A reasoning-enhanced fine-tuned version of Qwen/Qwen3-1.7B, trained on DeepSeek-R1 distilled thought chains from the Bespoke-Stratos-17k dataset.
Qwen/Qwen3-1.7Br=16, alpha=16) via Unslothbespokelabs/Bespoke-Stratos-17k (1,500 high-quality reasoning samples)Q4_K_M (GGUF)<think> / <|begin_of_thought|>) before generating the final solution.Qwen/Qwen3-1.7B