adpretko/x86_to_armv8mac_qwen25coder_0p5b_full

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:0.5BQuant:BF16Ctx Length:32kPublished:Mar 26, 2026License:otherArchitecture:Transformer Warm

The adpretko/x86_to_armv8mac_qwen25coder_0p5b_full model is a 0.5 billion parameter language model, fine-tuned from Qwen/Qwen2.5-Coder-0.5B-Instruct. It specializes in code generation and transformation tasks, specifically focusing on converting x86 code to armv8mac architecture. This model is optimized for developers working on cross-architecture code migration, leveraging its base in the Qwen2.5-Coder family for robust code understanding and generation capabilities.

Loading preview...

Model Overview

This model, adpretko/x86_to_armv8mac_qwen25coder_0p5b_full, is a specialized fine-tuned version of the Qwen/Qwen2.5-Coder-0.5B-Instruct base model. With 0.5 billion parameters and a context length of 32768 tokens, it is designed for specific code transformation tasks.

Key Capabilities

  • Code Transformation: Primarily fine-tuned for converting code from x86 architecture to armv8mac. This makes it highly relevant for developers migrating applications or understanding cross-architecture compatibility.
  • Instruction-Following: Inherits instruction-following capabilities from its Qwen2.5-Coder-Instruct base, enabling it to respond to specific coding prompts.
  • Efficiency: As a 0.5B parameter model, it offers a balance between performance and computational efficiency, suitable for targeted code generation tasks.

Training Details

The model was fine-tuned on a series of custom datasets: x86_to_armv8mac_000 through x86_to_armv8mac_006. Key training hyperparameters included a learning rate of 2e-05, a total batch size of 8 (with gradient accumulation), and a cosine learning rate scheduler over 0.5 epochs. This focused training regimen aims to imbue the model with precise knowledge for its specialized code conversion domain.

Good For

  • Developers needing assistance with x86 to armv8mac code migration.
  • Tasks requiring small, efficient code generation models with specific architectural knowledge.
  • Research into cross-architecture code translation and optimization.