[Quality] Fix ufmt formatting in test_compose_and_env.py#3965
Open
vmoens wants to merge 1 commit into
Open
Conversation
The ufmt pre-commit hook collapses the torchrl._utils import back to a single line, so the lint CI job (pre-commit run --all-files) fails on main and on every PR. Apply the hook's formatting. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3965
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (3 Unrelated Failures)As of commit d099e7a with merge base 3761aea ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
BROKEN TRUNK - The following jobs failed but was 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. |
meta-cla
Bot
added
the
CLA Signed
Contributor
Benchmark Results: PR
|
| Benchmark | main ops | PR ops | Change |
|---|---|---|---|
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] |
186.45 | 56.72 | -69.58% |
benchmarks/test_objectives_benchmarks.py::test_redq_deprec_speed[False-None] |
59.28 | 87.60 | +47.79% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-64] |
7.5764 | 10.57 | +39.49% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
2,778 | 3,625 | +30.50% |
benchmarks/test_objectives_benchmarks.py::test_values[vec_td1_return_estimate-False-False] |
78.70 | 55.34 | -29.68% |
benchmarks/test_objectives_benchmarks.py::test_values[vec_generalized_advantage_estimate-True-True] |
78.47 | 55.44 | -29.35% |
benchmarks/test_objectives_benchmarks.py::test_values[vec_td_lambda_return_estimate-True-False] |
78.11 | 55.63 | -28.77% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-16] |
28.14 | 35.98 | +27.88% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-64] |
5.5845 | 7.0412 | +26.08% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] |
844.82 | 1,047 | +23.95% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] |
48.43 | 37.59 | -22.37% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-224-224-64] |
10.04 | 12.22 | +21.77% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
2,789 | 3,372 | +20.91% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[True-backward] |
822.86 | 989.94 | +20.30% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-backward] |
348.47 | 416.80 | +19.61% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
3,468 | 2,909 | -16.12% |
benchmarks/test_collectors_benchmark.py::test_sync |
16.10 | 13.72 | -14.82% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] |
517.31 | 445.99 | -13.79% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] |
2,141 | 1,855 | -13.34% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
3,126 | 2,711 | -13.25% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] |
2,179 | 1,934 | -11.24% |
benchmarks/test_collectors_benchmark.py::test_single_with_rb |
7.2259 | 6.4973 | -10.08% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] |
1,839 | 2,008 | +9.19% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-224-224-1] |
549.36 | 599.26 | +9.08% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] |
1,875 | 2,032 | +8.34% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[True-None] |
106.40 | 115.17 | +8.24% |
benchmarks/test_objectives_benchmarks.py::test_values[generalized_advantage_estimate-True-True] |
88.62 | 95.33 | +7.57% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[vec_generalized_advantage_estimate-True-32-512] |
33.70 | 31.15 | -7.56% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[100-img_shape1-atari] |
599.08 | 642.53 | +7.25% |
benchmarks/test_objectives_benchmarks.py::test_reinforce_speed[True-backward] |
116.08 | 124.33 | +7.11% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[True-backward] |
110.34 | 118.05 | +6.99% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] |
844.17 | 900.90 | +6.72% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-None] |
659.40 | 703.39 | +6.67% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] |
978.51 | 1,042 | +6.47% |
benchmarks/test_objectives_benchmarks.py::test_redq_speed[reduce-overhead-None] |
209.45 | 222.71 | +6.33% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
3,062 | 2,876 | -6.05% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[100-img_shape1-atari] |
4,907 | 5,202 | +5.99% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] |
519.19 | 549.06 | +5.75% |
benchmarks/test_objectives_benchmarks.py::test_redq_speed[True-None] |
211.10 | 222.98 | +5.63% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_stack_then_write[100-img_shape1-atari] |
262.30 | 276.74 | +5.51% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[vec_generalized_advantage_estimate-True-1-512] |
649.89 | 684.61 | +5.34% |
benchmarks/test_envs_benchmark.py::test_simple |
1.7929 | 1.6978 | -5.30% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-cudnn-False-0-gru] |
1.2940 | 1.3621 | +5.26% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[vec_generalized_advantage_estimate-False-32-512] |
