Update README.md

README.md

@@ -266,15 +266,15 @@ Please note that every package indexed by the parquet files can contan several i
 
 ## Data access
 
-A standalone example of how to access the episode data via streaming is included in [stream_access_example.py](stream_access_example.py).
+We recommend to use [bulk_download.py](bulk_download.py).
+
 To launch it you might need to install some dependencies, e.g., by:
 ```bash
-pip install h5py 'imageio[ffmpeg]' requests zstandard datasets huggingface_hub
+pip install zstandard datasets huggingface_hub tqdm
 ```
 
-### Bulk download
+A standalone example of how to access the episode data via streaming is also included in [stream_access_example.py](stream_access_example.py).
 
-We recommend to use [bulk_download.py](bulk_download.py).
 
 ### Dataset post-processing and stats