Overview
This model, adpretko/armv8mac_to_riscv_qwen25coder_1p5b_full, is a specialized 1.5 billion parameter language model fine-tuned from the Qwen2.5-Coder-1.5B-Instruct base. Its primary function is to facilitate the translation of ARMv8 macOS assembly code into RISC-V assembly. The model was trained on a series of armv8mac_to_riscv datasets, indicating a focused application in cross-architecture code conversion.
Key Characteristics
- Base Model: Qwen/Qwen2.5-Coder-1.5B-Instruct.
- Parameter Count: 1.5 billion parameters.
- Context Length: Supports a substantial context window of 32768 tokens, beneficial for processing larger code blocks.
- Specialization: Fine-tuned specifically for ARMv8 macOS to RISC-V assembly translation.
Training Details
The model underwent training with a learning rate of 2e-05, a batch size of 1 (with 8 gradient accumulation steps for an effective total batch size of 8), and 0.5 epochs. The optimizer used was adamw_torch with standard betas and epsilon, employing a cosine learning rate scheduler with a 0.03 warmup ratio. This configuration suggests an emphasis on precise fine-tuning for its niche task.
Intended Use Cases
This model is particularly suited for developers and researchers working on:
- Automated code translation between ARMv8 macOS and RISC-V architectures.
- Assisting in porting software or understanding architectural differences at the assembly level.
- Specialized code generation tasks where cross-architecture compatibility is critical.