From 95861e3a5fed7acd870ad7057cc55d73bc8a4cd4 Mon Sep 17 00:00:00 2001 From: Robert Jan Schlimbach Date: Thu, 23 Sep 2021 12:00:41 +0200 Subject: [PATCH] Update README.md --- README.md | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/README.md b/README.md index d971820..268ff89 100644 --- a/README.md +++ b/README.md @@ -1,23 +1,28 @@ # 2D-VQ-AE-2 2D Vector-Quantized Auto-Encoder for compression of Whole-Slide Images in Histopathology -# How to run -## Locally +## How to run +### Locally set `CAMELYON16_PATH` and run `train.py`: ```bash CAMELYON16_PATH= python train.py ``` -## Lisa +### Lisa set `CAMELYON16_PATH`, and append `--multirun` to automatically submit a `sbatch` job through `submitit`, which copies CAMELYON16 to `$SCRATCH` of the allocated node. ```bash CAMELYON16_PATH= python train.py --multirun ``` - Change node type by overwriting the node config, e.g.: ```bash CAMELYON16_PATH= python train.py hydra/launcher/node@hydra.launcher=gpu_titanrtx --multirun ``` +## WIP results +Top: original, bottom: reconstruction. +Input dimensionality: `256×256×3`, latent dimensionality: `32×32` (i.e. `99.47%` compression). + +![image](https://user-images.githubusercontent.com/5969044/134488209-4c1696d3-6478-41d0-a7bf-e7e99544382b.png) + ## Research