-
Notifications
You must be signed in to change notification settings - Fork 327
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[BugFix] Refactor reductions #1968
Merged
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/1968
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 2 Unrelated FailuresAs of commit 00fffab with merge base b8ad113 (): NEW FAILURE - The following job has failed:
BROKEN TRUNK - The following jobs failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
label
Feb 27, 2024
vmoens
changed the title
[Feature] Use non-default mp start method in ParallelEnv
[[BugFix] Refactor reductions
Feb 27, 2024
|
Name | Max | Mean | Ops | Ops on Repo HEAD
|
Change |
---|---|---|---|---|---|
test_single | 61.3571ms | 60.9329ms | 16.4115 Ops/s | 16.4994 Ops/s | |
test_sync | 34.5857ms | 32.9518ms | 30.3474 Ops/s | 29.4725 Ops/s | |
test_async | 60.7826ms | 29.9556ms | 33.3827 Ops/s | 31.6680 Ops/s | |
test_simple | 0.4847s | 0.4336s | 2.3063 Ops/s | 2.3167 Ops/s | |
test_transformed | 0.6246s | 0.5781s | 1.7297 Ops/s | 1.7429 Ops/s | |
test_serial | 1.4471s | 1.4028s | 0.7128 Ops/s | 0.7085 Ops/s | |
test_parallel | 1.4347s | 1.3894s | 0.7197 Ops/s | 0.7142 Ops/s | |
test_step_mdp_speed[True-True-True-True-True] | 0.1391ms | 21.6034μs | 46.2891 KOps/s | 47.3545 KOps/s | |
test_step_mdp_speed[True-True-True-True-False] | 35.4670μs | 13.2123μs | 75.6870 KOps/s | 77.1513 KOps/s | |
test_step_mdp_speed[True-True-True-False-True] | 34.3740μs | 12.6774μs | 78.8808 KOps/s | 80.7549 KOps/s | |
test_step_mdp_speed[True-True-True-False-False] | 31.6100μs | 7.7510μs | 129.0150 KOps/s | 133.2336 KOps/s | |
test_step_mdp_speed[True-True-False-True-True] | 74.4490μs | 23.2589μs | 42.9942 KOps/s | 44.9120 KOps/s | |
test_step_mdp_speed[True-True-False-True-False] | 48.0500μs | 14.6073μs | 68.4589 KOps/s | 71.2604 KOps/s | |
test_step_mdp_speed[True-True-False-False-True] | 38.5220μs | 14.0554μs | 71.1470 KOps/s | 74.4642 KOps/s | |
test_step_mdp_speed[True-True-False-False-False] | 28.7630μs | 9.0756μs | 110.1861 KOps/s | 114.5829 KOps/s | |
test_step_mdp_speed[True-False-True-True-True] | 57.5080μs | 24.4921μs | 40.8295 KOps/s | 42.1819 KOps/s | |
test_step_mdp_speed[True-False-True-True-False] | 59.8920μs | 15.8802μs | 62.9714 KOps/s | 64.5789 KOps/s | |
test_step_mdp_speed[True-False-True-False-True] | 35.9970μs | 13.9238μs | 71.8195 KOps/s | 73.7901 KOps/s | |
test_step_mdp_speed[True-False-True-False-False] | 29.1150μs | 9.0257μs | 110.7944 KOps/s | 114.6165 KOps/s | |
test_step_mdp_speed[True-False-False-True-True] | 65.7630μs | 25.4941μs | 39.2247 KOps/s | 40.1711 KOps/s | |
test_step_mdp_speed[True-False-False-True-False] | 48.6510μs | 17.1724μs | 58.2330 KOps/s | 60.4765 KOps/s | |
test_step_mdp_speed[True-False-False-False-True] | 36.0670μs | 14.9552μs | 66.8662 KOps/s | 67.9402 KOps/s | |
test_step_mdp_speed[True-False-False-False-False] | 35.7970μs | 10.1945μs | 98.0924 KOps/s | 101.8102 KOps/s | |
test_step_mdp_speed[False-True-True-True-True] | 54.3820μs | 24.1704μs | 41.3729 KOps/s | 42.3894 KOps/s | |
test_step_mdp_speed[False-True-True-True-False] | 39.8950μs | 15.8477μs | 63.1006 KOps/s | 65.0700 KOps/s | |
test_step_mdp_speed[False-True-True-False-True] | 55.1430μs | 15.9997μs | 62.5013 KOps/s | 62.9643 KOps/s | |
