Model Overview
reaperdoesntknow/Qwen3-1.7B-Coder-Distilled-SFT is a 1.7 billion parameter model built on the Qwen3 architecture, developed by Convergent Intelligence LLC: Research Division. Its unique training involves a two-stage distillation process. Initially, it was distilled from a 30B Qwen3-Coder teacher model, transferring a structured reasoning backbone for STEM derivations across 12 domains, including Physics, Linear Algebra, and Engineering. This stage utilized proof-weighted cross-entropy and knowledge distillation KL divergence. The second stage involved supervised fine-tuning on approximately 54,600 logical inference problems from the KonstantinDob/logic_inference_dataset, focusing on propositional logic and formal inference.
Key Capabilities
- Structured Reasoning: Inherits precise sequential logic, explicit state tracking, and compositional decomposition from its Coder teacher.
- Logical Inference: Proficient in propositional logic, logical entailment, and formal inference problems.
- STEM Derivation: Capable of rigorous derivations in various scientific and mathematical fields.
- Efficient Size: Achieves advanced reasoning capabilities within a compact 1.7B parameter count.
Good For
- Logical inference and propositional logic tasks.
- Formal reasoning and structured argumentation.
- STEM derivation and educational tutoring applications.
- Component in verification pipelines.
- Edge deployment via GGUF quantized versions.