Update human demonstration dataset guidance

docs/algorithms/demonstrations.md

@@ -14,12 +14,14 @@ See the [devices page](https://robosuite.ai/docs/modules/devices.html) for detai
 
 ## Replaying Human Demonstrations
 
-We have included an example script that illustrates how demonstrations can be loaded and played back. Our [playback_demonstrations_from_hdf5](https://github.com/ARISE-Initiative/robosuite/blob/master/robosuite/scripts/playback_demonstrations_from_hdf5.py) script selects demonstration episodes at random from a demonstration pickle file and replays them.
+We have included an example script that illustrates how demonstrations can be loaded and played back. Our [playback_demonstrations_from_hdf5](https://github.com/ARISE-Initiative/robosuite/blob/master/robosuite/scripts/playback_demonstrations_from_hdf5.py) script selects demonstration episodes at random from a `demo.hdf5` file collected with robosuite and replays them.
 
 
 ## Existing Datasets
 
-We have included some sample demonstrations for each task at `models/assets/demonstrations`.
+Sample demonstrations are no longer bundled with the robosuite repository. To create a `demo.hdf5` file that can be used directly with the playback script above, collect demonstrations with `collect_human_demonstrations.py`.
+
+Public robot manipulation datasets are available from the [robomimic v0.1 dataset collection](https://robomimic.github.io/docs/datasets/robomimic_v0.1.html), which provides a download script and direct download links for the tasks it supports. These datasets are distributed for use with robomimic, so consult its [robosuite dataset documentation](https://robomimic.github.io/docs/datasets/robosuite.html) for format conversion and observation extraction instructions.
 
 
 ## Structure of collected demonstrations
@@ -55,11 +57,11 @@ The reason for storing mujoco states instead of raw observations is to make it e
 
 ## Using Demonstrations for Learning
 
-The [robomimic](https://arise-initiative.github.io/robomimic-web/) framework makes it easy to train policies using your own [datasets collected with robosuite](https://arise-initiative.github.io/robomimic-web/docs/introduction/datasets.html#robosuite-hdf5-datasets). The framework also contains many useful examples for how to integrate hdf5 datasets into your own learning pipeline.
+The [robomimic](https://robomimic.github.io/docs/) framework makes it easy to train policies using your own [datasets collected with robosuite](https://robomimic.github.io/docs/datasets/robosuite.html). The framework also contains many useful examples for how to integrate hdf5 datasets into your own learning pipeline.
 
 The robosuite repository also has some utilities for using the demonstrations to alter the start state distribution of training episodes for learning RL policies - this have proved effective in [several](https://arxiv.org/abs/1802.09564) [prior](https://arxiv.org/abs/1807.06919) [works](https://arxiv.org/abs/1804.02717). For example, we provide a generic utility for setting various types of learning curriculums which dictate how to sample from demonstration episodes when doing an environment reset. For more information see the `DemoSamplerWrapper` class.
 
 ## Warnings
 We have verified that deterministic action playback works specifically when playing back demonstrations on the *same machine* that the demonstrations were originally collected upon. However, this means that deterministic action playback is NOT guaranteed (in fact, very unlikely) to work across platforms or even across different machines using the same OS.
 
-While action playback trajectories are quite similar even if not completely identical to the original collected state trajectories, they do tend to drift over time, and should not be relied upon to accurately replicate demonstrations. Instead, we recommend directly setting states to reproduce the collected trajectories, as shown in [playback_demonstrations_from_hdf5](https://github.com/ARISE-Initiative/robosuite/blob/master/robosuite/scripts/playback_demonstrations_from_hdf5.py).
+While action playback trajectories are quite similar even if not completely identical to the original collected state trajectories, they do tend to drift over time, and should not be relied upon to accurately replicate demonstrations. Instead, we recommend directly setting states to reproduce the collected trajectories, as shown in [playback_demonstrations_from_hdf5](https://github.com/ARISE-Initiative/robosuite/blob/master/robosuite/scripts/playback_demonstrations_from_hdf5.py).

robosuite/scripts/playback_demonstrations_from_hdf5.py

@@ -10,7 +10,7 @@
     --visualize-gripper (optional): If set, will visualize the gripper site
 
 Example:
-    $ python playback_demonstrations_from_hdf5.py --folder ../models/assets/demonstrations/lift/
+    $ python playback_demonstrations_from_hdf5.py --folder /path/to/demonstrations/lift/
 """
 
 import argparse
@@ -30,7 +30,7 @@
         type=str,
         required=True,
         help="Path to your demonstration folder that contains the demo.hdf5 file, e.g.: "
-        "'path_to_assets_dir/demonstrations/YOUR_DEMONSTRATION'",
+        "'/path/to/demonstrations/YOUR_DEMONSTRATION'",
     ),
     parser.add_argument(
         "--use-actions",