New
[dev-ai project fine-tuning qwen3-coder code-review cpu]
Created 2026-07-10

Dev-AI

Dev-AI

Fine-tuned 27B code review model for CPU inference.

Architecture

┌──────────────┐    ┌──────────────┐    ┌──────────────┐
│  Dataset     │───▶│  QLoRA Train │───▶│  Merge +     │
│  (Code, SO,  │    │  (CPU, 4-bit)│    │  llama.cpp   │
│  Docs, PRs)  │    │  ~14 days    │    │  Server      │
└──────────────┘    └──────────────┘    └──────────────┘
                                              │
                                              ▼
                                        ┌──────────────┐
                                        │ Code Review  │
                                        │ API (port    │
                                        │ 8081)        │
                                        └──────────────┘

Workflow

  1. Vincent writes code → 2. POST to fine-tuned model → 3. Model reviews/corrects → 4. Vincent iterates

Hardware

  • Training: Ryzen 9 9900X CPU, 60GB RAM (no GPU)
  • Inference: llama.cpp server, ~17GB RAM for Q4 27B model
  • GPU (4090): Reserved for Vincent's own inference, not touched

Project Structure

dev-ai/
├── data/       # Training data
├── model/      # Model weights
├── output/     # Inference output
├── scripts/    # Training and inference scripts
├── training/   # Training pipeline
└── PROGRESS.md # Current status

Status

Development — QLoRA training pipeline on CPU.