561.62 | 532.32 | -5.22% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_stack_then_write[200-img_shape3-large_batch] |
133.71 | 140.36 | +4.97% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] |
539.33 | 566.12 | +4.97% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-False-False] |
55,931 | 53,158 | -4.96% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-False-0-gru] |
2.8719 | 3.0101 | +4.81% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-False-True] |
35,247 | 33,636 | -4.57% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] |
2,486 | 2,594 | +4.35% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-True-0-lstm] |
3.1067 | 2.9730 | -4.30% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-True-False] |
29,659 | 28,390 | -4.28% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[200-img_shape3-large_batch] |
316.91 | 330.44 | +4.27% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[True-backward] |
109.21 | 113.87 | +4.27% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-480-640-4] |
19.40 | 20.19 | +4.08% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_without_rb[100-img_shape0-atari] |
28.32 | 29.47 | +4.04% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[16-same] |
1,613 | 1,548 | -4.04% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[reduce-overhead-None] |
435.22 | 452.42 | +3.95% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[200-img_shape3-large_batch] |
297.85 | 309.42 | +3.89% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_with_rb[200-img_shape1-large_batch] |
9.8720 | 10.24 | +3.76% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-False-True] |
29,354 | 28,259 | -3.73% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_without_rb[200-img_shape1-large_batch] |
14.37 | 14.91 | +3.71% |
benchmarks/test_objectives_benchmarks.py::test_redq_deprec_speed[reduce-overhead-None] |
262.17 | 271.46 | +3.54% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
2,977 | 3,082 | +3.53% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-4] |
159.28 | 153.73 | -3.48% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-True-False] |
28,183 | 27,204 | -3.47% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[reduce-overhead-None] |
1,755 | 1,816 | +3.44% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[torch.save] |
7,135 | 6,891 | -3.41% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] |
2,888 | 2,791 | -3.35% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[True-backward] |
59.06 | 57.12 | -3.28% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-False-True] |
30,867 | 29,863 | -3.25% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] |
0.5271 | 0.5105 | -3.15% |
benchmarks/test_replaybuffer_benchmark.py::TestWindowingTransformsBenchmark::test_catframes_offline |
541.46 | 558.25 | +3.10% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] |
46.16 | 47.57 | +3.05% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-False-False] |
44,933 | 43,570 | -3.03% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-False-True] |
36,855 | 37,924 | +2.90% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-False-False] |
50,076 | 48,639 | -2.87% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] |
51.57 | 50.11 | -2.82% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[100-img_shape1-atari] |
674.59 | 693.21 | +2.76% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[untyped_storage] |
8.0549 | 8.2742 | +2.72% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-True-False] |
43,386 | 42,223 | -2.68% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[safetensors] |
22,381 | 22,979 | +2.67% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[reduce-overhead-None] |
112.72 | 115.71 | +2.65% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-False-False] |
56,322 | 57,807 | +2.64% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[True-backward] |
239.88 | 246.11 | +2.60% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-1] |
449.18 | 460.52 | +2.52% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-True-True] |
18,127 | 17,679 | -2.48% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] |
24.79 | 25.40 | +2.45% |
benchmarks/test_objectives_benchmarks.py::test_values[td0_return_estimate-False-False] |
7,535 | 7,719 | +2.43% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-True] |
22.86 | 23.41 | +2.37% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-False-0-lstm] |
1.9158 | 1.9612 | +2.37% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-False-False] |
77,667 | 79,442 | +2.28% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[4-constant] |
4,093 | 4,183 | +2.21% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[reduce-overhead-None] |