test_step_mdp_speed[False-True-True-False-False] | 32.8320μs | 10.2493μs | 97.5677 KOps/s | 100.6280 KOps/s | |
test_step_mdp_speed[False-True-False-True-True] | 36.0270μs | 25.5931μs | 39.0730 KOps/s | 39.2061 KOps/s | |
test_step_mdp_speed[False-True-False-True-False] | 45.2650μs | 17.1333μs | 58.3657 KOps/s | 59.8954 KOps/s | |
test_step_mdp_speed[False-True-False-False-True] | 46.4270μs | 17.3287μs | 57.7077 KOps/s | 58.6944 KOps/s | |
test_step_mdp_speed[False-True-False-False-False] | 44.1130μs | 11.3878μs | 87.8135 KOps/s | 90.2582 KOps/s | |
test_step_mdp_speed[False-False-True-True-True] | 52.2480μs | 26.8012μs | 37.3118 KOps/s | 38.3473 KOps/s | |
test_step_mdp_speed[False-False-True-True-False] | 45.8060μs | 18.2527μs | 54.7865 KOps/s | 55.8271 KOps/s | |
test_step_mdp_speed[False-False-True-False-True] | 38.6420μs | 17.2924μs | 57.8288 KOps/s | 58.6546 KOps/s | |
test_step_mdp_speed[False-False-True-False-False] | 31.4490μs | 11.3347μs | 88.2244 KOps/s | 89.5033 KOps/s | |
test_step_mdp_speed[False-False-False-True-True] | 68.5780μs | 27.6848μs | 36.1209 KOps/s | 36.9673 KOps/s | |
test_step_mdp_speed[False-False-False-True-False] | 47.4580μs | 19.5299μs | 51.2035 KOps/s | 52.7571 KOps/s | |
test_step_mdp_speed[False-False-False-False-True] | 45.2350μs | 18.2483μs | 54.7997 KOps/s | 55.9557 KOps/s | |
test_step_mdp_speed[False-False-False-False-False] | 35.6270μs | 12.5517μs | 79.6703 KOps/s | 82.3678 KOps/s | |
test_values[generalized_advantage_estimate-True-True] | 12.3558ms | 9.4353ms | 105.9844 Ops/s | 108.6433 Ops/s | |
test_values[vec_generalized_advantage_estimate-True-True] | 36.6919ms | 33.1999ms | 30.1206 Ops/s | 28.7498 Ops/s | |
test_values[td0_return_estimate-False-False] | 0.2258ms | 0.1652ms | 6.0537 KOps/s | 6.0766 KOps/s | |
test_values[td1_return_estimate-False-False] | 30.7700ms | 23.5686ms | 42.4293 Ops/s | 43.3936 Ops/s | |
test_values[vec_td1_return_estimate-False-False] | 48.3085ms | 33.8465ms | 29.5452 Ops/s | 28.6066 Ops/s | |
test_values[td_lambda_return_estimate-True-False] | 37.0508ms | 33.9022ms | 29.4966 Ops/s | 30.0673 Ops/s | |
test_values[vec_td_lambda_return_estimate-True-False] | 34.7041ms | 33.2610ms | 30.0652 Ops/s | 28.5768 Ops/s | |
test_gae_speed[generalized_advantage_estimate-False-1-512] | 10.3006ms | 8.1775ms | 122.2861 Ops/s | 122.1459 Ops/s | |
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] | 2.1200ms | 1.8792ms | 532.1469 Ops/s | 542.0489 Ops/s | |
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] | 0.4262ms | 0.3516ms | 2.8440 KOps/s | 2.9292 KOps/s | |
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] | 41.3653ms | 38.7967ms | 25.7754 Ops/s | 23.5150 Ops/s | |
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] | 3.7827ms | 3.0327ms | 329.7415 Ops/s | 332.6822 Ops/s | |
test_dqn_speed | 6.7430ms | 1.3284ms | 752.7951 Ops/s | 740.0580 Ops/s | |
test_ddpg_speed | 3.0015ms | 2.6417ms | 378.5398 Ops/s | 367.0727 Ops/s | |
test_sac_speed | 73.0230ms | 8.6717ms | 115.3177 Ops/s | 123.7322 Ops/s | |
test_redq_speed | 14.6003ms | 12.9154ms | 77.4267 Ops/s | 76.5797 Ops/s | |
test_redq_deprec_speed | 13.6666ms | 12.9078ms | 77.4725 Ops/s | 77.9847 Ops/s | |
test_td3_speed | 8.5125ms | 8.0276ms | 124.5706 Ops/s | 124.1387 Ops/s | |
test_cql_speed | 37.0032ms | 35.7520ms | 27.9705 Ops/s | 27.8690 Ops/s | |
