[Refactor] composite_lp_aggregate to handle log-probs aggregates globally

[vmoens]

I propose a global flag composite_lp_aggregate to handle the issue of the aggregation of log-probs.

So far, we have dealt with this using kwargs everywhere (in ProbabilisticTDModule, ProbabilisticTDSequential, CompositeDistribution and subclasses). The hierarchy of these classes and what to do when args conflict isn't easy to handle. It's also confusing for users, who I suspect will usually want to work with either collapsed or non-collapsed log-probs.

A global flag (set to True for now and False in the future) will make things easier to handle.
Globally, the v0.6.2 behaviour will not be changed all users who rely on it will be informed about the upcoming change through a warning that will tell them to set the global var to False to accommodate upcoming changes. If they set it to True, nothing will change for them but that also means that bugs will not be solved (we won't maintain the True behaviour).

When composite_lp_aggregate() == True, we'll have aggregate_probabilities=True, include_sum=True and inplace=True by default everywhere. When composite_lp_aggregate() == False, all of these will be set to False, meaning that any call to whatever.log_prob(tensordict) will return another tensordict containing the leaf log-probs.

[github-actions]

$\color{#D29922}\textsf{\Large⚠\kern{0.2cm}\normalsize Warning}$ Result of CPU Benchmark Tests

