ONNX port of Unicom model from open-metric-learning.
This model is intended to be used for similarity search.
Usage
Here's an example of performing inference using the model with FastEmbed.
from fastembed import ImageEmbedding
images = [
"./path/to/image1.jpg",
"./path/to/image2.jpg",
]
model = ImageEmbedding(model_name="Qdrant/Unicom-ViT-B-32")
embeddings = list(model.embed(images))
# [
# array([ 0.04177791, 0.0550059 , 0.00025418, 0.0252876 , ..., dtype=float32),
# array([2.23932182e-03, 4.68995124e-02, 3.28772422e-03, 7.57176951e-02, ...], dtype=float32)
# ]