278.18 | 284.27 | +2.19% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[reduce-overhead-None] |
83.85 | 82.06 | -2.14% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-256-256-16] |
12.18 | 12.44 | +2.13% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] |
193.09 | 197.11 | +2.08% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-False-False] |
64,881 | 63,544 | -2.06% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[True-None] |
1,790 | 1,753 | -2.06% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] |
818.20 | 801.40 | -2.05% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-True] |
24.90 | 24.39 | -2.05% |
benchmarks/test_replaybuffer_benchmark.py::TestWindowingTransformsBenchmark::test_action_chunk_transform[done_aware] |
1,492 | 1,522 | +2.04% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[True-None] |
276.30 | 281.87 | +2.02% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[False-backward] |
31.81 | 32.45 | +1.99% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-224-224-64] |
4.4922 | 4.5791 | +1.94% |
benchmarks/test_replaybuffer_benchmark.py::TestPrioritizedReplayBufferBenchmark::test_sampler_sample_scale[1000000-cpu] |
96.40 | 98.25 | +1.93% |
benchmarks/test_objectives_benchmarks.py::test_redq_deprec_speed[True-backward] |
132.36 | 134.78 | +1.82% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-cudnn-False-0-lstm] |
0.8480 | 0.8632 | +1.79% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[True-backward] |
275.66 | 280.55 | +1.77% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-480-640-1] |
79.06 | 80.45 | +1.76% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-True-0-gru] |
4.1141 | 4.1843 | +1.71% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[False-backward] |
90.65 | 89.12 | -1.70% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-480-640-16] |
4.9790 | 5.0634 | +1.70% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-True-True] |
20,781 | 21,132 | +1.69% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-224-224-16] |
47.85 | 48.65 | +1.68% |
benchmarks/test_collectors_benchmark.py::test_single |
8.9927 | 8.8452 | -1.64% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[100-img_shape2-large_img] |
386.03 | 392.30 | +1.62% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[100-img_shape2-large_img] |
403.85 | 410.28 | +1.59% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-256-256-4] |
48.70 | 49.48 | +1.58% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-1] |
490.53 | 498.28 | +1.58% |
| ... | ... | ... | Showing 120 of 220 comparisons, sorted by absolute change. |
GPU
Compared 230 benchmarks. Regressions over 5%: 8. Improvements over 5%: 21.
| Benchmark | main ops | PR ops | Change |
|---|---|---|---|
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] |
43.81 | 196.78 | +349.20% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] |
856.73 | 2,917 | +240.47% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[100-img_shape1-atari] |
2,512 | 4,229 | +68.34% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[reduce-overhead-None] |
61.08 | 89.71 | +46.86% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
2,788 | 3,418 | +22.60% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
2,862 | 3,492 | +22.03% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[reduce-overhead-None] |
103.79 | 81.81 | -21.17% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[reduce-overhead-None] |
85.90 | 101.01 | +17.60% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
3,423 | 2,915 | -14.84% |
benchmarks/test_replaybuffer_benchmark.py::TestWindowingTransformsBenchmark::test_catframes_offline |
532.40 | 609.45 | +14.47% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] |
768.24 | 863.88 | +12.45% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-64] |
6.3011 | 7.0723 | +12.24% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] |
426.39 | 476.72 | +11.80% |
benchmarks/test_objectives_benchmarks.py::test_values[vec_generalized_advantage_estimate-True-True] |
417.17 | 462.71 | +10.92% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-backward] |
454.32 | 501.95 | +10.48% |
benchmarks/test_objectives_benchmarks.py::test_reinforce_speed[True-backward] |
354.79 | 390.33 | +10.02% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] |
2,096 | 2,300 | +9.73% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] |
727.05 | 797.69 | +9.71% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[True-backward] |
381.12 | 414.52 | +8.77% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[True-backward] |
325.67 | 352.83 | +8.34% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] |
838.69 | 770.32 | -8.15% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[True-backward] |
242.31 | 261.97 | +8.11% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] |
