Modelpublic

delimitter/qwen3-1.7b-synoema-iot

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face

0likes21downloads

Model Card

Qwen3-1.7B Synoema IoT — 7/7 Fine-tune

LoRA adapter on Qwen/Qwen3-1.7B trained to generate Synoema IoT automation rules.

Score: 7/7 (100%) — stable across cycles C13-C16.

Property	Value
Base model	unsloth/qwen3-1.7b-unsloth-bnb-4bit
Adapter	LoRA QLoRA 4-bit
LoRA rank	r=16, alpha=32
Epochs	3
Train examples	5,455
Training cycle	C16 (stable 7/7 from C13)
Hardware	AMD RX 7900 GRE (ROCm, unsloth)