[CI] Fix optional library dependency conflicts#3899
Merged
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3899
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ✅ No FailuresAs of commit bed8a1d with merge base 1f0b769 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
github-actions
Bot
added
CI
vmoens
force-pushed
the
codex/fix-chess-jumanji-ci
branch
from
40812de to
14b8cb9
Compare
vmoens
changed the title
Contributor
Benchmark Results: PR
|
| Benchmark | main ops | PR ops | Change |
|---|---|---|---|
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] |
198.68 | 54.21 | -72.71% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-backward] |
343.08 | 426.72 | +24.38% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
3,133 | 3,741 | +19.41% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] |
957.40 | 1,143 | +19.33% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[4-same] |
25.43 | 30.12 | +18.45% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[True-backward] |
256.04 | 288.19 | +12.56% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[16-same] |
23.80 | 21.09 | -11.40% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
2,898 | 2,586 | -10.78% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] |
859.73 | 951.67 | +10.69% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
3,227 | 2,888 | -10.51% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
3,186 | 2,871 | -9.88% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-True-False] |
26,120 | 28,280 | +8.27% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-True-True] |
17,256 | 18,651 | +8.09% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-False-True] |
32,419 | 35,014 | +8.01% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[untyped_storage] |
8.7236 | 9.4106 | +7.88% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
2,973 | 3,202 | +7.74% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-False-False] |
55,152 | 59,210 | +7.36% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] |
2,316 | 2,158 | -6.83% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-True-False] |
33,425 | 35,690 | +6.78% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-False-True-True] |
18,835 | 20,029 | +6.34% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-True-True] |
18,198 | 19,328 | +6.21% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[4-constant] |
4,541 | 4,262 | -6.15% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[100-img_shape2-large_img] |
405.64 | 429.97 | +6.00% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] |
2,044 | 1,924 | -5.85% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-False-False-False] |
48,214 | 50,992 | +5.76% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[True-backward] |
114.32 | 120.88 | +5.73% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-True-True] |
19,910 | 21,047 | +5.71% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] |
736.35 | 777.60 | +5.60% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-False-False] |
53,013 | 55,946 | +5.53% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-True-False] |
28,296 | 29,858 | +5.52% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-True-False] |
30,578 | 32,163 | +5.18% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-False-True-False] |
30,915 | 32,412 | +4.84% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-False-False-True] |
29,128 | 30,538 | +4.84% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[100-img_shape2-large_img] |
387.60 | 406.13 | +4.78% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-False-False] |
48,683 | 50,925 | +4.60% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[pickle] |
11,997 | 12,526 | +4.41% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] |
696.57 | 726.47 | +4.29% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_stack_then_write[100-img_shape1-atari] |
272.50 | 284.17 | +4.28% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-False-False] |
43,773 | 45,623 | +4.23% |
benchmarks/test_objectives_benchmarks.py::test_values[td1_return_estimate-False-False] |
37.42 | 38.96 | +4.13% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-False-True] |
27,919 | 29,067 | +4.11% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-True-True] |
18,310 | 19,058 | +4.08% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[True-backward] |
967.76 | 1,007 | +4.02% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-True-True] |
21,443 | 22,290 | +3.95% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] |
2,677 | 2,782 | +3.93% |
benchmarks/test_objectives_benchmarks.py::test_values[td_lambda_return_estimate-True-False] |
25.23 | 26.21 | +3.89% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[100-img_shape1-atari] |
5,197 | 5,399 | +3.88% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] |
2,128 | 2,046 | -3.86% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[numpy] |
356,063 | 369,554 | +3.79% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-True-False] |
33,916 | 35,188 | +3.75% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-False-False] |
62,437 | 64,710 | +3.64% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[False-backward] |
28.94 | 27.90 | -3.62% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-True-False] |
37,927 | 39,236 | +3.45% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[True-backward] |
57.70 | 59.68 | +3.43% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[True-None] |
1,814 | 1,752 | -3.42% |
benchmarks/test_envs_benchmark.py::test_transformed |
0.9066 | 0.9371 | +3.36% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-single-True] |
1.3825 | 1.3362 | -3.35% |
benchmarks/test_objectives_benchmarks.py::test_values[generalized_advantage_estimate-True-True] |
101.07 | 104.37 | +3.27% |
benchmarks/test_objectives_benchmarks.py::test_values[td0_return_estimate-False-False] |
8,090 | 7,828 | -3.24% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-False-True] |
