Giters
weidler
/
RLaSpa
Reinforcement Learning in Latent Space
Geek Repo:
Geek Repo
Github PK Tool:
Github PK Tool
Stargazers:
5
Watchers:
5
Issues:
34
Forks:
1
weidler/RLaSpa Issues
Make tensorboard logs get their own directories
Closed
5 years ago
Use 4 frames of Race/Evasion as one state
Updated
5 years ago
Add agent.save() and agent.load()
Closed
5 years ago
Double Check the Workings of all Networks
Closed
5 years ago
Check how DQN memory influence training when using ckpts, eventually save memory in ckpts as well.
Closed
5 years ago
Develop a better metric for the progress report
Closed
5 years ago
Comments count
5
Have rendered testruns automatically saved to a gif file
Closed
5 years ago
Make Race/Evasion tasks customizable (kwargs for the register)
Closed
5 years ago
Create template for experiments in Wiki
Closed
5 years ago
Prepare job script for Aachen cluster
Closed
5 years ago
Verify if GPU is indeed utilized
Closed
5 years ago
Exploration strategy
Closed
5 years ago
Allow the DQN Policy to backpropagate through the representation module
Closed
5 years ago
Include MultiTask Possibility into Framework
Closed
5 years ago
Create new Tasks
Closed
5 years ago
Comments count
1
Tensorize everything
Closed
5 years ago
Check if input is as expected and throw meaningful errors.
Updated
5 years ago
Comments count
1
Document Code, include type hinting!
Updated
5 years ago
Comments count
1
Look at the heads of the Janus and Cerberus architecture
Closed
5 years ago
Make tasks implement the gym interface
Closed
5 years ago
Comments count
1
Create statistics of training time and performance
Closed
5 years ago
Comments count
2
Being able to save and continue training a model
Closed
5 years ago
Comments count
4
Create different tasks/obstacle maps to "transfer"
Closed
5 years ago
Comments count
3
Batch Learning in Representation Learners
Closed
5 years ago
Modular Framework for different RL Approaches
Closed
5 years ago
Comments count
4
Make pathing task use numpy arrays for faster conversions
Closed
5 years ago
Comments count
3
Implement
Closed
5 years ago
Create DQN framework
Closed
5 years ago
Comments count
2
Learn ObstaclePathing in latent space
Closed
5 years ago
Implement SiameseAutoencoder
Closed
5 years ago
Comments count
1
Try to learn ObstaclePathing
Closed
5 years ago
Comments count
6
Test RL agent on (padded) two-task latent representation
Closed
5 years ago
qwdqwd
Closed
5 years ago
Try to learn Pathing in latent space
Closed
5 years ago