CLIP Embedding API (FastAPI)

This project is a FastAPI server that provides endpoints for generating text, image, and combined (text + image) embeddings using the clip-ViT-B-32 model from SentenceTransformers.

---

Requirements

Make sure you have Python 3.7+ installed.

Install all dependencies using the provided batch file:

requirementsInstall.bat

---

Running the Server

To start the FastAPI server, simply run:

start.bat

---

API Access

Once the server is running, you can access the API locally at:

http://127.0.0.1:8000

The automatic interactive Swagger UI is available at:

http://127.0.0.1:8000/docs

---

API Endpoints

1. /embed-text

• Method: POST

• Body (form):

  • text: string

• Returns: Normalized text embedding

---

2. /embed-image

• Method: POST

• Body (form-data):

  • file: image file (JPG, PNG, etc.)

• Returns: Normalized image embedding

---

3. /embed-both

• Method: POST

• Body (form-data):

  • text: string
  • file: image file

• Returns: Averaged embedding from both image and text

---

Notes

• Model used: clip-ViT-B-32
• Supports GPU acceleration if available (automatically detected).
• CORS is enabled for * (can be adjusted in code).

---

Example curl Request

curl -X POST http://127.0.0.1:8000/embed-text -F "text=A dog playing guitar"