37,245 | 38,445 | +3.22% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-False-0-gru] |
2.9616 | 3.0568 | +3.21% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_without_rb[200-img_shape1-large_batch] |
15.28 | 15.76 | +3.20% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_stack_then_write[50-img_shape0-small] |
891.39 | 862.97 | -3.19% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[True-None] |
288.21 | 297.40 | +3.19% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-False-0-lstm] |
1.9760 | 2.0373 | +3.10% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_without_rb[100-img_shape0-atari] |
30.12 | 31.02 | +2.99% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-False-False] |
62,821 | 64,695 | +2.98% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-True-False] |
41,916 | 43,162 | +2.97% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] |
54.92 | 53.30 | -2.95% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-False-True] |
29,370 | 30,225 | +2.91% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-True-True] |
20,472 | 21,066 | +2.90% |
benchmarks/test_objectives_benchmarks.py::test_redq_speed[False-None] |
98.06 | 95.23 | -2.89% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[200-img_shape3-large_batch] |
328.80 | 338.24 | +2.87% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[16-constant] |
2,723 | 2,647 | -2.82% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[False-None] |
359.52 | 349.47 | -2.80% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-True-True] |
23,249 | 23,876 | +2.70% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[reduce-overhead-None] |
115.08 | 118.15 | +2.67% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[reduce-overhead-None] |
700.90 | 719.41 | +2.64% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[True-None] |
485.00 | 472.68 | -2.54% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[100-img_shape2-large_img] |
565.97 | 580.06 | +2.49% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-None] |
696.26 | 713.51 | +2.48% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] |
530.68 | 517.76 | -2.43% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[True-backward] |
248.39 | 254.43 | +2.43% |
benchmarks/test_envs_benchmark.py::test_parallel |
1.0011 | 0.9768 | -2.43% |
benchmarks/test_envs_benchmark.py::test_serial |
0.5881 | 0.6022 | +2.41% |
benchmarks/test_objectives_benchmarks.py::test_reinforce_speed[True-backward] |
125.29 | 128.30 | +2.40% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] |
525.54 | 538.02 | +2.37% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-224-224-1] |
275.48 | 269.24 | -2.26% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_with_rb[100-img_shape0-atari] |
26.26 | 26.84 | +2.23% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[reduce-overhead-None] |
568.97 | 581.40 | +2.19% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[100-img_shape1-atari] |
638.90 | 652.72 | +2.16% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-False-True] |
31,925 | 32,575 | +2.04% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[200-img_shape3-large_batch] |
310.70 | 316.93 | +2.00% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[True-backward] |
59.55 | 58.40 | -1.94% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[False-None] |
183.91 | 180.37 | -1.93% |
benchmarks/test_objectives_benchmarks.py::test_reinforce_speed[False-None] |
210.55 | 214.51 | +1.88% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-cudnn-False-0-gru] |
1.3920 | 1.3665 | -1.84% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-cudnn-True-0-gru] |
1.4560 | 1.4298 | -1.80% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_stack_then_write[100-img_shape2-large_img] |
176.00 | 172.89 | -1.76% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-True-0-lstm] |
3.1177 | 3.1713 | +1.72% |
benchmarks/test_objectives_benchmarks.py::test_redq_speed[reduce-overhead-None] |
223.92 | 227.77 | +1.72% |
benchmarks/test_objectives_benchmarks.py::test_redq_speed[True-None] |
221.58 | 225.28 | +1.67% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-False-False] |
77,284 | 78,563 | +1.65% |
benchmarks/test_collectors_benchmark.py::test_sync |
17.02 | 17.29 | +1.59% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[False-backward] |
85.77 | 84.41 | -1.58% |
benchmarks/test_replaybuffer_benchmark.py::TestPrioritizedReplayBufferBenchmark::test_sampler_sample_scale[1000000-cpu] |
98.84 | 97.33 | -1.53% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] |
1,077 | 1,094 | +1.53% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[vec_generalized_advantage_estimate-True-1-512] |
658.30 | 648.64 | -1.47% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] |
0.2158 | 0.2189 | +1.44% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[False-backward] |
522.07 | 529.26 | +1.38% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] |
170.85 | 173.20 | +1.37% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[reduce-overhead-None] |
288.07 | 291.91 | +1.33% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] |
167.76 | 165.57 | -1.30% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
2,975 | 3,013 | +1.30% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[reduce-overhead-None] |
488.58 | 482.26 | -1.29% |
benchmarks/test_objectives_benchmarks.py::test_redq_deprec_speed[True-None] |
284.76 | 281.09 | -1.29% |
benchmarks/test_collectors_benchmark.py::test_single_with_rb |
8.8262 | 8.9383 | +1.27% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-False-True] |
37,637 | 37,163 | -1.26% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[True-None] |
262.68 | 265.98 | +1.26% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[True-backward] |
113.90 | 115.32 | +1.25% |
| ... | ... | ... | Showing 120 of 216 comparisons, sorted by absolute change. |
GPU
Compared 226 benchmarks. Regressions over 5%: 18. Improvements over 5%: 14.
