{"id":32522065,"url":"https://github.com/droxer/mlx-cookbook","last_synced_at":"2026-07-02T23:01:38.952Z","repository":{"id":321001034,"uuid":"1076946130","full_name":"droxer/mlx-cookbook","owner":"droxer","description":"Fine-tune Qwen3-0.6B with MLX LoRA for self-cognition.","archived":false,"fork":false,"pushed_at":"2025-10-27T07:12:41.000Z","size":133,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-10-27T08:34:14.328Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/droxer.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-10-15T15:00:55.000Z","updated_at":"2025-10-27T07:12:45.000Z","dependencies_parsed_at":null,"dependency_job_id":"06828529-9000-4383-88c0-b4030c4b05a6","html_url":"https://github.com/droxer/mlx-cookbook","commit_stats":null,"previous_names":["droxer/mlx-cookbook"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/droxer/mlx-cookbook","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/droxer%2Fmlx-cookbook","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/droxer%2Fmlx-cookbook/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/droxer%2Fmlx-cookbook/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/droxer%2Fmlx-cookbook/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/droxer","download_url":"https://codeload.github.com/droxer/mlx-cookbook/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/droxer%2Fmlx-cookbook/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":35065702,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-05-26T15:22:16.424Z","status":"online","status_checked_at":"2026-07-02T02:00:06.368Z","response_time":173,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-10-28T06:51:03.412Z","updated_at":"2026-07-02T23:01:38.926Z","avatar_url":"https://github.com/droxer.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# MLX LoRA Fine-Tuning Cookbook\n\nFine-tune Qwen3-0.6B with MLX LoRA for self-cognition.\n\n## Quick Start\n\n```bash\n# Install dependencies\nuv sync\n\n# Complete workflow\nmake setup \u0026\u0026 make finetune \u0026\u0026 make chat\n```\n\n## Commands\n\n| Command | What it does |\n|---------|-------------|\n| `make help` | Show all commands |\n| `make setup` | Download dataset + model |\n| `make finetune` | Fine-tune with LoRA |\n| `make chat` | Interactive chat |\n| `make server` | Start API server |\n| `make status` | Check progress |\n| `make clean` | Clean up files |\n\n## Customization\n\n```bash\nmake finetune MODEL_NAME=\"MyBot\" MODEL_AUTHOR=\"Your Name\"\n```\n\n## Manual Commands\n\nIf you prefer not to use the Makefile:\n\n```bash\n# Setup\nuv run hf download --repo-type dataset modelscope/self-cognition --local-dir ./datasets/self-cognition\nuv run python transform.py --name \"AIR\" --author \"TronClass AIR\"\nuv run hf download --repo-type model Qwen/Qwen3-0.6B\n\n# Fine-tune\nuv run mlx_lm.lora --config ft_qwen3_lora.yaml\n\n# Test\nuv run mlx_lm.generate --model ./models/fine-tuned_Qwen3-0.6B --prompt \"What is your name?\"\nuv run mlx_lm.chat --model ./models/fine-tuned_Qwen3-0.6B\n```\n\n## Files\n\n```\n├── Makefile                 # 6 essential commands\n├── ft_qwen3_lora.yaml      # Fine-tuning config\n├── transform.py            # Dataset prep\n├── datasets/               # Raw dataset\n├── mlx_data/               # MLX format data\n├── cog_adapters/           # LoRA adapters\n└── models/                 # Fine-tuned model\n```\n\n## Config\n\n- **Model**: Qwen3-0.6B\n- **Iterations**: 500\n- **Learning rate**: 1e-4\n- **LoRA rank**: 8\n\n## Troubleshooting\n\n```bash\nmake status  # Check what's done\nmake clean   # Remove generated files\nmake setup   # Start fresh\n```\n\nThe fine-tuned model will respond with your custom name and author! 🚀","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdroxer%2Fmlx-cookbook","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fdroxer%2Fmlx-cookbook","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fdroxer%2Fmlx-cookbook/lists"}