test_a2c_speed | 8.2958ms | 7.2784ms | 137.3934 Ops/s | 136.3433 Ops/s | |
test_ppo_speed | 8.9053ms | 7.5453ms | 132.5321 Ops/s | 128.1275 Ops/s | |
test_reinforce_speed | 7.4692ms | 6.4880ms | 154.1307 Ops/s | 153.6782 Ops/s | |
test_iql_speed | 32.8662ms | 32.1010ms | 31.1517 Ops/s | 30.7924 Ops/s | |
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] | 2.5921ms | 2.1797ms | 458.7726 Ops/s | 468.8643 Ops/s | |
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] | 0.8390ms | 0.4901ms | 2.0402 KOps/s | 2.0299 KOps/s | |
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] | 0.8723ms | 0.4683ms | 2.1355 KOps/s | 2.1544 KOps/s | |
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] | 2.5829ms | 2.1781ms | 459.1257 Ops/s | 469.1928 Ops/s | |
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] | 0.9823ms | 0.4842ms | 2.0651 KOps/s | 2.0727 KOps/s | |
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] | 0.6155ms | 0.4572ms | 2.1873 KOps/s | 2.1838 KOps/s | |
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] | 3.3469ms | 2.2845ms | 437.7265 Ops/s | 440.6586 Ops/s | |
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] | 1.2313ms | 0.5981ms | 1.6720 KOps/s | 1.6609 KOps/s | |
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] | 0.8391ms | 0.5783ms | 1.7292 KOps/s | 1.7339 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] | 3.3028ms | 2.1800ms | 458.7177 Ops/s | 463.1026 Ops/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] | 0.7310ms | 0.4903ms | 2.0395 KOps/s | 2.0561 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] | 0.6239ms | 0.4631ms | 2.1594 KOps/s | 1.7921 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] | 4.0126ms | 2.2105ms | 452.3883 Ops/s | 469.1312 Ops/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] | 0.9909ms | 0.4926ms | 2.0299 KOps/s | 2.0643 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] | 0.7803ms | 0.4604ms | 2.1721 KOps/s | 2.1685 KOps/s | |
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] | 3.4992ms | 2.2802ms | 438.5601 Ops/s | 439.0793 Ops/s | |
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] | 0.8506ms | 0.6087ms | 1.6430 KOps/s | 1.4316 KOps/s | |
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] | 91.8818ms | 0.6789ms | 1.4730 KOps/s | 1.7454 KOps/s | |
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] | 83.4421ms | 5.2779ms | 189.4689 Ops/s | 194.1432 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] | 14.5374ms | 12.1248ms | 82.4753 Ops/s | 82.9046 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] | 1.8026ms | 1.0549ms | 947.9649 Ops/s | 909.9193 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] | 87.4969ms | 6.9010ms | 144.9075 Ops/s | 147.2350 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] | 14.6395ms | 12.0894ms | 82.7168 Ops/s | 83.3640 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] | 1.8082ms | 1.0563ms | 946.7031 Ops/s | 944.0432 Ops/s | |
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] | 83.8242ms | 7.0758ms | 141.3268 Ops/s | 139.3802 Ops/s | |
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] | 14.9502ms | 12.3860ms | 80.7363 Ops/s | 81.4621 Ops/s | |
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] | 2.0888ms | 1.3591ms | 735.7942 Ops/s | 704.9272 Ops/s |
|
Name | Max | Mean | Ops | Ops on Repo HEAD
|
Change |
---|---|---|---|---|---|
