# QUIN3-1.7B-Plant-Based
## Model Summary
This model is a fine-tuned version of **Qwen3-1.7B (base)** using LoRA.
The fine-tune focuses on nutrition, diet, and related Q&A tasks.
- **Base model:** Qwen3-1.7B (Apache 2.0)
- **Fine-tune method:** LoRA, merged for ease of use
- **Parameters:** 1.7B
- **Dataset size:** ~1,255 instruction–response pairs
- **Data sources:**
- Synthetic data generated from transcripts of [NutritionFacts.org](https://nutritionfacts.org) videos and blog posts
- Mix of adversarial examples and trivia questions
## Intended Use
- Research and educational purposes
- Personal experimentation
- **Not suitable for clinical or commercial use**
## Limitations
- Trained on a small dataset
- Not a medical authority
- Occasionally outputs non-English characters
- Long outputs may truncate or degrade in quality
## Usage
Example (Transformers):
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
tokenizer = AutoTokeni…
