Release v1.x

New issue

Open

No due date

•Last updated Aug 3, 2023

In this milestone we track everything that is not part of v1.0 but should not introduce any breaking changes.

11% complete

Humanoid Benchmark
HumanCompatibleAI/imitation#678 · opened on Feb 9, 2023 · Release v1.x
1
Benchmark Regression Test Suite
HumanCompatibleAI/imitation#647 · opened on Dec 28, 2022 · Release v1.x
Add support to warm-start GAIL and AIRL using behavioral cloning
HumanCompatibleAI/imitation#587 · opened on Oct 23, 2022 · Release v1.x
Add support for saving videos of policies on a environment for evaluation during and after training
HumanCompatibleAI/imitation#523 · opened on Aug 11, 2022 · Release v1.x
2
Use data acquired by users
HumanCompatibleAI/imitation#701 · opened on Apr 25, 2023 · Release v1.x
10
Support human preferences in “Deep RL from human preferences” (RLHP) implementation
HumanCompatibleAI/imitation#696 · opened on Mar 29, 2023 · Release v1.x
4
A suggestion for ”SimpleDAggerTrainer“ function
HumanCompatibleAI/imitation#693 · opened on Mar 23, 2023 · Release v1.x
Support for Dictionary Observation Space
HumanCompatibleAI/imitation#681 · opened on Feb 14, 2023 · Release v1.x
7
Would like to be able to run eval_policy for atari environments
HumanCompatibleAI/imitation#679 · opened on Feb 10, 2023 · Release v1.x
1
Inconsistent naming of expert demonstrations
HumanCompatibleAI/imitation#408 · opened on Feb 17, 2022 · Release v1.x
4
Embed code examples in docs
HumanCompatibleAI/imitation#606 · opened on Nov 3, 2022 · Release v1.x
3
save videos during training
Unread
#597 opened Oct 28, 2022 · Updated Jul 12, 2026
55
Support asynchronous human preference gathering in RLHP implementation
Unread
#716 opened May 15, 2023 · Updated May 28, 2024
13
DAgger Refactoring
HumanCompatibleAI/imitation#725 · opened on Jun 1, 2023 · Release v1.x
Refactor trajectory handling
HumanCompatibleAI/imitation#758 · opened on Jul 19, 2023 · Release v1.x
Consistent interface for algorithms & single training script
HumanCompatibleAI/imitation#79 · opened on Aug 14, 2019 · Release v1.x
17

Navigation Menu

Release v1.x

Humanoid Benchmark

Benchmark Regression Test Suite

Add support to warm-start GAIL and AIRL using behavioral cloning

Add support for saving videos of policies on a environment for evaluation during and after training

Use data acquired by users

Support human preferences in “Deep RL from human preferences” (RLHP) implementation

A suggestion for ”SimpleDAggerTrainer“ function

Support for Dictionary Observation Space

Would like to be able to run `eval_policy` for atari environments

Inconsistent naming of expert demonstrations

Embed code examples in docs

save videos during training

Support asynchronous human preference gathering in RLHP implementation

DAgger Refactoring

Refactor trajectory handling

Consistent interface for algorithms & single training script

Issues