The model can be used to generate recipes directly via text prompts like:
Generate a high-protein, low-fat recipe with around 400 calories.
Out-of-Scope Use
This model is not intended for medical diagnosis, treatment planning, or diet prescriptions requiring professional approval.
Bias, Risks, and Limitations
The model was trained on a custom dataset built by the author. It may not generalize well to all types of cuisines, dietary needs, or nutritional guidelines. It does not replace professional dietary advice.
Recommendations
Always consult a certified nutritionist or dietitian before following specific diets, especially if you have health conditions.
How to Get Started with the Model
python
1from transformers import pipeline
23generator = pipeline("text-generation", model="samdak93/qrit-2")4prompt ="Healthy dinner recipe under 400 calories, high protein"5output = generator(prompt, max_new_tokens=200)6print(output[0]["generated_text"])
Training Details
Training Data
The model was trained on a custom dataset of food recipes with nutrition tags and instructions built by the author.
Training Procedure
Platform: Google Colab (free tier)
Compute: Colab-provided GPU and RAM
Training regime: fp16 mixed precision
Evaluation
The model's output was evaluated manually for relevance, nutrition tag accuracy, and coherence of recipe instructions.
Environmental Impact
Hardware Type: Google Colab (free tier GPU)
Hours used: Approx. 6 hours
Cloud Provider: Google
Compute Region: Unknown
Carbon Emitted: Low (estimated via shared environment and short training time)
Technical Specifications
Model Architecture and Objective
The model is a fine-tuned version of GPT-2 (openai-community/gpt2) trained to generate nutrition-based recipes.
Compute Infrastructure
Hardware: Google Colab free GPU
Software: Python, Transformers, PyTorch
Citation
BibTeX:
@misc{qrit2,
author = {samdak93},
title = {qrit-2: Nutrition-based Recipe Generator},
year = {2025},
howpublished = {\url{https://huggingface.co/samdak93/qrit-2}},
}
