Merge branch 'develop' into end-to-end-planning

opendr-eu · Mar 22, 2022 · 9da581d · 9da581d
2 parents 4afedfc + 3a08d24
commit 9da581d
Show file tree

Hide file tree

Showing 79 changed files with 454 additions and 311 deletions.
diff --git a/.github/workflows/test_packages.yml b/.github/workflows/test_packages.yml
@@ -39,7 +39,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           # - perception/object_detection_3d
           # - control/mobile_manipulation
@@ -82,7 +87,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           - perception/object_detection_3d
           - control/mobile_manipulation

diff --git a/.github/workflows/tests_suite.yml b/.github/workflows/tests_suite.yml
@@ -71,7 +71,12 @@ jobs:
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
           - control/mobile_manipulation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - simulation/human_model_generation
           - perception/facial_expression_recognition
           - control/single_demo_grasp
@@ -169,7 +174,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           # - perception/object_detection_3d
           # - control/mobile_manipulation
@@ -195,7 +205,7 @@ jobs:
     - name: Test Wheel
       run: |
         export DISABLE_BCOLZ_AVX2=true
-        sudo apt -y install python3.8-venv libfreetype6-dev git build-essential cmake python3-dev wget libopenblas-dev libsndfile1 libboost-dev libeigen3-dev
+        sudo apt -y install python3.8-venv libfreetype6-dev git build-essential cmake python3-dev wget libopenblas-dev libsndfile1 libboost-dev python3-dev
         python3 -m venv venv
         source venv/bin/activate
         pip install wheel
@@ -233,7 +243,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           # - perception/object_detection_3d
           # - control/mobile_manipulation
@@ -303,7 +318,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           - perception/object_detection_3d
           - control/mobile_manipulation

diff --git a/.github/workflows/tests_suite_develop.yml b/.github/workflows/tests_suite_develop.yml
@@ -70,7 +70,12 @@ jobs:
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
           - control/mobile_manipulation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - simulation/human_model_generation
           - perception/facial_expression_recognition
           - control/single_demo_grasp
@@ -169,7 +174,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           # - perception/object_detection_3d
           # - control/mobile_manipulation
@@ -234,7 +244,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           # - perception/object_detection_3d
           # - control/mobile_manipulation
@@ -305,7 +320,12 @@ jobs:
           - perception/speech_recognition
           - perception/skeleton_based_action_recognition
           - perception/semantic_segmentation
-          - perception/object_detection_2d
+          - perception/object_detection_2d/centernet
+          - perception/object_detection_2d/detr
+          - perception/object_detection_2d/gem
+          - perception/object_detection_2d/ssd
+          - perception/object_detection_2d/yolov3
+          - perception/object_detection_2d/retinaface
           - perception/facial_expression_recognition
           - perception/object_detection_3d
           - control/mobile_manipulation

diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -9,7 +9,8 @@ Released on XX, XXth, 2022.
     - Added support for modular pip packages allowing tools to be installed separately ([#201](https://github.com/opendr-eu/opendr/pull/201)).
     - Simplified the installation process for pip by including the appropriate post-installation scripts ([#201](https://github.com/opendr-eu/opendr/pull/201)).
     - Improved the structure of the toolkit by moving `io` from `utils` to `engine.helper` ([#201](https://github.com/opendr-eu/opendr/pull/201)).
-    - Added support for `post-install` scripts and `opendr` dependencies in `.ini` files  ([#201](https://github.com/opendr-eu/opendr/pull/201)). 
+    - Added support for `post-install` scripts and `opendr` dependencies in `.ini` files  ([#201](https://github.com/opendr-eu/opendr/pull/201)).
+    - Updated toolkit to support CUDA 11.2 and improved GPU support ([#215](https://github.com/opendr-eu/opendr/pull/215)).
   - Bug Fixes:
     - Updated wheel building pipeline to include missing files and removed unnecessary dependencies ([#200](https://github.com/opendr-eu/opendr/pull/200)).
     - `panoptic_segmentation/efficient_ps`: updated dataset preparation scripts to create correct validation ground truth ([#221](https://github.com/opendr-eu/opendr/pull/221)).

diff --git a/Dockerfile-cuda b/Dockerfile-cuda
@@ -1,45 +1,17 @@
-FROM nvidia/cuda:11.0-base
+FROM nvidia/cuda:11.2.0-cudnn8-devel-ubuntu20.04
 
