thalos-hazmat-safety-v1

Thalos HazMat Safety Detection – v1.0 (Roboflow → HuggingFace export)

This repository contains the HazMat Safety Detection model used in the Thalos Safety Intelligence platform.

It identifies hazardous materials–related risks including:

• hazardous material placards
• chemical containers
• flammable / explosive warning labels
• spill / leak indicators
• hazmat storage violations
• environmental hazard markings
• DOT/OSHA/EPA classified hazard signage

This model was originally trained in Roboflow (Object Detection Large / YOLOv8-derived)
and exported as a PyTorch .pt file.
It is now hosted on HuggingFace for independent, cost-controlled inference inside the Thalos YOLO service.

---

🚀 Usage (Python)

from ultralytics import YOLO

model = YOLO("thalostech2025/thalos-hazmat-safety-v1/hazmat_safety_weights.pt")

results = model("example.jpg")
results.show()


⸻

🔗 Load From URL (Used by the Thalos YOLO Service)

import torch
import requests
from io import BytesIO

HF_URL = "https://huggingface.co/thalostech2025/thalos-hazmat-safety-v1/resolve/main/hazmat_safety_weights.pt"

response = requests.get(HF_URL)
model = torch.load(BytesIO(response.content), map_location="cpu")


⸻

🏷️ Labels / Classes

Based on the Roboflow v1.0 dataset:

hazmat
flammable
gas_cylinder
explosive
chemical_container
hazard_sign
spill
leak

(Class names may vary slightly based on the dataset version.)

⸻

📄 License

This model is distributed under the AGPL-3.0 license, consistent with Roboflow export requirements.

⸻

🔒 Important Notes
	•	This repository contains only the trained weights, not training data.
	•	Thalos loads this model through an internal YOLO microservice for inference.
	•	Future versions (v2.0+) may add multimodal features or integrate placard OCR.
	•	This model replaces Roboflow cloud inference to eliminate credit-based costs.

⸻

👤 Maintainer

Thalos Tech (2025)
Safety Intelligence & Hazard Detection Platform