545.96 | 502.95 | -7.88% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
3,697 | 3,443 | -6.88% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-224-224-1] |
583.52 | 622.90 | +6.75% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[safetensors] |
24,705 | 23,311 | -5.64% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[generalized_advantage_estimate-False-1-512] |
47.35 | 49.90 | +5.39% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[100-img_shape1-atari] |
663.94 | 629.06 | -5.25% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
3,423 | 3,247 | -5.14% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[vec_generalized_advantage_estimate-False-1-512] |
1,311 | 1,374 | +4.83% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-True-True] |
20,922 | 21,900 | +4.68% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-224-224-16] |
46.27 | 48.41 | +4.63% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[reduce-overhead-None] |
44.18 | 46.14 | +4.42% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[False-None] |
340.50 | 355.15 | +4.30% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[100-img_shape2-large_img] |
593.00 | 567.81 | -4.25% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-16] |
40.63 | 42.29 | +4.10% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] |
3,462 | 3,321 | -4.09% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-False] |
54.91 | 52.68 | -4.05% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-4] |
153.05 | 159.05 | +3.92% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-True-False] |
35,441 | 36,808 | +3.86% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[True-backward] |
373.03 | 358.95 | -3.77% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-False-True] |
35,384 | 36,710 | +3.75% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] |
49.45 | 47.65 | -3.65% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
3,252 | 3,136 | -3.57% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[200-img_shape3-large_batch] |
738.69 | 712.44 | -3.55% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-True-0-gru] |
51.61 | 53.41 | +3.49% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-224-224-1] |
287.80 | 277.86 | -3.45% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-None] |
832.77 | 861.42 | +3.44% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[True-backward] |
361.22 | 373.60 | +3.43% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[False-backward] |
154.47 | 149.21 | -3.40% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] |
0.5343 | 0.5166 | -3.31% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] |
193.34 | 199.68 | +3.28% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_stack_then_write[200-img_shape3-large_batch] |
140.38 | 135.80 | -3.26% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-True-True] |
20,992 | 21,659 | +3.18% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-224-224-64] |
11.97 | 12.35 | +3.17% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-False-True-False] |
33,311 | 32,261 | -3.15% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[True-None] |
1,898 | 1,958 | +3.15% |
benchmarks/test_envs_benchmark.py::test_simple |
1.2686 | 1.2294 | -3.09% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[reduce-overhead-None] |
853.51 | 879.53 | +3.05% |
benchmarks/test_objectives_benchmarks.py::test_redq_deprec_speed[True-backward] |
281.87 | 290.44 | +3.04% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-256-256-1] |
195.45 | 189.58 | -3.00% |
benchmarks/test_replaybuffer_benchmark.py::TestPrioritizedReplayBufferBenchmark::test_sample_mixed_devices[1000000-memmap_cpu_storage_cud... |
967.93 | 995.94 | +2.89% |
benchmarks/test_collectors_benchmark.py::test_sync_pixels |
10.15 | 10.44 | +2.77% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-224-224-4] |
72.71 | 70.71 | -2.76% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[vec_generalized_advantage_estimate-False-32-512] |
1,341 | 1,304 | -2.76% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] |
173.04 | 177.74 | +2.72% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] |
53.18 | 51.74 | -2.70% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[True-backward] |
235.01 | 228.69 | -2.69% |
benchmarks/test_objectives_benchmarks.py::test_values[vec_td1_return_estimate-False-False] |
840.82 | 863.29 | +2.67% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[True-None] |
709.68 | 728.63 | +2.67% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[reduce-overhead-None] |
1,848 | 1,897 | +2.64% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] |
531.30 | 545.27 | +2.63% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-64] |
10.39 | 10.66 | +2.62% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-True-False] |