| Benchmark | main ops | PR ops | Change |
|---|---|---|---|
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] |
44.07 | 195.56 | +343.71% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] |
951.05 | 3,173 | +233.67% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] |
190.96 | 39.94 | -79.09% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] |
185.43 | 50.59 | -72.72% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] |
2,562 | 3,467 | +35.36% |
benchmarks/test_objectives_benchmarks.py::test_reinforce_speed[reduce-overhead-None] |
100.79 | 126.88 | +25.89% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
3,779 | 2,843 | -24.77% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
2,937 | 3,609 | +22.87% |
benchmarks/test_objectives_benchmarks.py::test_redq_deprec_speed[reduce-overhead-None] |
105.92 | 84.44 | -20.27% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[100-img_shape1-atari] |
3,692 | 4,358 | +18.04% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
3,544 | 2,927 | -17.38% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] |
2,108 | 1,803 | -14.43% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] |
2,220 | 1,961 | -11.65% |
benchmarks/test_collectors_benchmark.py::test_single_with_rb_pixels |
5.4134 | 4.7883 | -11.55% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] |
3,378 | 3,035 | -10.15% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[200-img_shape3-large_batch] |
802.20 | 731.86 | -8.77% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[100-img_shape2-large_img] |
595.23 | 548.69 | -7.82% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] |
952.18 | 880.19 | -7.56% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-backward] |
448.63 | 482.54 | +7.56% |
benchmarks/test_objectives_benchmarks.py::test_values[vec_generalized_advantage_estimate-True-True] |
299.81 | 277.90 | -7.31% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] |
1,997 | 1,851 | -7.28% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[pickle] |
12,419 | 11,645 | -6.23% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[True-backward] |
315.17 | 334.25 | +6.05% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[torch.save] |
7,344 | 6,905 | -5.99% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[safetensors] |
24,314 | 22,871 | -5.94% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-4] |
141.51 | 149.81 | +5.86% |
benchmarks/test_envs_benchmark.py::test_simple |
1.2495 | 1.1781 | -5.72% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] |
731.82 | 771.71 | +5.45% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] |
515.22 | 542.21 | +5.24% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-parallel-buffers-False] |
0.5796 | 0.6100 | +5.23% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[100-img_shape2-large_img] |
388.58 | 408.23 | +5.06% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[50-img_shape0-small] |
4,220 | 4,431 | +5.02% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[100-img_shape2-large_img] |
408.60 | 425.66 | +4.18% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-1] |
478.31 | 458.81 | -4.08% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-False-True] |
39,902 | 38,283 | -4.06% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-True-True] |
23,035 | 23,945 | +3.95% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-False-True] |
32,699 | 33,974 | +3.90% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-False-False] |
65,346 | 62,949 | -3.67% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-True-False] |
32,316 | 31,139 | -3.64% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[50-img_shape0-small] |
3,433 | 3,557 | +3.62% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-False-False] |
51,650 | 49,795 | -3.59% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] |
47.98 | 49.69 | +3.56% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-False-False] |
46,092 | 44,514 | -3.42% |
benchmarks/test_envs_benchmark.py::test_transformed |
0.7186 | 0.6949 | -3.30% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[reduce-overhead-None] |
794.53 | 820.10 | +3.22% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-1] |
501.43 | 517.35 | +3.17% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] |
849.56 | 823.60 | -3.06% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-False-False] |
59,302 | 57,492 | -3.05% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_stack_then_write[100-img_shape2-large_img] |
