Views
No views yet
.pt file.1from ultralytics import YOLO
2
3model = YOLO("thalostech2025/thalos-forklift-safety-v1/forklift_safety_weights.pt")
4
5results = model("example.jpg")
6results.show()
7
8
9⸻
10
11🔗 Load From URL (Used by the Thalos YOLO Service)
12
13import torch
14import requests
15from io import BytesIO
16
17HF_URL = "https://huggingface.co/thalostech2025/thalos-forklift-safety-v1/resolve/main/forklift_safety_weights.pt"
18
19response = requests.get(HF_URL)
20model = torch.load(BytesIO(response.content), map_location="cpu")
21
22
23⸻
24
25🏷️ Labels / Classes
26
27The following classes are included based on the v1.0 dataset:
28
29forklift
30forklift_near_person
31forklift_operation_zone
32mhe_vehicle
33mhe_near_pedestrian
34pedestrian
35
36(Class names may vary slightly depending on dataset configuration.)
37
38⸻
39
40📄 License
41
42This model is distributed under the AGPL-3.0 license, consistent with Roboflow export requirements.
43
44⸻
45
46🔒 Important Notes
47 • This repository contains only the model weights, not training data.
48 • Thalos uses this model through its internal YOLO inference service.
49 • Future versions (v2.0+) may expand to include forklift–vehicle detection and distance-based safety scoring.
50 • Migrating to HuggingFace helps eliminate Roboflow inference-credit costs.
51
52⸻
53
54👤 Maintainer
55
56Thalos Tech (2025)
57Safety Intelligence & Hazard Detection Platform