Merge pull request #3 from jacarvalho/fix-setup

Fix setup
This commit is contained in:
João Carvalho 2023-11-20 11:10:15 +01:00 committed by GitHub
commit 8caf8a4142
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
4 changed files with 33 additions and 10 deletions

6
.gitmodules vendored
View File

@ -3,10 +3,10 @@
url = https://git.ias.informatik.tu-darmstadt.de/common/experiment_launcher.git
[submodule "deps/motion_planning_baselines"]
path = deps/motion_planning_baselines
url = git@github.com:anindex/motion_planning_baselines.git
url = https://github.com/anindex/motion_planning_baselines.git
[submodule "deps/torch_robotics"]
path = deps/torch_robotics
url = git@github.com:anindex/torch_robotics.git
url = https://github.com/anindex/torch_robotics.git
[submodule "deps/storm"]
path = deps/storm
url = git@github.com:jacarvalho/storm.git
url = https://github.com/jacarvalho/storm.git

View File

@ -27,10 +27,27 @@ If you have any questions please let me know -- [joao@robot-learning.de](mailto:
---
## Installation
Download [IsaacGym Preview 4](https://developer.nvidia.com/isaac-gym) and extract it under `deps/isaacgym`.
Pre-requisites:
- Ubuntu 20.04 (maybe works with other OS)
- [miniconda](https://docs.conda.io/projects/miniconda/en/latest/index.html)
Clone this repository with
```bash
cd ~
git clone --recurse-submodules https://github.com/jacarvalho/mpd-public.git
cd mpd-public
```
Download [IsaacGym Preview 4](https://developer.nvidia.com/isaac-gym) and extract it under `deps/isaacgym`
```bash
mv ~/Downloads/IsaacGym_Preview_4_Package.tar.gz ~/mpd-public/deps/
cd ~/mpd-public/deps
tar -xvf IsaacGym_Preview_4_Package.tar.gz
```
Run the bash setup script to install everything.
```
cd ~/mpd-public
bash setup.sh
```
@ -40,9 +57,13 @@ bash setup.sh
To try out the MPD inference, first download the data and the trained models.
```bash
gdown 1mmJAFg6M2I1OozZcyueKp_AP0HHkCq2k
conda activate mpd
```
```bash
gdown --id 1mmJAFg6M2I1OozZcyueKp_AP0HHkCq2k
tar -xvf data_trajectories.tar.gz
gdown 1I66PJ5QudCqIZ2Xy4P8e-iRBA8-e2zO1
gdown --id 1I66PJ5QudCqIZ2Xy4P8e-iRBA8-e2zO1
tar -xvf data_trained_models.tar.gz
```
@ -73,6 +94,10 @@ The results will be saved under `data_trained_models/[model_id]/results_inferenc
We recommend running the follwowing in a SLURM cluster.
```bash
conda activate mpd
```
To regenerate the data:
```bash
cd scripts/generate_data

View File

@ -1,4 +1,4 @@
numpy==1.19.5
numpy
tensorboard
tqdm
ConfigArgParse

View File

@ -8,9 +8,6 @@ if [ ! -d $ISAACGYM_DIR ]; then
exit
fi
git submodule update --init --recursive --remote --progress
export SKLEARN_ALLOW_DEPRECATED_SKLEARN_PACKAGE_INSTALL=True
eval "$(~/miniconda3/bin/conda shell.bash hook)"
@ -49,4 +46,5 @@ cd ${ROOT_DIR} && pip install -e .
conda remove --force ncurses --yes
conda install -c "conda-forge/label/cf202003" gdown --yes
pip install --upgrade --no-cache-dir gdown