test_single | 0.1197s | 0.1163s | 8.5975 Ops/s | 8.6205 Ops/s | |
test_sync | 96.0696ms | 95.7709ms | 10.4416 Ops/s | 10.4598 Ops/s | |
test_async | 0.1819s | 91.8702ms | 10.8849 Ops/s | 10.8347 Ops/s | |
test_single_pixels | 0.1985s | 0.1383s | 7.2300 Ops/s | 7.6446 Ops/s | |
test_sync_pixels | 83.1087ms | 81.5281ms | 12.2657 Ops/s | 12.1585 Ops/s | |
test_async_pixels | 0.1544s | 77.2140ms | 12.9510 Ops/s | 15.7331 Ops/s | |
test_simple | 0.8894s | 0.8352s | 1.1973 Ops/s | 1.1664 Ops/s | |
test_transformed | 1.1173s | 1.0588s | 0.9445 Ops/s | 0.9154 Ops/s | |
test_serial | 2.5272s | 2.4645s | 0.4058 Ops/s | 0.3966 Ops/s | |
test_parallel | 2.1671s | 2.0873s | 0.4791 Ops/s | 0.4740 Ops/s | |
test_step_mdp_speed[True-True-True-True-True] | 0.1072ms | 32.7082μs | 30.5734 KOps/s | 30.3322 KOps/s | |
test_step_mdp_speed[True-True-True-True-False] | 43.7610μs | 19.6714μs | 50.8353 KOps/s | 51.1013 KOps/s | |
test_step_mdp_speed[True-True-True-False-True] | 37.4300μs | 18.8026μs | 53.1841 KOps/s | 54.2066 KOps/s | |
test_step_mdp_speed[True-True-True-False-False] | 29.2800μs | 10.9380μs | 91.4242 KOps/s | 90.4130 KOps/s | |
test_step_mdp_speed[True-True-False-True-True] | 65.8510μs | 33.4269μs | 29.9160 KOps/s | 29.0962 KOps/s | |
test_step_mdp_speed[True-True-False-True-False] | 44.8400μs | 20.8175μs | 48.0364 KOps/s | 47.2311 KOps/s | |
test_step_mdp_speed[True-True-False-False-True] | 49.2400μs | 19.8589μs | 50.3552 KOps/s | 49.9286 KOps/s | |
test_step_mdp_speed[True-True-False-False-False] | 38.3210μs | 12.8880μs | 77.5918 KOps/s | 77.4371 KOps/s | |
test_step_mdp_speed[True-False-True-True-True] | 66.7710μs | 35.8279μs | 27.9112 KOps/s | 27.5834 KOps/s | |
test_step_mdp_speed[True-False-True-True-False] | 47.9600μs | 23.0007μs | 43.4770 KOps/s | 43.1498 KOps/s | |
test_step_mdp_speed[True-False-True-False-True] | 51.6810μs | 20.0892μs | 49.7781 KOps/s | 49.1897 KOps/s | |
test_step_mdp_speed[True-False-True-False-False] | 36.4110μs | 12.9192μs | 77.4044 KOps/s | 76.4533 KOps/s | |
test_step_mdp_speed[True-False-False-True-True] | 83.7320μs | 37.5331μs | 26.6432 KOps/s | 25.9078 KOps/s | |
test_step_mdp_speed[True-False-False-True-False] | 49.3700μs | 24.6343μs | 40.5938 KOps/s | 39.2296 KOps/s | |
test_step_mdp_speed[True-False-False-False-True] | 58.6010μs | 22.1483μs | 45.1501 KOps/s | 46.2375 KOps/s | |
test_step_mdp_speed[True-False-False-False-False] | 45.7610μs | 14.7255μs | 67.9096 KOps/s | 67.5911 KOps/s | |
test_step_mdp_speed[False-True-True-True-True] | 92.8210μs | 35.8524μs | 27.8922 KOps/s | 27.0643 KOps/s | |
test_step_mdp_speed[False-True-True-True-False] | 55.7410μs | 22.9670μs | 43.5407 KOps/s | 42.5914 KOps/s | |
test_step_mdp_speed[False-True-True-False-True] | 51.7610μs | 23.6010μs | 42.3712 KOps/s | 41.3042 KOps/s | |
test_step_mdp_speed[False-True-True-False-False] | 36.9110μs | 14.7746μs | 67.6838 KOps/s | 66.4299 KOps/s | |
test_step_mdp_speed[False-True-False-True-True] | 77.3820μs | 37.8654μs | 26.4093 KOps/s | 25.4218 KOps/s | |
test_step_mdp_speed[False-True-False-True-False] | 60.7410μs | 25.5538μs | 39.1331 KOps/s | 38.4341 KOps/s | |
test_step_mdp_speed[False-True-False-False-True] | 61.6510μs | 25.9748μs | 38.4989 KOps/s | 37.4128 KOps/s | |
test_step_mdp_speed[False-True-False-False-False] | 46.0510μs | 16.6188μs | 60.1727 KOps/s | 57.8502 KOps/s | |
