The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.
YAML Metadata
Warning:
empty or missing yaml metadata in repo card
(https://huggingface.co/docs/hub/datasets-cards)
Dhruv AI - Training Progress Report
Developer: Koyeliya Ghosh (koyelog)
Date: October 6, 2025
Status: All 3 Models Complete ✅
Project Overview
Dhruv AI is an AI-powered travel assistant for India with 3 core systems:
- Chatbot - Fine-tuned Llama 3.2 (1B) with LoRA
- Recommender - Hybrid ML recommendation system
- Monument Recognition - Custom EfficientNet CNN
Models
1. Chatbot (LLM)
- Base: Llama 3.2 1B Instruct
- Method: LoRA 4-bit fine-tuning
- Framework: Unsloth
- Training: 10 minutes
- Status: ✅ Complete
2. Recommendation System
- Algorithm: Content + Collaborative Filtering
- Framework: Scikit-learn
- Destinations: 20 (expandable to 200+)
- Accuracy: 85-92% expected
- Status: ✅ Complete
3. Monument Recognition
- Architecture: Custom EfficientNetV2
- Parameters: 38.4M
- Classes: 20-24 monuments
- Expected Accuracy: 85-95%
- Status: ✅ Architecture complete
Technology Stack
AI/ML: PyTorch, TensorFlow, Scikit-learn, Transformers, Unsloth
Backend: FastAPI, Uvicorn
Frontend: Streamlit / HTML
Deployment: Docker, Google Cloud Run
Model Links
- dhruv-ai-chatbot: https://huggingface.co/koyelog/dhruv-ai-chatbot
- dhruv-ai-recommender: https://huggingface.co/koyelog/dhruv-ai-recommender
- dhruv-ai-monument-recognition: https://huggingface.co/koyelog/dhruv-ai-monument-recognition
Next Steps
- Expand chatbot training data (1000+ examples)
- Complete monument CNN training on full dataset
- Expand destination database (100+ locations)
- Deploy API to cloud
- Build mobile app
Contact
- Developer: Koyeliya Ghosh
- Profile: https://huggingface.co/koyelog
- Project: Dhruv AI Travel Assistant
Last Updated: October 6, 2025
- Downloads last month
- 8