cpnota/autonomous-learning-library
A PyTorch library for building deep reinforcement learning agents.
PythonMIT
Issues
- 1
Cleanup Parallel Agent API
#209 opened - 0
Refactor handling of infos in state
#208 opened - 1
- 1
Preset import structure
#199 opened - 0
Bodies Documentation
#198 opened - 0
Frames argument issue
#194 opened - 1
Support newer versions of PyTorch
#188 opened - 0
torch-testing not maintained
#180 opened - 2
Unused variables in atari_wrappers.py
#178 opened - 1
Imports in scripts
#177 opened - 3
[Question] Change output path
#174 opened - 1
Fix CI
#164 opened - 2
Make run_experiment more modular
#162 opened - 1
Saving agents
#161 opened - 2
GymEnvironment name is equal to environment
#159 opened - 1
Problems with Library Install
#158 opened - 1
Test every nth episode
#157 opened - 1
[VAC] No reset after episode end?
#156 opened - 3
GaussianPolicy uses tanh two times
#154 opened - 1
Motivation behind `nn.Linear0`
#153 opened - 1
Should env.step() return done and info?
#149 opened - 1
eval() or act()
#148 opened - 1
Is close currently used?
#147 opened - 1
- 5
CUDA: out of memory
#144 opened - 1
The design choices regarding
#141 opened - 2
- 3
Documentation on StochasticPolicy
#136 opened - 1
Command line scripts with pip install
#129 opened - 2
Feature request: multiprocess env runner
#126 opened - 1
Device attribute unused
#125 opened - 1
Hyperparameter Tuning Module
#85 opened - 1
Explicit usage of target network
#84 opened - 1
Evaluation Metrics
#83 opened - 1
Parameter scheduling
#81 opened - 1
Save and Load Models
#78 opened - 1
Refactor approximation module
#71 opened - 0
Simplify Experiment API
#70 opened - 0
Deep Deterministic Policy Gradient
#69 opened - 1
Soft-Actor Critic
#67 opened - 1
make `layer` module an extension of `nn`
#65 opened - 0
Make writer level togglable
#62 opened - 0
Improve feature handling
#61 opened - 0
`State` constructor should properly handle `done` vs `mask`, and should have a __len__ property
#58 opened - 0
vanilla series agents
#57 opened - 1
- 1
Add `scripts` folder
#52 opened - 0
Learning Curve
#18 opened - 0
Make package structure flatter
#16 opened - 0
README
#13 opened