diff --git a/README.md b/README.md
index c0566e14e1d8594038e49bc5930bfc01dee4b08e..1096eb47d4989681d94305520fe0e3f0c202ffb6 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,21 @@
 # Instant Neural Graphics Primitives - for Tomography ![](https://github.com/NVlabs/instant-ngp/workflows/CI/badge.svg)
 
+Generate a save point (35000 steps by default, takes a few minutes):
+``` sh
+python ./scripts/run.py --training_data ../path/to/data/transforms.json --save_snapshot ../data/shepp-50-35000.ingp [--n_steps 20000]
+```
+
+Visualise it in the gui:
+```sh
+python ./scripts/run.py --load_snapshot ../data/shepp-50-35000.ingp --gui
+```
+
+Test it on unseen data (gives a PSNR - peak signal noise ratio).  Only the first image in the test dataset will be generated and compared to the ground truth
+```sh
+python ./scripts/run.py --load_snapshot ../data/shepp-50-35000.ingp --test_transforms ../data/shepp_test/transforms.json --screenshot_dir ../data
+```
+
+
 <img src="docs/assets_readme/fox.gif" height="342"/> <img src="docs/assets_readme/robot5.gif" height="342"/>
 
 Ever wanted to train a NeRF model of a fox in under 5 seconds? Or fly around a scene captured from photos of a factory robot? Of course you have!