 # Install dependencies
 RUN apt-get update && \
     apt-get --yes install git sudo apt-utils
 RUN DEBIAN_FRONTEND="noninteractive" apt-get -y install tzdata
 
-# Install CUDA 10.2
-RUN apt-get --yes install gcc-8 g++-8 wget  && \
-    update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-8 100 --slave /usr/bin/g++ g++ /usr/bin/g++-8  && \
-    update-alternatives --auto gcc && \
-    apt-get --yes install zlib1g-dev libbz2-dev libreadline-dev libssl-dev libsqlite3-dev libffi-dev && \
-    wget https://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda_10.2.89_440.33.01_linux.run && \
-    wget http://people.cs.uchicago.edu/~kauffman/nvidia/cudnn/cudnn-10.2-linux-x64-v8.2.0.53.tgz
-
-RUN sudo apt-get --yes install libxml2 && \
-    sh cuda_10.2.89_440.33.01_linux.run --silent --toolkit --override  --librarypath=/usr/local/cuda-10.2 && \
-    ln -s /usr/local/cuda-10.2/ /usr/local/cuda && \
-    tar -xzvf cudnn-10.2-linux-x64-v8.2.0.53.tgz && \
-    cp cuda/include/cudnn.h /usr/local/cuda/include && \
-    cp cuda/lib64/libcudnn* /usr/local/cuda/lib64 && \
-    chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn* && \
-    bash -c 'echo "/usr/local/cuda-10.2/lib64" > /etc/ld.so.conf.d/nvidia.conf' && \
-    ln -sf /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.2.0 /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_adv_train.so.8 && \
-    ln -sf /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.2.0 /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8 && \
-    ln -sf /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.0.0 /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8 && \
-    ln -sf /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.2.0 /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8 && \
-    ln -sf /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.2.0 /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_ops_train.so.8 && \
-    ln -sf /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.2.0 /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8 && \
-    ln -sf /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn.so.8.2.0 /usr/local/cuda-10.2/targets/x86_64-linux/lib/libcudnn.so.8 && \
-    ldconfig
-
 # Add Tini
 ENV TINI_VERSION v0.19.0
 ADD https://github.com/krallin/tini/releases/download/${TINI_VERSION}/tini /tini
 RUN chmod +x /tini
 ENTRYPOINT ["/tini", "--"]
 
-# Avoid switching back to gcc9 when install build-essential
-RUN sudo apt-get --yes install build-essential && \
-    update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-8 100 --slave /usr/bin/g++ g++ /usr/bin/g++-8  && \
-    update-alternatives --auto gcc
+RUN sudo apt-get --yes install build-essential
 
 # Clone the repo and install the toolkit
 ENV OPENDR_DEVICE gpu

diff --git a/README.md b/README.md
@@ -37,7 +37,7 @@ OpenDR aims to develop, train, deploy and evaluate deep learning models that imp
 
 OpenDR can be installed in the following ways:
 1. By *cloning* this repository (CPU/GPU support)
-2. Using *pip* (CPU only)
+2. Using *pip* (CPU/GPU support only)
 3. Using *docker* (CPU/GPU support)
 
 You can find detailed installation instruction in the [documentation](docs/reference/installation.md).

diff --git a/bin/activate.sh b/bin/activate.sh
@@ -4,9 +4,4 @@ export PYTHONPATH=$OPENDR_HOME/src:$PYTHONPATH
 export PYTHON=python3
 export LD_LIBRARY_PATH=$OPENDR_HOME/lib:$LD_LIBRARY_PATH
 
-if [[ -z "${OPENDR_DEVICE}" ]]; then
-  echo "[INFO] Set available device to CPU. You can manually change this by running 'export OPENDR_DEVICE=gpu'."
-  export OPENDR_DEVICE=cpu
-fi
-
 source venv/bin/activate
diff --git a/bin/install.sh b/bin/install.sh
@@ -38,13 +38,19 @@ sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main
 # Build OpenDR
 make install_compilation_dependencies
 make install_runtime_dependencies
-make libopendr
 
