum-4b-dpo-python},
3 author = {Pavlichenko, Nikita and Nazarov, Iurii and Dolgov, Ivan and Garanina, Ekaterina and Lasocki, Karol and Reshetnikova, Julia and Boitsov, Sergei and Bondyrev, Ivan and Karaeva, Dariia and Sheptyakov, Maksim and Ustalov, Dmitry and Mukhin, Artem and Proshev, Semyon and Abramov, Nikita and Kolomyttseva, Olga and Lysaniuk, Kseniia and Zavidnyi, Ilia and Semenkin, Anton and Tankov, Vladislav and Sazanovich, Uladzislau},
4 year = {2025},
5}
For questions, collaborations and requests reach us out via
mellum@jetbrains.com