173.20 | 178.45 | +3.03% |
benchmarks/test_objectives_benchmarks.py::test_values[td1_return_estimate-False-False] |
19.91 | 20.49 | +2.91% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] |
0.2091 | 0.2150 | +2.82% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[True-None] |
1,934 | 1,880 | -2.77% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[4-same] |
6.4571 | 6.6358 | +2.77% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[False-backward] |
77.96 | 80.04 | +2.68% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-4] |
162.18 | 166.45 | +2.63% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[100-img_shape1-atari] |
709.01 | 690.37 | -2.63% |
benchmarks/test_objectives_benchmarks.py::test_values[generalized_advantage_estimate-True-True] |
47.63 | 48.86 | +2.57% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[False-backward] |
233.41 | 239.28 | +2.51% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] |
2,894 | 2,821 | -2.51% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-False-False] |
55,929 | 54,569 | -2.43% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[reduce-overhead-None] |
851.88 | 872.49 | +2.42% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[True-None] |
732.24 | 749.87 | +2.41% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] |
1,932 | 1,885 | -2.40% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] |
167.03 | 171.04 | +2.40% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-single-True] |
1.3172 | 1.3488 | +2.39% |
benchmarks/test_objectives_benchmarks.py::test_values[td_lambda_return_estimate-True-False] |
12.07 | 12.35 | +2.35% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[reduce-overhead-None] |
103.91 | 106.28 | +2.29% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-16] |
43.04 | 44.02 | +2.28% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] |
545.56 | 533.16 | -2.27% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-480-640-16] |
36.48 | 37.31 | +2.26% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-False-False] |
77,562 | 75,821 | -2.25% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-False-True-True] |
19,979 | 19,545 | -2.17% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[True-None] |
369.64 | 377.44 | +2.11% |
benchmarks/test_envs_benchmark.py::test_serial |
0.4245 | 0.4157 | -2.07% |
benchmarks/test_compressed_storage_benchmark.py::TestCompressedStorageBenchmark::test_tensor_to_bytestream_speed[untyped_storage] |
8.5244 | 8.7001 | +2.06% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] |
164.93 | 168.31 | +2.05% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[True-None] |
814.01 | 830.56 | +2.03% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-True-True-True] |
20,480 | 20,893 | +2.02% |
benchmarks/test_envs_benchmark.py::test_cat_frames_functional[16-same] |
5.3843 | 5.4887 | +1.94% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] |
169.18 | 172.30 | +1.84% |
benchmarks/test_objectives_benchmarks.py::test_dqn_speed[True-backward] |
888.95 | 905.16 | +1.82% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-True-False-True] |
30,561 | 31,114 | +1.81% |
benchmarks/test_objectives_benchmarks.py::test_values[vec_td_lambda_return_estimate-True-False] |
852.53 | 837.42 | -1.77% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-True-False-True-False] |
32,355 | 31,782 | -1.77% |
benchmarks/test_objectives_benchmarks.py::test_a2c_speed[True-backward] |
343.96 | 350.04 | +1.77% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_contiguous[50-img_shape0-small] |
6,035 | 6,140 | +1.74% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-224-224-64] |
4.5185 | 4.5965 | +1.73% |
benchmarks/test_objectives_benchmarks.py::test_td3_speed[False-None] |
111.17 | 113.07 | +1.71% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[False-backward] |
126.40 | 128.53 | +1.69% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] |
499.57 | 491.17 | -1.68% |
benchmarks/test_objectives_benchmarks.py::test_ddpg_speed[reduce-overhead-None] |
823.85 | 837.65 | +1.67% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-True-False-True] |
42,590 | 43,301 | +1.67% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-True-False-True-False] |
38,960 | 38,328 | -1.62% |
benchmarks/test_storage_write_benchmark.py::TestCollectorIntegrationBenchmark::test_collector_without_rb_cuda[100-img_shape0-atari] |
17.59 | 17.31 | -1.58% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] |
