issues
search
lifelong-learning-systems
/
rlblocks
Reinforcement Learning Blocks for Researchers
MIT License
0
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
DqnScp agent doesn't reset `self.epsilon` like other Dqn agents
#51
coreylowman
opened
2 years ago
0
Adding example dqn in cartpole
#50
coreylowman
closed
2 years ago
0
Add docs
#49
ginoperrotta
closed
2 years ago
0
Experiment readme
#48
coreylowman
closed
2 years ago
2
Add documentation & plots showing agent behavior in minigrid_crossing_experiment
#47
coreylowman
closed
2 years ago
5
#29 Citing EWC & SCP
#46
coreylowman
closed
2 years ago
0
Fixing shape bugs in PPO blocks, adding example/docstrings
#45
coreylowman
closed
2 years ago
0
Store computations inside of the blocks as attributes on the object
#44
coreylowman
opened
2 years ago
0
Distribution.log_prob expects actions in the shape of Distribution.batch_shape
#43
coreylowman
closed
2 years ago
0
DqnCoverage agent should have both episodic memory buffer & short term buffer
#42
coreylowman
opened
2 years ago
2
Request for comments: #38 Adding new replay buffer
#41
coreylowman
closed
2 years ago
12
Resolves #37. Rename states to observations
#40
ginoperrotta
closed
2 years ago
0
Buffer re-prioritization method
#39
ginoperrotta
closed
2 years ago
0
Dataset redesign
#38
coreylowman
closed
2 years ago
11
Transition components named differently compared to tella
#37
ginoperrotta
closed
2 years ago
5
Resolves #30. Iterate in DQN experiment for shorter code
#36
ginoperrotta
closed
2 years ago
0
Coverage sampling
#35
ginoperrotta
closed
2 years ago
0
Should SCP aggregate synaptic matrices across batches with sum or mean?
#34
coreylowman
opened
2 years ago
0
switches to smaller logo image
#33
cash
closed
2 years ago
0
Add buffer priority function for state space coverage.
#32
ginoperrotta
closed
2 years ago
1
correctly storing synaptic matrix
#31
neilfendley
closed
2 years ago
3
Make minigrid_cross_experiment train function take the agent class and iterate over agent classes in main
#30
coreylowman
closed
2 years ago
0
Add citations in comments on SCP, EWC modules
#29
coreylowman
closed
2 years ago
0
SCP doesn't aggregate matrices across batches in DqnScp
#28
coreylowman
closed
2 years ago
2
Add DQN agent with reservoir sampled memory
#27
ginoperrotta
closed
2 years ago
2
Task memory agent performs differently on first block
#26
ginoperrotta
opened
2 years ago
4
Seed DQN rng for network init and for SCP computation.
#25
ginoperrotta
closed
2 years ago
0
adding scp loss to torchblocks
#24
neilfendley
closed
2 years ago
3
Current config used for DQN experiment
#23
ginoperrotta
closed
2 years ago
0
Uncomment all of the methods in the experiment.main
#22
coreylowman
closed
2 years ago
0
Cleanup
#21
coreylowman
closed
2 years ago
5
Fix random start loc and overhead view for other grid sizes
#20
ginoperrotta
closed
2 years ago
0
Fix dqn with task memory crash on first eval.
#19
ginoperrotta
closed
2 years ago
0
WIP changes to agent and curriculum to get DQN to learn minigrid
#18
ginoperrotta
closed
2 years ago
3
Adding two task minigrid
#17
coreylowman
closed
2 years ago
0
#8 Adding dqn with task specific buffers
#16
coreylowman
closed
2 years ago
0
Fixes #13
#15
coreylowman
closed
2 years ago
0
Add minigrid curriculum using custom unlock and door key
#14
coreylowman
closed
2 years ago
0
eval block contributes to epsilon greedy counter
#13
coreylowman
closed
2 years ago
0
Add tella DQN
#12
ginoperrotta
closed
2 years ago
3
Patch EWC loss call to use all anchors
#11
ginoperrotta
closed
2 years ago
0
Add example agent using optuna
#10
coreylowman
opened
2 years ago
0
Implement experiment with DQN and DQN with EWC
#9
cash
closed
2 years ago
1
Implement comparison between DQN and DQN with replay
#8
cash
closed
2 years ago
3
Make multiple cartpole variant curriculum
#7
cash
closed
2 years ago
1
Implement DQN as tella agent
#6
cash
closed
2 years ago
0
Figure out network sizes for minigrid tasks
#5
cash
closed
2 years ago
0
Add parameter tuning documentation
#4
cash
opened
2 years ago
0
Adding cartpole curriculums
#3
coreylowman
closed
2 years ago
0
Draft: datasets & samplers api
#2
coreylowman
closed
2 years ago
2
Next