101 lines
3.9 KiB
Markdown
101 lines
3.9 KiB
Markdown
|
---
|
|||
|
|
license: apache-2.0
|
|||
|
|
base_model:
|
|||
|
|
- prithivMLmods/Muscae-Qwen3-UI-Code-4B
|
|||
|
|
language:
|
|||
|
|
- en
|
|||
|
|
pipeline_tag: text-generation
|
|||
|
|
library_name: transformers
|
|||
|
|
tags:
|
|||
|
|
- trl
|
|||
|
|
- text-generation-inference
|
|||
|
|
- code
|
|||
|
|
- ui
|
|||
|
|
---
|
|||
|
|
|
|||
|
|
|
|||
|
|
|
|||
|
|
# **Octans-Qwen3-UI-Code-4B**
|
|||
|
|
|
|||
|
|
> **Octans-Qwen3-UI-Code-4B** is an **optimized successor** of **Muscae-Qwen3-UI-Code-4B**, fine-tuned for enhanced **UI reasoning precision**, **layout structuring**, and **frontend code synthesis**.
|
|||
|
|
> Built upon **Qwen3-4B** and refined through **Abliterated Reasoning Optimization**, it delivers **balanced, structured, and production-grade** UI code outputs for experimental and research use.
|
|||
|
|
> Ideal for **frontend developers**, **UI engineers**, and **design system researchers** exploring next-generation code synthesis.
|
|||
|
|
|
|||
|
|
> [!note]
|
|||
|
|
> GGUF: [https://huggingface.co/prithivMLmods/Octans-Qwen3-UI-Code-4B-GGUF](https://huggingface.co/prithivMLmods/Octans-Qwen3-UI-Code-4B-GGUF)
|
|||
|
|
|
|||
|
|
## **Key Features**
|
|||
|
|
|
|||
|
|
1. **Enhanced UI-Oriented Reasoning**
|
|||
|
|
Upgraded reasoning calibration from *Muscae* with deeper token optimization for **frontend logic**, **layout reasoning**, and **component cohesion**.
|
|||
|
|
|
|||
|
|
2. **Refined Web UI Component Generation**
|
|||
|
|
Generates **responsive, accessible**, and **semantic** UI components with **Tailwind**, **React**, and **HTML5**, ensuring cleaner syntax and reduced redundancy.
|
|||
|
|
|
|||
|
|
3. **Improved Layout-Aware Structure**
|
|||
|
|
Demonstrates superior understanding of **hierarchical design**, **stateful components**, and **responsive alignment**, enhancing multi-device compatibility.
|
|||
|
|
|
|||
|
|
4. **Optimized Hybrid Reasoning Engine**
|
|||
|
|
Integrates symbolic and probabilistic logic for **event-driven** UI workflows, conditional rendering, and state synchronization in code outputs.
|
|||
|
|
|
|||
|
|
5. **Structured Output Excellence**
|
|||
|
|
Produces consistent results in **HTML**, **React**, **Markdown**, **JSON**, and **YAML**, suitable for **UI prototyping**, **design systems**, and **auto-documentation**.
|
|||
|
|
|
|||
|
|
6. **Lightweight and Deployable**
|
|||
|
|
Maintains a **4B parameter** scale, optimized for **mid-range GPUs**, **edge inference**, or **offline environments**, without compromising structure or reasoning depth.
|
|||
|
|
|
|||
|
|
## **Quickstart with Transformers**
|
|||
|
|
|
|||
|
|
```python
|
|||
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|||
|
|
|
|||
|
|
model_name = "prithivMLmods/Octans-Qwen3-UI-Code-4B"
|
|||
|
|
|
|||
|
|
model = AutoModelForCausalLM.from_pretrained(
|
|||
|
|
model_name,
|
|||
|
|
torch_dtype="auto",
|
|||
|
|
device_map="auto"
|
|||
|
|
)
|
|||
|
|
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
|||
|
|
|
|||
|
|
prompt = "Generate a responsive dashboard layout with Tailwind and modular React components."
|
|||
|
|
|
|||
|
|
messages = [
|
|||
|
|
{"role": "system", "content": "You are a frontend coding assistant skilled in UI generation, semantic HTML, and structured React components."},
|
|||
|
|
{"role": "user", "content": prompt}
|
|||
|
|
]
|
|||
|
|
|
|||
|
|
text = tokenizer.apply_chat_template(
|
|||
|
|
messages,
|
|||
|
|
tokenize=False,
|
|||
|
|
add_generation_prompt=True
|
|||
|
|
)
|
|||
|
|
|
|||
|
|
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
|
|||
|
|
|
|||
|
|
generated_ids = model.generate(
|
|||
|
|
**model_inputs,
|
|||
|
|
max_new_tokens=512
|
|||
|
|
)
|
|||
|
|
generated_ids = [
|
|||
|
|
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
|
|||
|
|
]
|
|||
|
|
|
|||
|
|
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
|||
|
|
print(response)
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
## **Intended Use**
|
|||
|
|
|
|||
|
|
* Advanced web UI and component code generation
|
|||
|
|
* Responsive frontend prototyping with Tailwind/React
|
|||
|
|
* Research on **structured reasoning** in code synthesis
|
|||
|
|
* Semantic, design-system-aligned component generation
|
|||
|
|
* Experimental projects exploring **UI intelligence modeling**
|
|||
|
|
|
|||
|
|
## **Limitations**
|
|||
|
|
|
|||
|
|
* Research-focused model – not fine-tuned for production-critical pipelines
|
|||
|
|
* Specialized for **UI code** – not suitable for general text generation or long-form reasoning
|
|||
|
|
* May exhibit variability with **cross-framework** or **overextended prompts**
|
|||
|
|
* Prioritizes **code structure and logic clarity** over aesthetic or creative expression.
|