164.11 | 166.70 | +1.58% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[True-False-True-True-True] |
21,252 | 20,921 | -1.56% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] |
3,746 | 3,689 | -1.52% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-True-False] |
28,054 | 27,636 | -1.49% |
benchmarks/test_envs_benchmark.py::test_step_mdp_speed[False-False-False-False-True] |
29,004 | 28,574 | -1.48% |
benchmarks/test_objectives_benchmarks.py::test_ppo_speed[True-None] |
688.60 | 698.62 | +1.46% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] |
1,011 | 1,026 | +1.45% |
benchmarks/test_objectives_benchmarks.py::test_sac_speed[True-None] |
610.76 | 619.57 | +1.44% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-224-224-64] |
12.62 | 12.80 | +1.42% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_storage_write_lazystack[200-img_shape3-large_batch] |
328.72 | 333.30 | +1.39% |
benchmarks/test_objectives_benchmarks.py::test_gae_speed[vec_generalized_advantage_estimate-False-1-512] |
1,324 | 1,306 | -1.36% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-256-256-16] |
12.10 | 12.27 | +1.35% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[torchvision-256-256-64] |
10.88 | 11.03 | +1.33% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-False-0-gru] |
22.55 | 22.84 | +1.30% |
benchmarks/test_objectives_benchmarks.py::test_cql_speed[reduce-overhead-None] |
88.58 | 89.68 | +1.25% |
benchmarks/test_objectives_benchmarks.py::test_reinforce_speed[True-None] |
758.76 | 768.10 | +1.23% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-480-640-1] |
77.55 | 78.50 | +1.22% |
benchmarks/test_vla_preprocessing_benchmark.py::test_openvla_preprocessing_throughput[pil-256-256-4] |
48.08 | 48.66 | +1.21% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[True-None] |
513.96 | 520.16 | +1.21% |
benchmarks/test_non_tensor_env_benchmark.py::test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] |
0.5224 | 0.5286 | +1.19% |
benchmarks/test_replaybuffer_benchmark.py::test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] |
162.47 | 164.31 | +1.14% |
benchmarks/test_rnn_reset_backends_benchmark.py::test_rnn_rollout_with_intermediate_resets[b256-t128-i32-h512-scan-False-0-lstm] |
21.22 | 21.46 | +1.12% |
benchmarks/test_storage_write_benchmark.py::TestStorageWriteBenchmark::test_collector_lazystack_then_write[200-img_shape3-large_batch] |
308.02 | 311.30 | +1.06% |
benchmarks/test_objectives_benchmarks.py::test_iql_speed[False-backward] |
68.28 | 68.99 | +1.03% |
benchmarks/test_replaybuffer_benchmark.py::TestPrioritizedReplayBufferBenchmark::test_sample_mixed_devices[1000000-cuda_storage_cpu_sampler] |
90.64 | 89.72 | -1.01% |
benchmarks/test_objectives_benchmarks.py::test_redq_deprec_speed[False-backward] |
70.66 | 71.37 | +1.01% |
| ... | ... | ... | Showing 120 of 226 comparisons, sorted by absolute change. |
vmoens
changed the title
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.
Summary
--no-depsin chess and Jumanji CI jobs to avoid resolver failures from stale nightly wheel metadatatorchvision.io.decode_imagemujoco<3.10.0for robosuite 1.4.0 compatibility with themj_fullMbindingTests
python -m compileall -q torchrl/envs/custom/chess.py torchrl/envs/libs/jumanji.pybash -n .github/unittest/linux_libs/scripts_chess/install.sh .github/unittest/linux_libs/scripts_jumanji/install.shbash -n .github/unittest/linux_libs/scripts_libero/install.sh .github/unittest/linux_libs/scripts_libero/run_test.shpython -m pip install --dry-run --pre torch --index-url https://download.pytorch.org/whl/nightly/cu128 --platform manylinux_2_28_x86_64 --platform manylinux_2_27_x86_64 --platform manylinux_2_24_x86_64 --platform manylinux_2_17_x86_64 --platform manylinux2014_x86_64 --python-version 3.10 --implementation cp --abi cp310 --only-binary=:all:python -m pip install --dry-run --pre torchvision --index-url https://download.pytorch.org/whl/nightly/cu128 --no-deps --platform manylinux_2_28_x86_64 --platform manylinux_2_27_x86_64 --platform manylinux_2_24_x86_64 --platform manylinux_2_17_x86_64 --platform manylinux2014_x86_64 --python-version 3.10 --implementation cp --abi cp310 --only-binary=:all:uv pip install --dry-run --target /tmp/rl-libero-dryrun --python-version 3.10 --python-platform x86_64-manylinux_2_28 "bddl==1.0.1" easydict "gym==0.25.2" h5py imageio matplotlib "mujoco<3.10.0" "numpy<2" opencv-python "robosuite==1.4.0" termcolor tqdm