Updating ARCC deepracer

.gitignore

@@ -1,6 +1,17 @@
-
-aws-deepracer-workshops/
-
-deepracer/
-
+.vscode/
+custom_files/
+logs/
 docker/volumes/
+recording/
+recording
+/*.env
+/*.bak
+/*.tar
+/*.json
+DONE
+data/
+tmp/
+autorun.s3url
+nohup.out
+/*.sh
+_

LICENSE

@@ -0,0 +1,14 @@
+Copyright 2019-2023 AWS DeepRacer Community. All Rights Reserved.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy of this
+software and associated documentation files (the "Software"), to deal in the Software
+without restriction, including without limitation the rights to use, copy, modify,
+merge, publish, distribute, sublicense, and/or sell copies of the Software, and to
+permit persons to whom the Software is furnished to do so.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
+INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
+PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
+HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

README.md

@@ -1,235 +1,70 @@
-# DeepRacer-For-Dummies
-Provides a quick and easy way to get up and running with a local deepracer training environment using Docker Compose.
-This repo just creates a wrapper around the amazing work done by Chris found here: https://github.com/crr0004/deepracer
-Please refer to his repo to understand more about what's going on under the covers.
+# DeepRacer-For-Cloud
+Provides a quick and easy way to get up and running with a DeepRacer training environment using a cloud virtual machine or a local compter, such [AWS EC2 Accelerated Computing instances](https://aws.amazon.com/ec2/instance-types/?nc1=h_ls#Accelerated_Computing) or the Azure [N-Series Virtual Machines](https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes-gpu).
 
-# Video Instructions
-[![Video Instructions](https://img.youtube.com/vi/CFNcKmtVRSI/0.jpg)](https://www.youtube.com/watch?v=CFNcKmtVRSI)
+DRfC runs on Ubuntu 22.04 and 24.04. GPU acceleration requires a NVIDIA GPU, preferrably with more than 8GB of VRAM.
 
-# Getting Started
+## Introduction
 
----
-#### Prerequisites
+DeepRacer-For-Cloud (DRfC) started as an extension of the work done by Alex (https://github.com/alexschultz/deepracer-for-dummies), which is again a wrapper around the amazing work done by Chris (https://github.com/crr0004/deepracer). With the introduction of the second generation Deepracer Console the repository has been split up. This repository contains the scripts needed to *run* the training, but depends on Docker Hub to provide pre-built docker images. All the under-the-hood building capabilities are in the [Deepracer Simapp](https://github.com/aws-deepracer-community/deepracer-simapp) repository.
 
-* This project is specifically built to run on Ubuntu 18.04 with an **Nvidia GPU**. It is assumed you already have **CUDA/CUDNN** installed and configured.
+As if December 2025 the original DeepRacer service in the AWS console is no longer available, and is replaced by [DeepRacer-on-AWS](https://aws.amazon.com/solutions/implementations/deepracer-on-aws/) which you can install in your own AWS environment. DeepRacer-For-Cloud is independent of any AWS service, so it is not directly impacted by this change.
 
-* You also need to have **Docker** installed as well as the **Nvidia-Docker** runtime.
+## Main Features
 
-* You should have an AWS account with the **AWS cli** installed. The credentials should be located in your home directory (~/.aws/credentials)
+DRfC supports a wide set of features to ensure that you can focus on creating the best model:
+* User-friendly
+	* Based on the continously updated community [Robomaker](https://github.com/aws-deepracer-community/deepracer-simapp) container, supporting a wide range of CPU and GPU setups.
+	* Wide set of scripts (`dr-*`) enables effortless training.
+	* Detection of your AWS DeepRacer Console models; allows upload of a locally trained model to any of them.
+* Modes
+	* Time Trial
+	* Object Avoidance
+	* Head-to-Bot
+* Training
+	* Multiple Robomaker instances per Sagemaker (N:1) to improve training progress.
+	* Multiple training sessions in parallel - each being (N:1) if hardware supports it - to test out things in parallel.
+	* Connect multiple nodes together (Swarm-mode only) to combine the powers of multiple computers/instances.
+* Evaluation
+	* Evaluate independently from training.
+	* Save evaluation run to MP4 file in S3.
+* Logging
+	* Training metrics and trace files are stored to S3.
+	* Optional integration with AWS CloudWatch.
+	* Optional exposure of Robomaker internal log-files.
+* Technology
+	* Supports both Docker Swarm (used for connecting multiple nodes together) and Docker Compose
 
-* ensure you have **vncviewer** installed
+## Tech Stack
 
-#### NOTE: If you already have these prerequisites setup then you can simply run the init.sh script described in the **Initialization** section. If you are setting everything up for the first time, then the information provided here can help you to get your environment ready to use this repo.
+DRfC is built on top of the [AWS DeepRacer Simapp](https://github.com/aws-deepracer-community/deepracer-simapp) — a single Docker image used for three purposes:
 
+* **Robomaker** — one or more containers providing robotics simulation via ROS and Gazebo
+* **Sagemaker** — container running the model training job
+* **RL Coach** — container that bootstraps the Sagemaker container using the Sagemaker SDK and Sagemaker Local
 
