---
base_model:
- Qwen/Qwen3-4B
tags:
- text-generation-inference
- transformers
- unsloth
- qwen3
- code
- cot
- thinking
- reasoning
- small
license: apache-2.0
language:
- en
datasets:
- nvidia/OpenCodeReasoning
- nvidia/OpenCodeInstruct
- ChavyvAkvar/qwen3-coder-480b-distill-mini-Converted
---

# Qwen3-4b-code-reasoning-v2
A small qwen 3 model with 4 billion parameters trained on `nvidia/OpenCodeReasoning` and `nvidia/OpenCodeInstruct`. The data is filtered only up to 32 thousands long.
Additionally it is anchored with a small subset of `ChavyvAkvar/qwen3-coder-480b-distill-mini-Converted`.

## Recommended parameters:
```
temp: 0.6-0.7
repeat_penalty: 1.2
top_p: 0.95
top_k: 20
min_p: 0.0
enable_thinking: True
```

# Uploaded finetuned  model

- **Developed by:** ertghiu256
- **License:** apache-2.0
- **Finetuned from model :** unsloth/qwen3-4b-unsloth-bnb-4bit

This qwen3 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.

[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)