Pre-Release v0.6.0 (#39)

* Prepare release

* Update docker images
This commit is contained in:
Antonin RAFFIN 2020-06-01 13:09:47 +02:00 committed by GitHub
parent bb01253261
commit 403fff5d50
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 19 additions and 7 deletions

View file

@ -1,4 +1,4 @@
image: stablebaselines/stable-baselines3-cpu:0.6.0a7 image: stablebaselines/stable-baselines3-cpu:0.6.0
type-check: type-check:
script: script:
@ -15,5 +15,4 @@ doc-build:
lint-check: lint-check:
script: script:
- pip install flake8 # TODO: remove when new version on Pypi
- make lint - make lint

View file

@ -33,6 +33,7 @@ RUN \
cd ${CODE_DIR}/stable-baselines3 3&& \ cd ${CODE_DIR}/stable-baselines3 3&& \
pip install -e .[extra,tests,docs] && \ pip install -e .[extra,tests,docs] && \
# Use headless version for docker # Use headless version for docker
pip uninstall -y opencv-python && \
pip install opencv-python-headless && \ pip install opencv-python-headless && \
rm -rf $HOME/.cache/pip rm -rf $HOME/.cache/pip

View file

@ -5,7 +5,6 @@
**WARNING: Stable Baselines3 is currently in a beta version, breaking changes may occur before 1.0 is released** **WARNING: Stable Baselines3 is currently in a beta version, breaking changes may occur before 1.0 is released**
Note: most of the documentation of [Stable Baselines](https://github.com/hill-a/stable-baselines) should be still valid though.
# Stable Baselines3 # Stable Baselines3
@ -35,7 +34,6 @@ These algorithms will make it easier for the research community and industry to
| High code coverage | :heavy_check_mark: | | High code coverage | :heavy_check_mark: |
| Type hints | :heavy_check_mark: | | Type hints | :heavy_check_mark: |
<!-- | Tensorboard support | :heavy_check_mark: | -->
### Roadmap to V1.0 ### Roadmap to V1.0
@ -45,7 +43,6 @@ Planned features:
- [ ] DQN (almost ready, currently in testing phase) - [ ] DQN (almost ready, currently in testing phase)
- [ ] DDPG (you can use its successor TD3 for now) - [ ] DDPG (you can use its successor TD3 for now)
- [ ] HER - [ ] HER
- [ ] Support for MultiDiscrete and MultiBinary action spaces
### Planned features (v1.1+) ### Planned features (v1.1+)

View file

@ -3,9 +3,11 @@
Changelog Changelog
========== ==========
Pre-Release 0.6.0a11 (WIP) Pre-Release 0.6.0 (2020-06-01)
------------------------------ ------------------------------
**Tensorboard support, refactored logger**
Breaking Changes: Breaking Changes:
^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^
- Remove State-Dependent Exploration (SDE) support for ``TD3`` - Remove State-Dependent Exploration (SDE) support for ``TD3``
@ -47,6 +49,7 @@ Others:
- Added ``.readthedoc.yml`` file - Added ``.readthedoc.yml`` file
- Added ``flake8`` and ``make lint`` command - Added ``flake8`` and ``make lint`` command
- Added Github workflow - Added Github workflow
- Added warning when passing both ``train_freq`` and ``n_episodes_rollout`` to Off-Policy Algorithms
Documentation: Documentation:
^^^^^^^^^^^^^^ ^^^^^^^^^^^^^^

View file

@ -3,6 +3,7 @@ import os
import io import io
import zipfile import zipfile
import pickle import pickle
import warnings
from typing import Union, Type, Optional, Dict, Any, List, Tuple, Callable from typing import Union, Type, Optional, Dict, Any, List, Tuple, Callable
from abc import ABC, abstractmethod from abc import ABC, abstractmethod
from collections import deque from collections import deque
@ -800,6 +801,17 @@ class OffPolicyRLModel(BaseRLModel):
assert isinstance(env, VecEnv), "You must pass a VecEnv" assert isinstance(env, VecEnv), "You must pass a VecEnv"
assert env.num_envs == 1, "OffPolicyRLModel only support single environment" assert env.num_envs == 1, "OffPolicyRLModel only support single environment"
if n_episodes > 0 and n_steps > 0:
# Note we are refering to the constructor arguments
# that are named `train_freq` and `n_episodes_rollout`
# but correspond to `n_steps` and `n_episodes` here
warnings.warn("You passed a positive value for `train_freq` and `n_episodes_rollout`."
"Please make sure this is intended. "
"The agent will collect data by stepping in the environment "
"until both conditions are true: "
"`number of steps in the env` >= `train_freq` and "
"`number of episodes` > `n_episodes_rollout`")
if self.use_sde: if self.use_sde:
self.actor.reset_noise() self.actor.reset_noise()

View file

@ -1 +1 @@
0.6.0a11 0.6.0