-#### Local Environment Setup
+### Core Technologies
 
-If you are running Windows and would like to use this repo, you will need to modify the process to get everything to run on Windows (not recommended as you will not be able to take advantage of the GPU during training) Many users have found it useful to dual-boot (Windows/Linux). There are many tutorials online for how to do this. You can follow the instructions provided below as guidance.
+| Component | Version |
+|-----------|---------|
+| Ubuntu | 24.04 |
+| Python | 3.12 |
+| TensorFlow | 2.20 |
+| CUDA | 12.6 (GPU only) |
+| Redis | 8.0.4 |
+| ROS | 2 Jazzy |
+| Gazebo | Harmonic |
 
-##### * Installing Ubuntu 18.04 with Windows 10
+### Images
 
-https://medium.com/bigdatarepublic/dual-boot-windows-and-linux-aa281c3c01f9
+Pre-built images are available on [Docker Hub](https://hub.docker.com/repository/docker/awsdeepracercommunity/deepracer-simapp) as `awsdeepracercommunity/deepracer-simapp:<VERSION>-cpu` (CPU) and `awsdeepracercommunity/deepracer-simapp:<VERSION>-gpu` (CUDA GPU). Both support OpenGL acceleration.
 
-When it gets to the Disk Management part, to make space for your Ubuntu installation, followed this guide and specifically look at the 2nd method (MiniTool Partition Wizard):
+During installation DRfC will automatically pull the latest image based on whether you have a GPU or CPU installation.
 
-https://win10faq.com/shrink-partition-windows-10/?source=post_page---------------------------
+## Documentation
 
-=======
-##### * Installing the AWS CLI
+Full documentation can be found on the [Deepracer-for-Cloud GitHub Pages](https://aws-deepracer-community.github.io/deepracer-for-cloud).
 
-	pip install -U awscli
-
-Then Follow this: https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-configure.html
+## Support
 
