## What do these changes do?
1. Separate the log related code to logger.py from services.py.
2. Allow users to modify logging formatter in `ray start`.
## Related issue number
https://github.com/ray-project/ray/pull/2664
* add noisy network
* distributional q-learning in dev
* add distributional q-learning
* validated rainbow module
* add some comments
* supply some comments
* remove redundant argument to pass CI test
* async replay optimizer does NOT need annealing beta
* ignore rainbow specific arguments for DDPG and Apex
* formatted by yapf
* Update dqn_policy_graph.py
* Update dqn_policy_graph.py
* added ars
* functioning ars with regression test
* added regression tests for ARs
* fixed default config for ARS
* ARS code runs, now time to test
* ARS working and tested, changed std deviation of meanstd filter to initialize to 1
* ARS working and tested, changed std deviation of meanstd filter to initialize to 1
* pep8 fixes
* removed unused linear model
* address comments
* more fixing comments
* post yapf
* fixed support failure
* Update LICENSE
* Update policies.py
* Update test_supported_spaces.py
* Update policies.py
* Update LICENSE
* Update test_supported_spaces.py
* Update policies.py
* Update policies.py
* Update filter.py
* + Compatibility fix under py2 on ray.tune
* + Revert changes on master branch
* + Use default JsonEncoder in ray.tune.logger
* + Add UT for infinity support
A bunch of minor rllib fixes:
pull in latest baselines atari wrapper changes (and use deepmind wrapper by default)
move reward clipping to policy evaluator
add a2c variant of a3c
reduce vision network fc layer size to 256 units
switch to 84x84 images
doc tweaks
print timesteps in tune status
This PR makes it so that when Ray is started via ray.init() (as opposed to via ray start) the Redis servers will be started in "protected mode" (which means that clients can only connect by connecting to localhost).
In practice, we actually connect redis clients by passing in the node IP address (not localhost), so I need to create a redis config file on the fly to allow both localhost and the node's actual IP address (it would have been nice to find a way to do this from the Python redis client, but I couldn't find one).
This adds some experimental (undocumented) support for launching Ray on existing nodes. You have to provide the head ip, and the list of worker ips.
There are also a couple additional utils added for rsyncing files and port-forward.
This PR introduces the following changes:
* Ray Tune -> Tune
* [breaking] Creation of `schedulers/`, moving PBT, HyperBand into a submodule
* [breaking] Search Algorithms now must take in experiment configurations via `add_configurations` rather through initialization
* Support `"run": (function | class | str)` with automatic registering of trainable
* Documentation Changes
The goal of this PR is to allow custom policies to perform model-based rollouts. In the multi-agent setting, this requires access to not only policies of other agents, but also their current observations.
Also, you might want to return the model-based trajectories as part of the rollout for efficiency.
compute_actions() now takes a new keyword arg episodes
pull out internal episode class into a top-level file
add function to return extra trajectories from an episode that will be appended to the sample batch
documentation
ray exec CLUSTER CMD [--screen] [--start] [--stop]
ray attach CLUSTER [--start]
Example:
ray exec sgd.yaml 'source activate tensorflow_p27 && cd ~/ray/python/ray/rllib && ./train.py --run=PPO --env=CartPole-v0' --screen --start --stop
This will in one command create a cluster and run the command on it in a screen session. The screen can later be attached to via ray attach. After the command finishes, the cluster workers will be terminated and the head node stopped.
to support TF version < 1.5
to support rmsprop optimizer in Impala
Before TF1.5, tf.reduce_sum() and tf.reduce_max() has an argument keep_dims which has been renamed as keepdims in later versions.
In the original paper of Impala, they use rmsprop algorithm to optimize the model. We'd better also support it so that users can reproduce their experiments. Without any tuning, say that using the same hyper-parameters as AdamOptimizer, it reaches "episode_reward_mean": 19.083333333333332 in Pong after consume 3,610,350 samples.
This PR adds a driver table for the new GCS, which enables cleanup functionality associated with monitoring driver death.
Some testing in `monitor_test.py` is restored, but redis sharding for xray is needed to enable remaining tests.
Rename AsyncSamplesOptimizer -> AsyncReplayOptimizer
Add AsyncSamplesOptimizer that implements the IMPALA architecture
integrate V-trace with a3c policy graph
audit V-trace integration
benchmark compare vs A3C and with V-trace on/off
PongNoFrameskip-v4 on IMPALA scaling from 16 to 128 workers, solving Pong in <10 min. For reference, solving this env takes ~40 minutes for Ape-X and several hours for A3C.
This also removes the async resetting code in VectorEnv. While that improves benchmark performance slightly, it substantially complicates env configuration and probably isn't worth it for most envs.
This makes it easy to efficiently support setups like Joint PPO: https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/retro-contest/gotta_learn_fast_report.pdf
For example, for 188 envs, you could do something like num_envs: 10, num_envs_per_worker: 19.
The dict merge prevents crashes when tune is trying to get resource requests for agents and you override a config subkey. The min iter time prevents iterations from getting too small, incurring high overhead. This is easy to run into on Ape-X since throughput can get very high.
We should use episode ids instead of the timestep to determine when sequences should be cut, since when batches are concatenated, increasing t does not guarantee we are part of the same episode.
