metadata
language:
- en
license: apache-2.0
tags:
- duoneural
- sft
- qwen
- qwen2.5-coder
base_model: Qwen/Qwen2.5-Coder-3B-Instruct
datasets:
- DuoNeural/Gemma4-E2B-SFT-WebCode
Qwen2.5-Coder-3B-SFT-WebCode
📊 Recorded — SFT fine-tune by DuoNeural.
Base model: Qwen/Qwen2.5-Coder-3B-Instruct
Dataset: DuoNeural/Gemma4-E2B-SFT-WebCode
Training: LoRA rank=16 α=32, 3 epochs, lr=2e-4, effective batch=16
Eval: GSM8K + ARC-Challenge via lm_eval 0.4.x
Benchmark Results
| Model | GSM8K flex | ARC-norm | ARC-acc |
|---|---|---|---|
| Baseline | 0.5807 | 0.4957 | 0.4590 |
| Qwen2.5-Coder-3B-SFT-WebCode | 0.3207 | 0.4957 | 0.4590 |
| Δ | -0.2600 | +0.0000 | +0.0000 |
About DuoNeural
Post-training research lab exploring emergent behaviors in small language models. We publish datasets, models, and research papers.
Generated by Archon — DuoNeural lab AI