-##### * Installing Docker-ce (steps from https://docs.docker.com/install/linux/docker-ce/ubuntu/ )
-
-	sudo apt-get remove docker docker-engine docker.io containerd runc
-	sudo apt-get update
-
-	  sudo apt-get install \
-    apt-transport-https \
-    ca-certificates \
-    curl \
-    gnupg-agent \
-    software-properties-common
-
-    curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
-    sudo apt-key fingerprint 0EBFCD88
-
-    sudo add-apt-repository \
-    "deb [arch=amd64] https://download.docker.com/linux/ubuntu \
-    $(lsb_release -cs) \
-    stable"
-
-	  sudo apt-get update
-    sudo apt-get install docker-ce docker-ce-cli containerd.io
-
-Verify docker works
-
-	sudo docker run hello-world
-
-##### 3. Installing Docker-compose (from https://docs.docker.com/compose/install/#install-compose )
-
-    curl -L https://github.com/docker/compose/releases/download/1.24.1/docker-compose-`uname -s`-`uname -m` -o     /usr/local/bin/docker-compose
-    sudo chmod +x /usr/local/bin/docker-compose
-
-Verify installation
-
-    docker-compose --version
-
-###### NOTE: You can also choose to install docker-compose via another package manager (i.e. pip or conda), but if you do, make sure to do so in a virtual env.  Many OS’s have python system packages that conflict with docker-compose dependencies. ######
-
-Additionally, make sure your user-id can run docker without sudo (from https://docs.docker.com/install/linux/linux-postinstall/ )
-
-	sudo groupadd docker
-	sudo usermod -aG docker $USER
-
-Log out and log back in so that your group membership is re-evaluated.
-
-And configure Docker to start on boot.
-
-    sudo systemctl enable docker
-
-##### * Preparing for nvidia-docker
-
-The NVIDIA Container Toolkit allows users to build and run GPU accelerated Docker containers.  
-Nvidia-docker essentially exposes the GPU to the containers to use:  https://github.com/NVIDIA/nvidia-docker
-
-You may want to note what you have installed currently.
-
-        sudo apt list --installed | grep nvidia
-
-Then prepare for clean installation of Nvidia drivers.
-
-        sudo apt-get purge nvidia*  
-
-##### Installing nvidia-docker runtime (from https://github.com/NVIDIA/nvidia-docker/wiki/Installation-(version-2.0) )
-
-    distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
-	curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
-	curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
-	sudo apt-get update
-	sudo apt-get install nvidia-docker2
-	sudo pkill -SIGHUP dockerd
-
-##### * Installing the proper nvidia drivers
-
-Check for driver version here according to your GPU(s):  https://www.nvidia.com/Download/index.aspx?lang=en-us
-In the dropdown for OS, choose “show all OS’s” to see if there are Ubuntu specific choices.  Otherwise choose Linux.
-If you get a dropdown for “cuda toolkit”, choose 10.0)
-
-	sudo add-apt-repository ppa:graphics-drivers
-	sudo apt-get update
-	sudo apt install nvidia-driver-410 && sudo reboot 
-
-###### NOTE: 410 is a driver version that is compatible with the GPU I selected on the Nvidia website. ######
-
-Verify the driver installation:
-
-	nvidia-smi
-	nvcc --version  
-
-##### * Installing VNC viewer on your local machine
-
-This doc is straight forward: https://www.techspot.com/downloads/5760-vnc-viewer.html
-
-##### * Installing the Nvidia deep learning libraries (CUDA/CUDNN) for GPU hardware:
-
-This guide goes through how to install CUDA & CUDNN : https://medium.com/@zhanwenchen/install-cuda-and-cudnn-for-tensorflow-gpu-on-ubuntu-79306e4ac04e
-
-###### NOTE: You can apparently use Anaconda instead to install CUDA/CUDNN. I have not tried this, however some users have and have reported that this method is much easier. If you use this approach, you will need to first install Anaconda.  Once installed you can then use the conda package manager to install the desired versions of CUDA and cuDNN.  The following installation configuration has been reported to work together successfully ######
-
-##### Downloading Anaconda
-
-	sudo apt-get update -y && sudo apt-get upgrade -y
-	cd /tmp/
-	sudo wget https://repo.anaconda.com/archive/Anaconda3-2019.03-Linux-x86_64.sh
-
-##### Installing Anaconda
-
-	bash Anaconda3-2019.03-Linux-x86_64.sh
-	"yes" for using the default directory location
-	“yes” for running conda init
-
-##### Activating Anaconda  
-
-	source ~/.bashrc
-
-##### Verifying the conda package manager works
-
-    conda list
-
-##### Installing CUDA/CUDNN
-
-    conda install cudnn==7.3.1 && conda install -c fragcolor cuda10.0
-
-
-#### Initialization (After all prerequisites have been installed)
-
-
-##### 11. Run Init.sh from this repo (refer to the rest of this doc for script details) 
-
-In a command prompt, simply run "./init.sh".
-This will set everything up so you can run the deepracer local training environment.
-
-
-**init.sh** performs these steps so you don't have to do them manually:
-1. Clones Chris's repo:  https://github.com/crr0004/deepracer.git
-2. Does a mkdir -p ~/.sagemaker && cp config.yaml ~/.sagemaker
-3. Sets the image name in rl_deepracer_coach_robomaker.py  to "crr0004/sagemaker-rl-tensorflow:nvidia”
-4. Also sets the instance_type in rl_deepracer_coach_robomaker.py to “local_gpu”
-5. Copies the reward.py and model-metadata files into your Minio bucket
-
-
-To start or stop the local deepracer training, use the scripts found in the scripts directory.
-
-Here is a brief overview of the available scripts:
-
-#### Scripts
-
-* training
-	* start.sh
-		* starts the whole environment using docker compose
-		* it will also open a terminal window where you can monitor the log output from the sagemaker training directory
-		* it will also automatically open vncviewer so you can watch the training happening in Gazebo
-	* stop.sh
-		* stops the whole environment
-		* automatically finds and stops the training container which was started from the sagemaker container
-	* upload-snapshot.sh
-		* uploads a specific snapshot to S3 in AWS. If no checkpoint is provided, it attempts to retrieve the latest snapshot
-	* set-last-run-to-pretrained.sh
-		* renames the last training run directory from ***rl-deepracer-sagemaker*** to ***rl-deepracer-pretrained*** so that you can use it as a starting point for a new training run.
-	* delete-last-run.sh
-		* (WARNING: this script deletes files on your system. I take no responsibility for any resulting actions by running this script. Please look at what the script is doing before running it so that you understand)
-		* deletes the last training run including all of the snapshots and log files. You will need sudo to run this command.
-
-
-* evaluation
-	* start.sh
-		* starts the whole environment using docker compose to run an evaluation run
-		* it will also open a terminal window where you can monitor the log output from the sagemaker training directory
-		* it will also automatically open vncviewer so you can watch the training happening in Gazebo
-	* stop.sh
-		* stops the whole environment
-		* automatically finds and stops the training container which was started from the sagemaker container
-
-* log-analysis
-	* start.sh
-		* starts a container with Nvidia-Docker running jupyter labs with the log analysis notebooks which were originally provided by AWS and then extended by  Tomasz Ptak
-		* the logs from robomaker are automatically mounted in the container so you don't have to move any files around
-		* in order to get to the container, look at the log output from when it starts. You need to grab the URL including the token query parameter and then paste it into the brower at **localhost:8888**.
-	* stop.sh
-		* stops the log-analysis container
-
-
-#### Hyperparameters
-
-You can modify training hyperparameters from the file **rl_deepracer_coach_robomaker.py**.
-
-#### Action Space & Reward Function
-
-The action-space and reward function files are located in the **deepracer-for-dummies/docker/volumes/minio/bucket/custom_files** directory
-
-#### Track Selection
-
-The track selection is controled via an environment variable in the **.env** file located in the **deepracer-for-dummies/docker** directory
+* For general support it is suggested to join the [AWS DeepRacing Community](https://deepracing.io/). The Community Slack has a channel #dr-training-local where the community provides active support.
+* Create a GitHub issue if you find an actual code issue, or where updates to documentation would be required.