YOLOv5-Streamlit-Deployment/README.md

🚀 YOLOv5 Streamlit Deployment

A Easy way to deploy YOLOv5 object detection model with Streamlit.

Please feel free to use/edit.

code modified by GitHub/thepbordin from GitHub/zhoroh

✨ Features

• YOLO Weights Source
  • Load from Local
  • Download Weights from URL
• Example Dataset
  • Videos

  • Images
    

• Upload Data
  • Video
  • Image
• Select computing device (cuda/cpu)

⚙️ Installation

Local Use

1. Install Requirements pip install -r requirements.txt
2. Install ffmpeg (for video inferencing)
   • For Windows read here
   • For Mac (brew) brew install ffmpeg
3. Strart Stremlit

   cd YOLOv5-Streamlit-Deployment
   streamlit run app.py
   

Streamlit Cloud

1. Edit a configuration in app.py (read ⚙️ Config Instruction)
2. (Optional) Upload example datas in
   • example_images
   • example_videos
3. Deploy on Streamlit

⚙️ Config Instruction

Download model from URL

1. Upload model to Internet Archive

2. Go to your uploaded file page.

3. From DOWNLOAD OPTIONS select SHOW ALL

4. Right click at .pt and Copy link address.

5. Edit config in app.py

   cfg_enable_url_download = True
   url = "your_model_url"
   

Use local .pt file:

Edit config in app.py

## CFG
cfg_model_path = "models/your_model_name.pt" 

Reference

Yolov5 Real-time Inference using Streamlit