Image Embeddings and Interactive Visualization

This repo is a minimal starter project for creating image embeddings (here using UNICOM), and then visualizing and exploring those embeddings with static and interactive plots:

• Rich feature representation through UNICOM
• Interactive plotting using the Bokeh Python package - also see my blog post

Quickstart

1. Clone this repo, and the UNICOM repo:

Make sure you have git and git lfs installed on your machine (e.g. via git lfs install).

git clone https://github.com/mluerig/demo-image-embeddings
git clone https://github.com/deepglint/unicom

2. Install mamba via miniforge, or conda (Miniconda/Anaconda).

3. Open a terminal at the repo root dir, and create the environment from environment.yml:

mamba env create -f environment.yml -n unicom1

Then activate it:

mamba activate unicom1

4. Optional: install PyTorch with GPU (NVIDIA) support. If you have a compatible CUDA GPU and drivers, install the CUDA build from the official channels:

mamba install nvidia::cuda-toolkit==12.6
pip install torch torchvision --index-url https://download.pytorch.org/whl/cu126 --force-reinstall

If you do not have a GPU, you can skip this step. Embedding generation will still run on CPU, but it will be much slower.

5. Prepare the example data and run the notebooks:

• create embeddings in scripts/01_create_embeddings.ipynb
• visualize and explore them in scripts/02_make_cluster_plots.ipynb

Notes

• The provided environment.yml keeps PyTorch out by default so you can choose CPU or GPU builds explicitly.
• The example workflow uses UNICOM embeddings, but the plotting notebooks can also be adapted to embeddings from other models.