update readme

README.md

@@ -68,26 +68,25 @@ This dataset is useful for various computer vision tasks, including depth estima
 ```python
 from datasets import load_dataset
 
-# Load the dataset
 dataset = load_dataset("0jl/NYUv2", trust_remote_code=True, split="train")
 ```
 
 
 ### Common Errors
 
-#### `fsspec.exceptions.FSTimeoutError`
+* `fsspec.exceptions.FSTimeoutError`
 
-Can occur for `datasets==3.0` when the download takes more than 5 minutes.
+  Can occur for `datasets==3.0` when the download takes more than 5 minutes. This increases the timeout to 1 hour:
 
-```python
-import datasets, aiohttp
-dataset = datasets.load_dataset(
-    "0jl/NYUv2",
-    trust_remote_code=True,
-    split="train",
-    storage_options={'client_kwargs': {'timeout': aiohttp.ClientTimeout(total=3600)}}
-)
-```
+  ```python
+  import datasets, aiohttp
+  dataset = datasets.load_dataset(
+      "0jl/NYUv2",
+      trust_remote_code=True,
+      split="train",
+      storage_options={'client_kwargs': {'timeout': aiohttp.ClientTimeout(total=3600)}}
+  )
+  ```
 
 
 ## Dataset Structure
@@ -101,7 +100,7 @@ The dataset contains only one training split with the following features:
 - `scene_type`: Scene type (string)
 
 
-### Citation Information
+## Citation Information
 
 If you use this dataset, please cite the original paper:
 
@@ -112,4 +111,4 @@ If you use this dataset, please cite the original paper:
   booktitle = {Proceedings of the European Conference on Computer Vision},
   year      = {2012}
 }
-```
+```