https://github.com/koulanurag/vistorch
Interacting with Latent Space of AutoEncoder
https://github.com/koulanurag/vistorch
autoencoder dashboard latent-space plotly pytorch visualization
Last synced: 8 months ago
JSON representation
Interacting with Latent Space of AutoEncoder
- Host: GitHub
- URL: https://github.com/koulanurag/vistorch
- Owner: koulanurag
- Created: 2019-04-24T06:17:50.000Z (about 7 years ago)
- Default Branch: master
- Last Pushed: 2022-11-22T05:16:49.000Z (over 3 years ago)
- Last Synced: 2024-03-21T02:10:45.043Z (about 2 years ago)
- Topics: autoencoder, dashboard, latent-space, plotly, pytorch, visualization
- Language: Python
- Homepage:
- Size: 411 KB
- Stars: 21
- Watchers: 3
- Forks: 2
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# visTorch
visTorch is a simple tool for interactive visualization of latent space of an Autoencoder. It's built using [Plotly's Dash framework](https://plot.ly/products/dash/) and has been tested with Pytorch.
## Installation
```bash
pip install -r requirements.txt
python setup.py install
```
## Usage
```python
from visTorch import visboard
...
# your code
...
vis_board = visboard()
vis_board.add_ae(model,
dataset,
latent_options={'n': model.latent_size, 'min': -1, 'max': 1, 'step': 0.05},
model_paths={'Best': ae_path})
vis_board.run_server('127.0.0.1','8080')
```
Refer to ```.\examples``` for more details
## Demo
