Modelpublic

Danny-jin/math-sft-a17-lora

sourceHugging Faceupdated 4mo agoView on Hugging Face

0likes8downloads

Model Card

Final A17 LoRA Adapter

This repository contains the final A17 LoRA adapter used by Danny-jin/math-sft-final-inference.

The base model is Qwen/Qwen3-4B-Thinking-2507. In the inference repository, set QWEN_BASE_MODEL to either that HuggingFace id or a local snapshot path, then load this adapter with vLLM LoRA support.

Required files:

—adapter_config.json
—adapter_model.safetensors
—tokenizer files copied from the training run