test_step_mdp_speed[False-False-True-True-True] | 66.6510μs | 40.0073μs | 24.9955 KOps/s | 23.9469 KOps/s | |
test_step_mdp_speed[False-False-True-True-False] | 48.6610μs | 27.1393μs | 36.8469 KOps/s | 36.1599 KOps/s | |
test_step_mdp_speed[False-False-True-False-True] | 51.9610μs | 25.2675μs | 39.5765 KOps/s | 38.1678 KOps/s | |
test_step_mdp_speed[False-False-True-False-False] | 46.2510μs | 16.5512μs | 60.4185 KOps/s | 58.8860 KOps/s | |
test_step_mdp_speed[False-False-False-True-True] | 69.2910μs | 41.2098μs | 24.2661 KOps/s | 23.3573 KOps/s | |
test_step_mdp_speed[False-False-False-True-False] | 55.4210μs | 28.9215μs | 34.5763 KOps/s | 33.4372 KOps/s | |
test_step_mdp_speed[False-False-False-False-True] | 52.7010μs | 27.6230μs | 36.2017 KOps/s | 35.4049 KOps/s | |
test_step_mdp_speed[False-False-False-False-False] | 40.3700μs | 18.3526μs | 54.4882 KOps/s | 52.3527 KOps/s | |
test_values[generalized_advantage_estimate-True-True] | 25.9339ms | 25.2210ms | 39.6495 Ops/s | 40.7670 Ops/s | |
test_values[vec_generalized_advantage_estimate-True-True] | 84.6647ms | 3.2635ms | 306.4213 Ops/s | 301.8175 Ops/s | |
test_values[td0_return_estimate-False-False] | 96.9020μs | 60.3129μs | 16.5802 KOps/s | 16.9435 KOps/s | |
test_values[td1_return_estimate-False-False] | 55.7105ms | 53.8822ms | 18.5590 Ops/s | 19.2954 Ops/s | |
test_values[vec_td1_return_estimate-False-False] | 1.9890ms | 1.7640ms | 566.8953 Ops/s | 568.7573 Ops/s | |
test_values[td_lambda_return_estimate-True-False] | 88.8871ms | 86.5326ms | 11.5563 Ops/s | 11.4801 Ops/s | |
test_values[vec_td_lambda_return_estimate-True-False] | 3.8199ms | 1.7922ms | 557.9816 Ops/s | 560.2772 Ops/s | |
test_gae_speed[generalized_advantage_estimate-False-1-512] | 24.4089ms | 23.6592ms | 42.2669 Ops/s | 44.2776 Ops/s | |
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] | 0.9143ms | 0.7155ms | 1.3977 KOps/s | 1.4623 KOps/s | |
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] | 0.8026ms | 0.6593ms | 1.5167 KOps/s | 1.5425 KOps/s | |
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] | 1.5603ms | 1.4614ms | 684.2706 Ops/s | 692.7310 Ops/s | |
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] | 0.9916ms | 0.6883ms | 1.4528 KOps/s | 1.5219 KOps/s | |
test_dqn_speed | 8.1883ms | 1.4905ms | 670.9227 Ops/s | 653.6460 Ops/s | |
test_ddpg_speed | 3.1072ms | 2.7627ms | 361.9709 Ops/s | 358.9984 Ops/s | |
test_sac_speed | 81.9151ms | 8.6172ms | 116.0473 Ops/s | 125.8513 Ops/s | |
test_redq_speed | 11.2316ms | 10.1747ms | 98.2831 Ops/s | 96.7165 Ops/s | |
test_redq_deprec_speed | 11.6480ms | 11.0091ms | 90.8340 Ops/s | 91.2263 Ops/s | |
test_td3_speed | 8.3927ms | 7.9757ms | 125.3808 Ops/s | 125.4887 Ops/s | |
test_cql_speed | 26.2358ms | 25.4412ms | 39.3063 Ops/s | 39.3745 Ops/s | |
test_a2c_speed | 5.8739ms | 5.5191ms | 181.1892 Ops/s | 180.5996 Ops/s | |
test_ppo_speed | 5.9982ms | 5.7887ms | 172.7506 Ops/s | 171.1791 Ops/s | |
test_reinforce_speed | 6.1469ms | 4.5697ms | 218.8350 Ops/s | 223.4037 Ops/s | |
test_iql_speed | 19.9472ms | 19.3770ms | 51.6076 Ops/s | 51.1282 Ops/s | |
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] | 3.5617ms | 2.8865ms | 346.4387 Ops/s | 351.3294 Ops/s | |
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] | 0.6696ms | 0.5552ms | 1.8012 KOps/s | 1.8147 KOps/s | |
