---
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- output:
    url: images/vlcsnap-2025-09-11-22h02m37s694.png
  text: '-'
- output:
    url: images/vlcsnap-2025-09-11-22h02m27s197.png
  text: '-'
- output:
    url: images/vlcsnap-2025-09-11-22h02m08s816.png
  text: '-'
base_model: openai/gpt-oss-20b
instance_prompt: gpt-oss, react components, react, frontend javascript
license: mit
---
# RTune GPT OSS 20B Fine-Tune

<Gallery />

## Model description 

This model is a fine-tuned version of GPT-OSS 20B, specifically trained to generate high-quality React components with modern best practices. Using a teacher-student approach, it was trained on a curated dataset of 78+ React components generated by GPT-OSS 120B and scored above 75&#x2F;100 for quality. The model excels at creating production-ready JSX with React hooks, Tailwind CSS styling, semantic HTML, and realistic mock data across four aesthetic styles (business, indie, blue, art). Fine-tuned using QLoRA on an RTX 5090 with Unsloth optimization, it achieves GPT-4 level component quality while running efficiently on consumer hardware for completely private, offline code generation.

## Trigger words

You should use `gpt-oss` to trigger the image generation.

You should use `react components` to trigger the image generation.

You should use `react` to trigger the image generation.

You should use `frontend javascript` to trigger the image generation.


## Download model


[Download](/Dogestella/rtune-gpt-oss-20b-finetune/tree/main) them in the Files & versions tab.