- Prerequisites
- Pulling the image
- Building the image
- Running the docker container
- Attach to the container
- Clone the implemetation of the World Models concept
- Running AppliedDataSciencePartners - WorldModels
- Running hardmaru - WordlModels
This specific image needs a CUDA capable GPU with CUDA version 10.0 and nvidia-docker version 2 installed on the host machine
-
mkdir ffabi_shared_folder
-
cd ffabi_shared_folder
-
git clone https://github.com/ffabi/SemesterProject.git
-
cd SemesterProject/docker_setup
-
docker build -f Dockerfile -t ffabi/gym:10 .
-
cd ../../..
Or pull from Dockerhub:
docker pull ffabi/gym:10
-
nvidia-docker create -p 8192:8192 -p 8193:22 -p 8194:8194 --name ffabi_gym -v $(pwd)/ffabi_shared_folder:/root/ffabi_shared_folder ffabi/gym:10
-
nvidia-docker start ffabi_gym
-
docker exec -it ffabi_gym bash
-
cd ffabi_shared_folder/SemesterProject/applied_worldmodel
-
mkdir data log
I recommend to use screen:
-
screen -R train
to enter screen 'train' (press Ctrl+A and Ctrl+D to exit) -
xvfb-run -a python3 01_generate_random_data.py
-
xvfb-run -a python3 01_generate_random_data.py --validation
-
xvfb-run -a python3 02_train_vae.py --num_files 10
-
xvfb-run -a python3 03_generate_rnn_data.py --num_files 10
-
xvfb-run -a python3 03_generate_rnn_data.py --num_files 10 --validation
-
xvfb-run -a python3 04_train_rnn.py --start_batch 0 --max_batch 0 --new_model
-
xvfb-run -a python3 05_train_controller.py car_racing --num_worker 1 --num_worker_trial 2 --num_episode 4 --max_length 1000 --eval_steps 25
python3 model.py car_racing --filename controller.json --render_mode