test

DATASET.md

@@ -0,0 +1,19 @@
+## 📖 Dataset
+
+- Go to the official website of the [NuScenes dataset](https://www.nuscenes.org/download). You will need to create an 
+account. 
+- Download the _Full dataset (v1.0)_. This includes the _Mini_ dataset and the _Trainval_ dataset with all the 
+metadata and sensor file blobs.
+- Extract the tar files in order to obtain the following folder structure.
+
+  nuscenes
+    |
+     -- mini
+        |
+         -- maps
+        |
+         -- samples
+        |
+         -- sweeps
+        |
+         -- v1.0-mini

README.md

@@ -49,7 +49,7 @@ Or locally:
 them to an `output_vis` folder.
 
 ### Evaluation
-- Download the [NuScenes dataset](https://www.nuscenes.org/download) (Full dataset (v1.0)).
+- Download the [NuScenes dataset](https://www.nuscenes.org/download). For detailed instructions, see [DATASET.md](DATASET.md).
 - Download [pre-trained weights](https://github.com/wayveai/fiery/releases/download/v1.0/fiery.ckpt).
 - Run `python evaluate.py --checkpoint ${CHECKPOINT_PATH} --dataroot ${NUSCENES_DATAROOT}`.