Fixes minor issues in cosmos documentation and evaluation script (#506)

# Description

Fixes for QA bugs:
- Update to Mimic-Cosmos pipeline doc to point users to the
Cosmos-Transfer1 setup instructions.
- Added a CLI argument to the robustness evaluation script for what
checkpoint to start the evaluation from. Updated the corresponding doc.

## Type of change

- Bug fix (non-breaking change which fixes an issue)
- This change requires a documentation update

## Checklist

- [x] I have run the [`pre-commit` checks](https://pre-commit.com/) with
`./isaaclab.sh --format`
- [x] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] I have updated the changelog and the corresponding version in the
extension's `config/extension.toml` file
- [x] I have added my name to the `CONTRIBUTORS.md` or my name already
exists there

---------
Signed-off-by: rwiltz <165190220+rwiltz@users.noreply.github.com>
Signed-off-by: Kelly Guo <kellyguo123@hotmail.com>
Signed-off-by: Ashwin Varghese Kuruttukulam <123109010+ashwinvkNV@users.noreply.github.com>
Signed-off-by: Kelly Guo <kellyg@nvidia.com>
Signed-off-by: Michael Gussert <michael@gussert.com>
Signed-off-by: samibouziri <79418773+samibouziri@users.noreply.github.com>
Signed-off-by: Mayank Mittal <12863862+Mayankm96@users.noreply.github.com>
Signed-off-by: Kyle Morgenstein <34984693+KyleM73@users.noreply.github.com>
Signed-off-by: Hongyu Li <lihongyu0807@icloud.com>
Signed-off-by: Toni-SM <toni.semu@gmail.com>
Signed-off-by: James Tigue <166445701+jtigue-bdai@users.noreply.github.com>
Signed-off-by: Pascal Roth <57946385+pascal-roth@users.noreply.github.com>
Signed-off-by: Victor Khaustov <3192677+vi3itor@users.noreply.github.com>
Signed-off-by: AlvinC <alvincny529@gmail.com>
Signed-off-by: Tyler Lum <tylergwlum@gmail.com>
Signed-off-by: Miguel Alonso Jr. <76960110+miguelalonsojr@users.noreply.github.com>
Signed-off-by: renaudponcelet <renaud.poncelet@gmail.com>
Co-authored-by: lotusl-code <lotusl@nvidia.com>
Co-authored-by: Kelly Guo <kellyg@nvidia.com>
Co-authored-by: jaczhangnv <jaczhang@nvidia.com>
Co-authored-by: rwiltz <165190220+rwiltz@users.noreply.github.com>
Co-authored-by: Yanzi Zhu <yanziz@nvidia.com>
Co-authored-by: nv-mhaselton <mhaselton@nvidia.com>
Co-authored-by: cosmith-nvidia <141183495+cosmith-nvidia@users.noreply.github.com>
Co-authored-by: Michael Gussert <michael@gussert.com>
Co-authored-by: CY Chen <cyc@nvidia.com>
Co-authored-by: oahmednv <oahmed@Nvidia.com>
Co-authored-by: Ashwin Varghese Kuruttukulam <123109010+ashwinvkNV@users.noreply.github.com>
Co-authored-by: Rafael Wiltz <rwiltz@nvidia.com>
Co-authored-by: Peter Du <peterd@nvidia.com>
Co-authored-by: matthewtrepte <mtrepte@nvidia.com>
Co-authored-by: chengronglai <chengrongl@nvidia.com>
Co-authored-by: pulkitg01 <pulkitg@nvidia.com>
Co-authored-by: Connor Smith <cosmith@nvidia.com>
Co-authored-by: Ashwin Varghese Kuruttukulam <ashwinvk@nvidia.com>
Co-authored-by: Kelly Guo <kellyguo123@hotmail.com>
Co-authored-by: Mayank Mittal <12863862+Mayankm96@users.noreply.github.com>
Co-authored-by: samibouziri <79418773+samibouziri@users.noreply.github.com>
Co-authored-by: James Smith <142246516+jsmith-bdai@users.noreply.github.com>
Co-authored-by: Shundo Kishi <syundo0730@gmail.com>
Co-authored-by: Sheikh Dawood <sabdulajees@nvidia.com>
Co-authored-by: Toni-SM <aserranomuno@nvidia.com>
Co-authored-by: Gonglitian <70052908+Gonglitian@users.noreply.github.com>
Co-authored-by: James Tigue <166445701+jtigue-bdai@users.noreply.github.com>
Co-authored-by: Mayank Mittal <mittalma@leggedrobotics.com>
Co-authored-by: Kyle Morgenstein <34984693+KyleM73@users.noreply.github.com>
Co-authored-by: Johnson Sun <20457146+j3soon@users.noreply.github.com>
Co-authored-by: Pascal Roth <57946385+pascal-roth@users.noreply.github.com>
Co-authored-by: Hongyu Li <lihongyu0807@icloud.com>
Co-authored-by: Jean-Francois-Lafleche <57650687+Jean-Francois-Lafleche@users.noreply.github.com>
Co-authored-by: Wei Jinqi <changshanshi@outlook.com>
Co-authored-by: Louis LE LAY <le.lay.louis@gmail.com>
Co-authored-by: Harsh Patel <hapatel@theaiinstitute.com>
Co-authored-by: Kousheek Chakraborty <kousheekc@gmail.com>
Co-authored-by: Victor Khaustov <3192677+vi3itor@users.noreply.github.com>
Co-authored-by: AlvinC <alvincny529@gmail.com>
Co-authored-by: Felipe Mohr <50018670+felipemohr@users.noreply.github.com>
Co-authored-by: AdAstra7 <87345760+likecanyon@users.noreply.github.com>
Co-authored-by: gao <ziqi.gao@iff-extern.fraunhofer.de>
Co-authored-by: Tyler Lum <tylergwlum@gmail.com>
Co-authored-by: -T.K.- <t_k_233@outlook.com>
Co-authored-by: Clemens Schwarke <96480707+ClemensSchwarke@users.noreply.github.com>
Co-authored-by: Miguel Alonso Jr. <76960110+miguelalonsojr@users.noreply.github.com>
Co-authored-by: Miguel Alonso Jr. <miguel.alonso@nfinite.app>
Co-authored-by: renaudponcelet <renaud.poncelet@gmail.com>
Co-authored-by: Ales Borovicka <aborovicka@nvidia.com>
Co-authored-by: nv-mm <mmagdics@nvidia.com>
Co-authored-by: Antoine RICHARD <antoiner@nvidia.com>

