Artur Niederfahrenhorst bccabb5ebf [RLlib] More fixes around PPO release tests. (#36045) 1 年之前
..
checkpointing_tests 83179ab1db [RLlib] Load state from load_state_path for rlmodule spec. (#35180) 1 年之前
learning_tests bccabb5ebf [RLlib] More fixes around PPO release tests. (#36045) 1 年之前
multi_gpu_learning_tests 7401b39720 [RLlib] Fix double '::' in RLlib release test yaml files. (#34865) 1 年之前
multi_gpu_with_attention_learning_tests 7401b39720 [RLlib] Fix double '::' in RLlib release test yaml files. (#34865) 1 年之前
multi_gpu_with_lstm_learning_tests 7401b39720 [RLlib] Fix double '::' in RLlib release test yaml files. (#34865) 1 年之前
smoke_tests 5073be7ad6 [RLlib] Don't add a cpu to bundle for learner when using gpu (#35529) 1 年之前
stress_tests 3d335e1b22 [CI][GCE][RLlib] Add GCE variations to RLlib release tests (#34080) 1 年之前
unit_gpu_tests 65d9a410f7 [ci] Clean up ci/ directory (refactor ci/travis) (#23866) 2 年之前
1gpu_16cpus.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
1gpu_16cpus_gce.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
1gpu_24cpus.yaml 66aae4cb32 [Release Test] Make sure to delete all EBS volumes (#28707) 2 年之前
1gpu_24cpus_gce.yaml 3d335e1b22 [CI][GCE][RLlib] Add GCE variations to RLlib release tests (#34080) 1 年之前
1gpu_32cpus.yaml 3d335e1b22 [CI][GCE][RLlib] Add GCE variations to RLlib release tests (#34080) 1 年之前
1gpu_32cpus_gce.yaml 3d335e1b22 [CI][GCE][RLlib] Add GCE variations to RLlib release tests (#34080) 1 年之前
1gpu_4cpus.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
1gpu_4cpus_gce.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
2gpus_32cpus.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
2gpus_32cpus_gce.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
2gpus_64cpus.yaml 3a181229cb Allow tf and torch apex release tests to run in parallel (#30882) 1 年之前
2gpus_64cpus_gce.yaml 3d335e1b22 [CI][GCE][RLlib] Add GCE variations to RLlib release tests (#34080) 1 年之前
32cpus.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
32cpus_gce.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
4gpus_512_cpus_gce.yaml 3d335e1b22 [CI][GCE][RLlib] Add GCE variations to RLlib release tests (#34080) 1 年之前
4gpus_544_cpus.yaml 66aae4cb32 [Release Test] Make sure to delete all EBS volumes (#28707) 2 年之前
4gpus_64cpus.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
4gpus_64cpus_gce.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
8gpus_96cpus.yaml 1a0989a1c0 [Release Test] Add GCE variation for core release tests [2/n] (#34337) 1 年之前
8gpus_96cpus_gce.yaml fb5cefd99a [RLlib] Try 8gpus_96cpus_gce with n1 and t4 nodes (#34459) 1 年之前
README.rst c9d220bcda [RLlib] Upgrade RLlib regression test scripts to new testing tool - RLlib release logs for 1.4. (#16080) 3 年之前
app_config.yaml 4995e14d35 [RLlib] Learner group checkpointing (#34379) 1 年之前
debug_app_config.yaml d50f528b4d [ci][rllib][core] Make rllib_multi_gpu_with_attention_learning_tests.gce run with debug wheels (#34897) 1 年之前
multi_node_checkpointing_compute_config.yaml 5073be7ad6 [RLlib] Don't add a cpu to bundle for learner when using gpu (#35529) 1 年之前
multi_node_checkpointing_compute_config_gce.yaml 5073be7ad6 [RLlib] Don't add a cpu to bundle for learner when using gpu (#35529) 1 年之前

README.rst

RLlib Tests
===========

This directory contains various RLlib release tests.

You should run these tests with the `releaser `_ tool.

Overview
--------
Currently, there are 3 RLlib tests:

1. ``learning_tests`` - Tests, whether major algos (tf+torch) can learn in Atari or PyBullet envs in ~30-60min.
1. ``stress_tests`` - Runs 4 IMPALA Atari jobs, each one using 1GPU and 128CPUs (needs autoscaling to succeed).
1. ``unit_gpu_tests`` - Tests, whether all of RLlib's example scripts can be run on a GPU.

Generally the releaser tool will run all tests in parallel.

Acceptance criteria
-------------------
These tests are considered passing when they throw no error at the end of the output log.