d3-lite body a4
A Zils d3-lite body (Qwen3.5-0.8B decision model). Fine-tuned from the current d3-lite soup with a rank-16 LoRA (one epoch, AdamW lr 1e-4, linear schedule) on function-calling decisions (call next, which call, required arguments) built from public function-calling conversations, with a KL anchor to the starting soup's answers; the LoRA is merged into the weights.
- Downloads last month
- 12
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐ Ask for provider support