SynGen-Nano-1.7B

<center> <img src="title.svg" style="height: 50px"> (Dataset based on <a href="https://huggingface.co/datasets/Pinkstack/syngen-reasoning-0.6b-dataset">Pinkstack/syngen-reasoning-0.6b-dataset</a>) </center>

This is a 1.7B parameter LLM designed for synthetic grounded reasoning generation between the user prompt and the final model output, specifically for dataset modifications, but can be used for multiple use cases which require reasoning.

For example, this model allows you to turn any chat dataset into a reasoning dataset as if it was generated by DeepSeek R1 or OpenAI's GPT OSS!

Sampler Settings: Pretty standard, temp = 0.7, top_p = 0.95.

---

Prompt Format

System Message

<reasoning_style>deepseek_r1</reasoning_style> # Can replace deepseek_r1 with gpt_oss
<system_prompt>Original System Prompt</system_prompt>

Prompt Message

<user>User Message Here</user>
<assistant>Assistant Final Response Here (without reasoning)</assistant>

Output Format

<think>Generated Reasoning</think>

---

Training Details

• Base Model: Qwen/Qwen3-1.7B
• Training Epochs: 0.4
• Learning Rate: 2e-5
• Batch Size: 32
• Training Method: Full Fine-Tune (FFT)
• Training Hardware: 2 x A100-40GB
• Training Platform: Burla
• Total Cost: $0.00 USD
• Seed: 42