 # If working on GPU install GPU dependencies as needed
 if [[ "${OPENDR_DEVICE}" == "gpu" ]]; then
-  echo "[INFO] Installing  mxnet-cu102==1.8.0. You can override this later if you are using a different CUDA version."
-  pip3 install mxnet-cu102==1.8.0
+  pip3 uninstall -y mxnet
+  pip3 uninstall -y torch
+  echo "[INFO] Replacing  mxnet-cu112==1.8.0post0 to enable CUDA acceleration."
+  pip3 install mxnet-cu112==1.8.0post0
+  echo "[INFO] Replacing torch==1.9.0+cu111 to enable CUDA acceleration."
+  pip3 install torch==1.9.0+cu111 torchvision==0.10.0+cu111 torchaudio==0.9.0 -f https://download.pytorch.org/whl/torch_stable.html
+  echo "[INFO] Reinstalling detectronv2."
+  pip3 install 'git+https://github.com/facebookresearch/detectron2.git'
 fi
 
-deactivate
+make libopendr
 
+deactivate
diff --git a/dependencies/dependencies.ini b/dependencies/dependencies.ini
@@ -1,7 +1,7 @@
 [runtime]
 # 'python' and 'python-dependencies' keys expect a value in the Python requirements file format
 #  https://pip.pypa.io/en/stable/reference/pip_install/#requirements-file-format
-python=torch==1.7.1
+python=torch==1.9.0
        wheel
 
 [device]

diff --git a/docs/reference/installation.md b/docs/reference/installation.md
@@ -36,9 +36,9 @@ Using dockerfiles is strongly advised (please see below), unless you know what y
 Please also make sure that you have enough RAM available for the installation (about 4GB of free RAM is needed for the full installation/compilation).
 
