Model Overview

This model, riscv_to_armv8mac_qwen25coder_1p5b_full, is a specialized fine-tuned version of the Qwen/Qwen2.5-Coder-1.5B-Instruct base model. Developed by adpretko, it focuses on the niche task of translating RISC-V assembly code to ARMv8-A assembly, specifically for macOS environments. It leverages a 1.5 billion parameter architecture and supports a substantial context length of 32,768 tokens, making it suitable for processing moderately sized code blocks.

Key Capabilities

• Cross-Architecture Code Translation: Primarily designed for converting RISC-V assembly to ARMv8-A (macOS) assembly.
• Code Generation: Inherits code generation capabilities from its Qwen2.5-Coder-Instruct base.
• Specialized Fine-tuning: Trained on a series of riscv_to_armv8mac datasets (000-006) to enhance its proficiency in this specific translation task.

Training Details

The model was trained with a learning rate of 2e-05, a batch size of 1 (with 8 gradient accumulation steps), and for 0.5 epochs. It utilized an AdamW optimizer and a cosine learning rate scheduler with a 0.03 warmup ratio. The training was conducted using Transformers 4.46.1 and PyTorch 2.5.1+cu121.