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] | 0.7246ms | 0.5324ms | 1.8783 KOps/s | 1.8966 KOps/s | |
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] | 3.1763ms | 2.8932ms | 345.6400 Ops/s | 347.9830 Ops/s | |
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] | 0.8540ms | 0.5458ms | 1.8323 KOps/s | 1.8359 KOps/s | |
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] | 4.3827ms | 0.5272ms | 1.8969 KOps/s | 1.9113 KOps/s | |
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] | 3.3082ms | 3.0347ms | 329.5219 Ops/s | 335.0267 Ops/s | |
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] | 0.8035ms | 0.6853ms | 1.4592 KOps/s | 1.4876 KOps/s | |
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] | 4.2985ms | 0.6610ms | 1.5128 KOps/s | 1.5294 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] | 3.0864ms | 2.8835ms | 346.8035 Ops/s | 349.3170 Ops/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] | 0.8133ms | 0.5524ms | 1.8103 KOps/s | 1.8040 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] | 4.2095ms | 0.5302ms | 1.8861 KOps/s | 1.8810 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] | 3.2488ms | 2.9148ms | 343.0817 Ops/s | 342.1059 Ops/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] | 0.6595ms | 0.5488ms | 1.8221 KOps/s | 1.8289 KOps/s | |
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] | 4.2387ms | 0.5266ms | 1.8990 KOps/s | 1.9137 KOps/s | |
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] | 3.3461ms | 3.0343ms | 329.5684 Ops/s | 333.7818 Ops/s | |
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] | 0.9742ms | 0.6756ms | 1.4802 KOps/s | 1.2919 KOps/s | |
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] | 0.1060s | 0.7509ms | 1.3317 KOps/s | 1.5248 KOps/s | |
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] | 0.1033s | 6.7429ms | 148.3041 Ops/s | 150.2873 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] | 17.5692ms | 15.1089ms | 66.1862 Ops/s | 65.1672 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] | 2.1111ms | 1.0827ms | 923.5801 Ops/s | 830.0409 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] | 0.1012s | 8.6182ms | 116.0339 Ops/s | 116.6915 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] | 17.6358ms | 15.0454ms | 66.4655 Ops/s | 66.0502 Ops/s | |
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] | 2.1598ms | 1.0918ms | 915.9190 Ops/s | 771.7874 Ops/s | |
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] | 0.1011s | 7.1092ms | 140.6625 Ops/s | 145.3824 Ops/s | |
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] | 0.1108s | 17.3567ms | 57.6147 Ops/s | 64.9777 Ops/s | |
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] | 1.7752ms | 1.4297ms | 699.4426 Ops/s | 625.8402 Ops/s |
Great thanks! The current errors don't seem related to the reduction parameter. |
I have added the on-policy tests here: #1972 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Labels
bug
Something isn't working
CLA Signed
This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
cc @albertbou92
The tests are failing because when collecting
loss.action_loss(...)
the tensor is not reduced.I think if the sub-loss is a public method the reduction should happen there.
Also we should only reduce the losses, not the metadata.
The PR solves both these issues