docs/source/overview/augmented_imitation.rst

@@ -92,14 +92,14 @@ We use the RGB, depth and shaded segmentation videos from the previous step as i
    :align: center
    :alt: RGB, depth and segmentation control inputs to Cosmos
 
-We provide an example augmentation output from `Cosmos Transfer1 <https://github.com/nvidia-cosmos/cosmos-transfer1/tree/1dca0bd1b3060d112bfec23c287572beabff9758>`_ below:
+We provide an example augmentation output from `Cosmos Transfer1 <https://github.com/nvidia-cosmos/cosmos-transfer1/tree/e4055e39ee9c53165e85275bdab84ed20909714a>`_ below:
 
 .. figure:: https://download.isaacsim.omniverse.nvidia.com/isaaclab/images/cosmos_output.gif
    :width: 100%
    :align: center
    :alt: Cosmos Transfer1 augmentation output
 
-We recommend using the `Cosmos Transfer1 <https://github.com/nvidia-cosmos/cosmos-transfer1/tree/1dca0bd1b3060d112bfec23c287572beabff9758>`_ model for visual augmentation as we found it to produce the best results in the form of a highly diverse dataset with a wide range of visual variations. You can refer to `this example <https://github.com/nvidia-cosmos/cosmos-transfer1/blob/1dca0bd1b3060d112bfec23c287572beabff9758/examples/inference_cosmos_transfer1_7b.md#example-2-multimodal-control>`_ for reference on how to use Transfer1 for this usecase. We further recommend the following settings to be used with the Transfer1 model for this task:
+We recommend using the `Cosmos Transfer1 <https://github.com/nvidia-cosmos/cosmos-transfer1/tree/e4055e39ee9c53165e85275bdab84ed20909714a>`_ model for visual augmentation as we found it to produce the best results in the form of a highly diverse dataset with a wide range of visual variations. You can refer to the installation instructions `here <https://github.com/nvidia-cosmos/cosmos-transfer1/blob/e4055e39ee9c53165e85275bdab84ed20909714a/INSTALL.md#environment-setup>`_, the checkpoint download instructions `here <https://github.com/nvidia-cosmos/cosmos-transfer1/blob/e4055e39ee9c53165e85275bdab84ed20909714a/examples/inference_cosmos_transfer1_7b.md#download-checkpoints>`_ and `this example <https://github.com/nvidia-cosmos/cosmos-transfer1/blob/e4055e39ee9c53165e85275bdab84ed20909714a/examples/inference_cosmos_transfer1_7b.md#example-2-multimodal-control>`_ for reference on how to use Transfer1 for this usecase. We further recommend the following settings to be used with the Transfer1 model for this task:
 
 .. rubric:: Hyperparameters
 
@@ -348,6 +348,8 @@ Below is an explanation of the different settings used for evaluation:
     :widths: 30 70
     :header-rows: 0
 
+    * - ``--start_epoch``
+      - Epoch of the checkpoint to start the evaluation from. (default: 100)
     * - ``--horizon``
       - Step horizon of each rollout. (default: 400)
     * - ``--num_rollouts``

scripts/imitation_learning/robomimic/robust_eval.py

@@ -37,6 +37,9 @@ parser.add_argument(
 )
 parser.add_argument("--task", type=str, default=None, help="Name of the task.")
 parser.add_argument("--input_dir", type=str, default=None, help="Directory containing models to evaluate.")
+parser.add_argument(
+    "--start_epoch", type=int, default=100, help="Epoch of the checkpoint to start the evaluation from."
+)
 parser.add_argument("--horizon", type=int, default=400, help="Step horizon of each rollout.")
 parser.add_argument("--num_rollouts", type=int, default=15, help="Number of rollouts for each setting.")
 parser.add_argument("--num_seeds", type=int, default=3, help="Number of random seeds to evaluate.")
@@ -279,7 +282,7 @@ def main() -> None:
                 for model in model_checkpoints:
                     # Skip early checkpoints
                     model_epoch = int(model.split(".")[0].split("_")[-1])
-                    if model_epoch <= 100:
+                    if model_epoch <= args_cli.start_epoch:
                         continue
 
                     model_path = os.path.join(args_cli.input_dir, model)