Views
No views yet
| Metric | Value |
|---|---|
| Base Model | Qwen3-0.6B |
| Fine-tuning | LoRA (r=16, alpha=16) |
| Training Data | 16,000 synthetic notifications |
| Folder Accuracy | 94.0% |
| Priority Accuracy | 83.0% |
| Quantization | Q8_0 |
| Model Size | ~650 MB |
{"folder": "Work", "priority": "high"}./llama-cli -m qwen3-0.6b-notifai-Q8_0.gguf -p "<prompt>"1@misc{notifai2026,
2 title={NotifAI: On-Device Notification Classification},
3 author={封一},
4 year={2026},
5 url={https://github.com/charfeng1/notifai}
6}