https://github.com/robmarkcole/umap-image-embedding-streamlit-app

App to explore umap image embeddings for MNIST class datasets
https://github.com/robmarkcole/umap-image-embedding-streamlit-app

bokeh matplotlib mnist streamlit umap

Last synced: 3 days ago
JSON representation

App to explore umap image embeddings for MNIST class datasets

• Host: GitHub
• URL: https://github.com/robmarkcole/umap-image-embedding-streamlit-app
• Owner: robmarkcole
• License: apache-2.0
• Created: 2021-10-15T03:30:54.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2021-10-15T06:44:59.000Z (over 3 years ago)
• Last Synced: 2025-01-29T12:39:44.446Z (10 days ago)
• Topics: bokeh, matplotlib, mnist, streamlit, umap
• Language: Jupyter Notebook
• Homepage:
• Size: 12.1 MB
• Stars: 2
• Watchers: 2
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# umap-image-embedding-streamlit-app

App to explore umap image embeddings for MNIST class datasets.







## UMAP

Umap depends on numba which itself uses a pinned version of numpy. This dependency limitation can be avoided by splitting the generation of embeddings and the plotting of embeddings into different envs if required. Simply use the exported `embedding.npy` file

## Dev

* `python3 -m venv venv`

* `source venv/bin/activate`

* `pip install -r requirements.txt`

* `pip install jupyterlab` for dev or `streamlit run app.py` for the app