adpretko/armv8_to_riscv_qwen25coder_3p0b_full
adpretko/armv8_to_riscv_qwen25coder_3p0b_full is a fine-tuned version of the Qwen2.5-Coder-3B-Instruct model, developed by adpretko. This model specializes in code translation, specifically from ARMv8 to RISC-V architectures. It was trained on a series of dedicated armv8_to_riscv datasets, making it suitable for tasks involving cross-architecture code conversion. Its primary application is in facilitating the migration or understanding of code between these two distinct instruction set architectures.
Loading preview...
Model Overview
This model, adpretko/armv8_to_riscv_qwen25coder_3p0b_full, is a specialized fine-tuned variant of the Qwen2.5-Coder-3B-Instruct base model. Its core purpose is to translate or convert code snippets from the ARMv8 instruction set architecture to RISC-V. This makes it a highly focused tool for developers working on cross-architecture compatibility or migration projects.
Key Characteristics
- Base Model: Fine-tuned from Qwen/Qwen2.5-Coder-3B-Instruct.
- Specialization: Dedicated to ARMv8 to RISC-V code translation.
- Training Data: Trained on multiple
armv8_to_riscvdatasets (000 through 006).
Training Details
The model underwent a focused training regimen with the following hyperparameters:
- Learning Rate: 2e-05
- Optimizer: AdamW with default betas and epsilon
- Epochs: 0.5
- Batch Size: 1 (train), 8 (eval) with 8 gradient accumulation steps
- Scheduler: Cosine learning rate scheduler with 0.03 warmup ratio
Intended Use Cases
This model is particularly well-suited for:
- Code Migration: Assisting in the conversion of ARMv8 assembly or low-level code to RISC-V.
- Cross-Architecture Development: Supporting developers who need to understand or adapt code between these two architectures.
- Research: Exploring automated code translation techniques for different instruction sets.