 
-You can set the inference/training device using the `OPENDR_DEVICE` variable.
+If you want to install GPU-related dependencies, then you can appropriately set the `OPENDR_DEVICE` variable.
 The toolkit defaults to using CPU.
-If you want to use GPU, please set this variable accordingly:
+Therefore, if you want to use GPU, please set this variable accordingly *before* running the installation script:
 ```bash
 export OPENDR_DEVICE=gpu
 ```
@@ -49,24 +49,20 @@ source ./bin/activate.sh
 ```
 Then, you are ready to use the toolkit!
 
+**NOTE:** `OPENDR_DEVICE` does not alter the inference/training device at *runtime*. 
+It only affects the dependency installation.
+You can use OpenDR API to change the inference device.
+
 You can also verify the installation by using the supplied Python and C unit tests:
 ```bash
 make unittest
 make ctests
 ```
 
-If you plan to use GPU-enabled functionalities, then you are advised to install [CUDA 10.2](https://developer.nvidia.com/cuda-10.2-download-archive).
-To do so, you can follow these steps:
-```bash
-sudo apt install gcc-8 g++-8 gcc-9 g++-9
-sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-8 8 --slave /usr/bin/g++ g++ /usr/bin/g++-8
-sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-9 9 --slave /usr/bin/g++ g++ /usr/bin/g++-9
-echo "Please switch to GCC 8"
-sudo update-alternatives --config gcc
-```
-Then, you can install CUDA, along CuDNN.
-You can also refer to this [dockerfile](https://github.com/opendr-eu/opendr/blob/master/Dockerfile-cuda) for installation instructions.
-Note that NVIDIA 30xx GPUs may not be fully supported, due to CUDA limitations.
+If you plan to use GPU-enabled functionalities, then you are advised to install [CUDA 11.2](https://developer.nvidia.com/cuda-11.2.0-download-archive), along with [CuDNN](https://developer.nvidia.com/cudnn).
+
+**HINT:** All tests probe for the `TEST_DEVICE` enviromental variable when running.
+If this enviromental variable is set during testing, it allows for easily running all tests on a different device (e.g., setting `TEST_DEVICE=cuda:0` runs all tests on the first GPU of the system).
 
 # Installing using *pip*
 
@@ -95,14 +91,15 @@ This is not needed for newer CPUs.
 The same OpenDR package is used for both CPU and GPU systems. 
 However, you need to have the appropriate GPU-enabled dependencies installed to use a GPU with OpenDR.
 If you plan to use GPU, then you should first install [mxnet-cuda](https://mxnet.apache.org/versions/1.4.1/install/index.html?platform=Linux&language=Python&processor=CPU) and [detectron2](https://detectron2.readthedocs.io/en/latest/tutorials/install.html).
-For example, if you stick with the default PyTorch version (1.7) and use CUDA10.2, then you can simply follow:
+For example, if you stick with the default PyTorch version (1.8) and use CUDA11.2, then you can simply follow:
 ```bash
 sudo apt install python3.8-venv libfreetype6-dev git build-essential cmake python3-dev wget libopenblas-dev libsndfile1 libboost-dev libeigen3-dev 
 python3 -m venv venv
 source venv/bin/activate
 pip install wheel
-pip install detectron2==0.5 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html
-pip install mxnet-cu102==1.8.0 
+pip install torch==1.9.0+cu111 torchvision==0.10.0+cu111 torchaudio==0.9.0 -f https://download.pytorch.org/whl/torch_stable.html
+pip install 'git+https://github.com/facebookresearch/detectron2.git'
+pip install mxnet-cu112==1.8.0post0
 pip install opendr-toolkit-engine
 pip install opendr-toolkit
 ```

diff --git a/docs/reference/object-detection-2d-yolov3.md b/docs/reference/object-detection-2d-yolov3.md
@@ -5,7 +5,7 @@ The *yolov3* module contains the *YOLOv3DetectorLearner* class, which inherits f
 ### Class YOLOv3DetectorLearner
 Bases: `engine.learners.Learner`
 
-The *YOLOv3DetectorLearner* class is a wrapper of the SSD detector[[1]](#yolo-1)
+The *YOLOv3DetectorLearner* class is a wrapper of the YOLO detector[[1]](#yolo-1)
 [GluonCV implementation](https://github.com/dmlc/gluon-cv/blob/master/gluoncv/model_zoo/yolo/yolo3.py).
 It can be used to perform object detection on images (inference) as well as train new object detection models.
 

diff --git a/src/opendr/control/mobile_manipulation/dependencies.ini b/src/opendr/control/mobile_manipulation/dependencies.ini
@@ -10,18 +10,16 @@ python=vcstool
 [runtime]
 # 'python' key expects a value using the Python requirements file format
 #  https://pip.pypa.io/en/stable/reference/pip_install/#requirements-file-format
-python=torch==1.7.1
+python=torch==1.9.0
        tensorboard
        numpy
        pyyaml
        matplotlib
        pyparsing
        pillow>=8.3.2
        scipy
-       stable-baselines3>=0.10.0
-       gym>=0.19.0
+       stable-baselines3==1.1.0
        cloudpickle>=1.5.0
-       defusedxml
        netifaces
 
 opendr=opendr-toolkit-engine
diff --git a/src/opendr/control/single_demo_grasp/Makefile b/src/opendr/control/single_demo_grasp/Makefile
@@ -16,12 +16,6 @@ ifeq ($(MAKECMDGOALS),)
 MAKECMDGOALS = release
 endif
 
-ifeq ($(OPENDR_DEVICE),gpu)
-DETECTRON_WHEEL=https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.7/index.html
-else
-DETECTRON_WHEEL=https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html
-endif
-
 .PHONY: release install_compilation_dependencies install_runtime_dependencies
 
 release: install_compilation_dependencies
@@ -31,7 +25,7 @@ install_runtime_dependencies:
 
 install_compilation_dependencies:
 	@+echo "#"; echo "# * Install Compilation Dependencies for single demonstration grasping *"; echo "#"
-	@+python3 -m pip install detectron2==0.5 -f \$(DETECTRON_WHEEL)
+	@+python3 -m pip install 'git+https://github.com/facebookresearch/detectron2.git'
 	@./install_single_demo_grasp.sh
 
 help:

diff --git a/src/opendr/control/single_demo_grasp/dependencies.ini b/src/opendr/control/single_demo_grasp/dependencies.ini
@@ -1,12 +1,16 @@
+[compilation]
+python=torch==1.9.0
+       torchvision==0.10.0
+
 [runtime]
 # 'python' key expects a value using the Python requirements file format
 #  https://pip.pypa.io/en/stable/reference/pip_install/#requirements-file-format
-python=torch==1.7.1
-       torchvision==0.8.2
+python=torch==1.9.0
+       torchvision==0.10.0
        matplotlib>=2.2.2
        imgaug==0.4.0
        pillow>=8.3.2
 
 opendr=opendr-toolkit-engine
 
-post-install=python3 -m pip install detectron2==0.5 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cpu/torch1.7/index.html
+post-install=python3 -m pip install 'git+https://github.com/facebookresearch/detectron2.git'