Our implementation of "Dream to Control: Learning Behaviors by Latent Imagination" in Pytorch

Key components

Debug error with fixed episode length that returns done and some phases done is not included into calculating the reward structure which can be weird

Name		Name	Last commit message	Last commit date
Latest commit History 94 Commits
__pycache__		__pycache__
.gitignore		.gitignore
.python-version		.python-version
Dreamer.py		Dreamer.py
RSSM.py		RSSM.py
Readme.md		Readme.md
ReplayBuffer.py		ReplayBuffer.py
config.yaml		config.yaml
conv_env_dec.py		conv_env_dec.py
decoded_observation_300.png		decoded_observation_300.png
decoded_observation_400.png		decoded_observation_400.png
decoded_observation_500.png		decoded_observation_500.png
decoded_observation_600.png		decoded_observation_600.png
dm.py		dm.py
main.py		main.py
pyproject.toml		pyproject.toml
requirements.txt		requirements.txt
ssh-remote+132.145.195.195:.inetloc		ssh-remote+132.145.195.195:.inetloc
training_video.mp4_1000.mp4		training_video.mp4_1000.mp4
training_video.mp4_1100.mp4		training_video.mp4_1100.mp4
training_video.mp4_1200.mp4		training_video.mp4_1200.mp4
training_video.mp4_1300.mp4		training_video.mp4_1300.mp4
training_video.mp4_1400.mp4		training_video.mp4_1400.mp4
training_video.mp4_300.mp4		training_video.mp4_300.mp4
training_video.mp4_400.mp4		training_video.mp4_400.mp4
training_video.mp4_500.mp4		training_video.mp4_500.mp4
training_video.mp4_600.mp4		training_video.mp4_600.mp4
training_video.mp4_700.mp4		training_video.mp4_700.mp4
training_video.mp4_800.mp4		training_video.mp4_800.mp4
training_video.mp4_900.mp4		training_video.mp4_900.mp4
uv.lock		uv.lock
value_functions.py		value_functions.py