SafeVixAI/SafeVixAI-Dataset-Hub
SafeVixAI Dataset Hub 🛡️ The Intelligence Layer for the SafeVixAI platform — IIT Madras Road Safety Hackathon 2026 This repository hosts all datasets, pre-trained models, notebooks, and reproducible data acquisition scripts that power the SafeVixAI application. It is designed to be cloned directly into Google Colab or any research environment. Main Application Repo: SafeVixAI/SafeVixAI ⚡ Quickstart (Google Colab) # Clone the entire intelligence layer !git… See the full description on the dataset page: https://huggingface.co/datasets/SafeVixAI/SafeVixAI-Dataset-Hub.
1147
1# SafeVixAI — Research Notebooks 📓2 3These notebooks are the **research and training layer** of SafeVixAI. Each one processes raw data from the Hub and produces a model, index, or processed dataset used by the live application.4 5---6 7## ⚡ One-Click Colab Setup8 9Run this at the top of **any** notebook to mount the full dataset:10 11```python12# Step 1 — Clone the Hub (only run once per session)13!git clone https://huggingface.co/datasets/SafeVixAI/SafeVixAI-Dataset-Hub /content/hub14 15# Step 2 — Install dependencies16!pip install -q pdfplumber chromadb sentence-transformers ultralytics onnx17 18# Step 3 — Confirm data is accessible19import os20print("Data folders:", os.listdir("/content/hub"))21```22 23---24 25## 📓 Notebook Index26 27| # | Notebook | What it Produces | Input Data |28|---|---|---|---|29| 1 | `YOLOv8_Pothole_Detector_Training` | ONNX road damage model | `pothole_training/road_damage_2025/` |30| 2 | `ChromaDB_RAG_Vectorstore_Build` | ChromaDB index for legal RAG | `legal/`, `medical/` PDFs |31| 3 | `Accident_EDA_&_Hotspot_Generator` | Blackspot seed CSV + heatmap | `accidents/kaggle_india_accidents.csv` |32| 4 | `Roads_Data_Processing` | Sampled PMGSY GeoJSON | `roads/pmgsy_roads.geojson` |33| 5 | `Risk_Model_ONNX_Training` | Risk scoring ONNX model | `accidents/` + `roads/` |34 35---36 37## 📁 Data Paths Used by These Notebooks38 39All notebooks expect data at these paths after cloning the Hub:40 41```42/content/hub/43├── chatbot_service/data/44│ ├── accidents/ ← Notebooks 3, 545│ ├── legal/ ← Notebook 246│ ├── medical/ ← Notebook 247│ ├── pothole_training/ ← Notebook 148│ └── roads/ ← Notebook 449├── backend/datasets/ ← Notebook 4, 550└── frontend/public/models/ ← Output for Notebooks 1, 551```52 53---54 55## 🏃 Running Order56 57For a full pipeline run, execute notebooks in this order:58 59```601 → Train pothole detector → produces ONNX model612 → Build RAG vectorstore → produces ChromaDB index623 → Accident EDA → produces blackspot_seed.csv634 → Roads processing → produces sampled GeoJSON645 → Risk model → combines accidents + roads → risk ONNX65```66 67---68 69## 💾 Saving Outputs Back70 71After training, outputs land in `/content/hub/`. To persist them:72 73```python74# Commit outputs back to Hub (requires HF token)75import os76os.environ["HUGGING_FACE_TOKEN"] = "your_token_here"77 78!cd /content/hub && git config user.email "you@example.com"79!cd /content/hub && git config user.name "SafeVixAI"80!cd /content/hub && git add . && git commit -m "chore: update trained model outputs"81!cd /content/hub && git push https://your_username:$HUGGING_FACE_TOKEN@huggingface.co/datasets/SafeVixAI/SafeVixAI-Dataset-Hub main82```83 84---85 86*Part of the [SafeVixAI](https://github.com/SafeVixAI/SafeVixAI) — IIT Madras Road Safety Hackathon 2026*87