HA-Siala/Mamba-full-v0.2

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:4kTool Calling:SupportedPublished:Jun 11, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

HA-Siala/Mamba-full-v0.2 is a 7 billion parameter Mistral-based language model developed by HA-Siala. This model is specifically fine-tuned to identify flaws within Mamba code and generate refactored versions. Its primary application is in Mamba code analysis and optimization.

Loading preview...

Overview

HA-Siala/Mamba-full-v0.2 is a 7 billion parameter language model built upon the Mistral architecture. Developed by HA-Siala, this model has undergone specialized fine-tuning to address a very specific technical challenge: the detection of flaws within Mamba codebases.

Key Capabilities

  • Mamba Code Flaw Detection: The model is trained to identify errors, inefficiencies, or other issues present in Mamba programming code.
  • Code Refactoring: Beyond identification, it can produce refactored versions of the flawed Mamba code, aiming to improve its quality, efficiency, or correctness.

Good For

  • Mamba Code Analysis: Developers working with Mamba code can leverage this model to automatically scan for potential problems.
  • Automated Code Improvement: It can assist in the refactoring process, suggesting or generating improved code segments for Mamba projects.

For a deeper dive into the dataset and source code used for this model's development, refer to the official GitHub repository.