A framework to explore the latent space of autoencoders implemented in torch

Project description

lightcone

A framework to explore the latent space of convolutional autoencoders implemented in pytorch.

Compose your decoder and your encoder into the lightcone autoencoder:

from lightcone.models import AutoEncoder

model = AutoEncoder(encoder=your_encoder, decoder=your_decoder)

After model has been training, the latent space can be explored in a Jupyter-Notebook as follows

model.explore(data_loader=your_data_loader)

Make sure to install and activate the Jupyter notebook extenstion

jupyter nbextension install --py jupyter_dash
jupyter nbextension enable --py jupyter_dash

This version

0.1.0

Jan 3, 2023

0.0.1

Dec 28, 2022

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Uploaded Jan 3, 2023 Source

Uploaded Jan 3, 2023 Python 3

Details for the file lightcone-0.1.0.tar.gz.

Hashes for lightcone-0.1.0.tar.gz
Algorithm	Hash digest
SHA256	`51ecdfe50385a35c35c1502e691e8487ffdd80465c540d82e1ffcdb6e2e93a89`
MD5	`18fe6a4b2b78744c0d9c55c15fee529c`
BLAKE2b-256	`856429831ea18946aae80840f0439c04eb73a6e49b6f71241ea13dd8e5809d0e`

Details for the file lightcone-0.1.0-py3-none-any.whl.

Hashes for lightcone-0.1.0-py3-none-any.whl
Algorithm	Hash digest
SHA256	`68887303c5ca3d717c7da121b3ded9e3eaa6932a3517e38942ff77fdbbca8d36`
MD5	`d7672ecdfba9bff3f3b0bc1ad7006761`
BLAKE2b-256	`f9a77de46fef5ac9f66c477bdf89e3e4414c2328cf2e5dbf7420e23ddddf5062`