40,885 | 39,819 | -2.61% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[100-img_shape1-atari] |
702.28 | 684.44 | -2.54% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-True-False] |
32,229 | 33,044 | +2.53% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-4] |
139.02 | 142.52 | +2.52% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[False-backward] |
69.32 | 71.03 | +2.47% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-False-True] |
30,510 | 31,254 | +2.44% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[True-None] |
519.69 | 531.91 | +2.35% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] |
1,022 | 1,046 | +2.35% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-False-True] |
43,501 | 42,496 | -2.31% |
benchmarks/test_envs_benchmark.py::test_serial |
0.4295 | 0.4394 | +2.30% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-256-256-64] |
3.0997 | 3.0295 | -2.27% |
benchmarks/test_collectors_benchmark.py::test_async |
11.03 | 11.28 | +2.26% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-False-False] |
59,432 | 58,127 | -2.20% |
benchmarks/test_envs_benchmark.py::test_parallel |
0.5325 | 0.5440 | +2.15% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-False-True] |
33,823 | 33,101 | -2.13% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-True-True] |
23,905 | 24,409 | +2.11% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[200-img_shape3-large_batch] |
327.57 | 320.76 | -2.08% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[False-backward] |
77.68 | 79.29 | +2.07% |
benchmarks/test_objectives_benchmarks.py::test_values[td_lambda_return_estimate-True-False] |
12.18 | 12.44 | +2.07% |
benchmarks/test_collectors_benchmark.py::test_sync_preempt |
10.64 | 10.42 | -2.05% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-16] |
35.44 | 36.15 | +2.01% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] |
165.45 | 168.76 | +2.00% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] |
187.37 | 191.06 | +1.97% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[True-None] |
396.18 | 403.94 | +1.96% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[False-backward] |
40.60 | 39.84 | -1.89% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[reduce-overhead-None] |
828.88 | 844.19 | +1.85% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] |
23.73 | 23.29 | -1.83% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-1] |
447.89 | 456.09 | +1.83% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[True-None] |
644.37 | 656.12 | +1.82% |
benchmarks/test_collectors_benchmark.py::test_async_pixels |
10.99 | 10.80 | -1.76% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-True-False] |
30,128 | 29,602 | -1.75% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] |
171.75 | 174.75 | +1.75% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[False-backward] |
83.09 | 84.53 | +1.73% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[100-img_shape2-large_img] |
407.40 | 400.47 | -1.70% |
benchmarks/test_envs_benchmark.py::test_transformed |
0.7292 | 0.7169 | -1.69% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] |
171.35 | 174.24 | +1.68% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[16-constant] |
4,849 | 4,931 | +1.68% |
benchmarks/test_objectives_benchmarks.py::test_reinforce_speed[True-None] |
772.24 | 784.95 | +1.65% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-True-False] |
43,563 | 42,854 | -1.63% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[16-same] |
3,695 | 3,636 | -1.60% |
benchmarks/test_objectives_benchmarks.py::test_values[td1_return_estimate-False-False] |
20.15 | 20.46 | +1.54% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] |
2,072 | 2,040 | -1.54% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_without_rb_cuda[100-img_shape0-atari] |
17.75 | 17.48 | -1.53% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] |
167.22 | 169.78 | +1.53% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-True-False] |
35,687 | 35,146 | -1.52% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[reduce-overhead-None] |
880.39 | 893.22 | +1.46% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-256-256-4] |
49.21 | 48.50 | -1.44% |
| ... | ... | ... | Showing 120 of 230 comparisons, sorted by absolute change. |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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.
The ufmt pre-commit hook reformats the
torchrl._utilsimport intest/transforms/test_compose_and_env.pyback to a single line, so thepython-source-and-configslint job (pre-commit run --all-files) currently fails on main and turnslint-donered on every open PR (e.g. #3963).This applies the hook's formatting;
pre-commit run ufmt --files test/transforms/test_compose_and_env.pynow passes.🤖 Generated with Claude Code