Total Benchmarks: 217. Improved: $\large\color{#35bf28}15$. Worsened: $\large\color{#d91a1a}7$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	47.6490μs	19.1288μs	52.2773 KOps/s	50.9612 KOps/s	$\color{#35bf28}+2.58\%$
test_plain_set_stack_nested	50.6140μs	19.4405μs	51.4391 KOps/s	50.1858 KOps/s	$\color{#35bf28}+2.50\%$
test_plain_set_nested_inplace	72.2750μs	21.0282μs	47.5551 KOps/s	46.3551 KOps/s	$\color{#35bf28}+2.59\%$
test_plain_set_stack_nested_inplace	62.2160μs	21.0214μs	47.5706 KOps/s	46.5642 KOps/s	$\color{#35bf28}+2.16\%$
test_items	39.4940μs	4.1608μs	240.3359 KOps/s	231.1232 KOps/s	$\color{#35bf28}+3.99\%$
test_items_nested	0.4813ms	0.3958ms	2.5263 KOps/s	2.4399 KOps/s	$\color{#35bf28}+3.54\%$
test_items_nested_locked	0.7288ms	0.3963ms	2.5233 KOps/s	2.4552 KOps/s	$\color{#35bf28}+2.77\%$
test_items_nested_leaf	0.1209ms	77.6177μs	12.8837 KOps/s	12.6881 KOps/s	$\color{#35bf28}+1.54\%$
test_items_stack_nested	0.5459ms	0.3998ms	2.5014 KOps/s	2.4422 KOps/s	$\color{#35bf28}+2.42\%$
test_items_stack_nested_leaf	0.1518ms	78.9272μs	12.6699 KOps/s	12.2968 KOps/s	$\color{#35bf28}+3.03\%$
test_items_stack_nested_locked	0.6043ms	0.3990ms	2.5064 KOps/s	2.4353 KOps/s	$\color{#35bf28}+2.92\%$
test_keys	28.3730μs	3.4943μs	286.1811 KOps/s	287.2819 KOps/s	$\color{#d91a1a}-0.38\%$
test_keys_nested	0.2928ms	0.1650ms	6.0602 KOps/s	5.9116 KOps/s	$\color{#35bf28}+2.51\%$
test_keys_nested_locked	0.6882ms	0.1696ms	5.8965 KOps/s	5.6655 KOps/s	$\color{#35bf28}+4.08\%$
test_keys_nested_leaf	0.2618ms	0.1431ms	6.9906 KOps/s	6.7656 KOps/s	$\color{#35bf28}+3.32\%$
test_keys_stack_nested	0.2618ms	0.1658ms	6.0307 KOps/s	6.0366 KOps/s	$\color{#d91a1a}-0.10\%$
test_keys_stack_nested_leaf	0.2190ms	0.1431ms	6.9857 KOps/s	7.0435 KOps/s	$\color{#d91a1a}-0.82\%$
test_keys_stack_nested_locked	0.2737ms	0.1709ms	5.8503 KOps/s	5.8339 KOps/s	$\color{#35bf28}+0.28\%$
test_values	9.9846μs	1.0267μs	974.0270 KOps/s	950.8590 KOps/s	$\color{#35bf28}+2.44\%$
test_values_nested	0.1225ms	62.4896μs	16.0027 KOps/s	15.8927 KOps/s	$\color{#35bf28}+0.69\%$
test_values_nested_locked	0.1135ms	62.4497μs	16.0129 KOps/s	15.7265 KOps/s	$\color{#35bf28}+1.82\%$
test_values_nested_leaf	0.1239ms	71.1562μs	14.0536 KOps/s	13.7923 KOps/s	$\color{#35bf28}+1.89\%$
test_values_stack_nested	0.1184ms	62.6220μs	15.9688 KOps/s	15.3605 KOps/s	$\color{#35bf28}+3.96\%$
test_values_stack_nested_leaf	0.1250ms	71.7296μs	13.9412 KOps/s	14.0569 KOps/s	$\color{#d91a1a}-0.82\%$
test_values_stack_nested_locked	0.1157ms	63.0234μs	15.8671 KOps/s	15.3180 KOps/s	$\color{#35bf28}+3.59\%$
test_membership	5.2881μs	0.6915μs	1.4461 MOps/s	1.1304 MOps/s	$\textbf{\color{#35bf28}+27.93\%}$
test_membership_nested	24.1050μs	2.9025μs	344.5357 KOps/s	340.9579 KOps/s	$\color{#35bf28}+1.05\%$
test_membership_nested_leaf	43.4110μs	2.9194μs	342.5396 KOps/s	342.3902 KOps/s	$\color{#35bf28}+0.04\%$
test_membership_stacked_nested	16.5110μs	2.8670μs	348.8025 KOps/s	339.4935 KOps/s	$\color{#35bf28}+2.74\%$
test_membership_stacked_nested_leaf	39.6540μs	2.8821μs	346.9653 KOps/s	340.6511 KOps/s	$\color{#35bf28}+1.85\%$
test_membership_nested_last	26.3990μs	4.3044μs	232.3202 KOps/s	226.6151 KOps/s	$\color{#35bf28}+2.52\%$
test_membership_nested_leaf_last	48.6010μs	4.3479μs	229.9974 KOps/s	225.6689 KOps/s	$\color{#35bf28}+1.92\%$
test_membership_stacked_nested_last	42.2480μs	4.2741μs	233.9692 KOps/s	73.8938 KOps/s	$\textbf{\color{#35bf28}+216.63\%}$
test_membership_stacked_nested_leaf_last	23.8750μs	4.3086μs	232.0963 KOps/s	74.5746 KOps/s	$\textbf{\color{#35bf28}+211.23\%}$
test_nested_getleaf	57.4670μs	10.5724μs	94.5856 KOps/s	92.2604 KOps/s	$\color{#35bf28}+2.52\%$
test_nested_get	54.4820μs	10.0525μs	99.4779 KOps/s	97.0814 KOps/s	$\color{#35bf28}+2.47\%$
test_stacked_getleaf	37.3100μs	10.5969μs	94.3674 KOps/s	93.4605 KOps/s	$\color{#35bf28}+0.97\%$
test_stacked_get	46.1570μs	10.0403μs	99.5984 KOps/s	98.2228 KOps/s	$\color{#35bf28}+1.40\%$
test_nested_getitemleaf	54.9830μs	11.2386μs	88.9791 KOps/s	87.8680 KOps/s	$\color{#35bf28}+1.26\%$
test_nested_getitem	46.6170μs	10.6394μs	93.9905 KOps/s	92.3678 KOps/s	$\color{#35bf28}+1.76\%$
test_stacked_getitemleaf	56.4360μs	11.2025μs	89.2661 KOps/s	88.2837 KOps/s	$\color{#35bf28}+1.11\%$
test_stacked_getitem	38.4020μs	10.6039μs	94.3052 KOps/s	92.0940 KOps/s	$\color{#35bf28}+2.40\%$
test_lock_nested	1.1256ms	0.4571ms	2.1875 KOps/s	1.7835 KOps/s	$\textbf{\color{#35bf28}+22.66\%}$
test_lock_stack_nested	0.5093ms	0.4237ms	2.3603 KOps/s	2.3935 KOps/s	$\color{#d91a1a}-1.39\%$
test_unlock_nested	0.9275ms	0.3756ms	2.6625 KOps/s	2.6387 KOps/s	$\color{#35bf28}+0.90\%$
test_unlock_stack_nested	0.5263ms	0.3444ms	2.9033 KOps/s	3.0112 KOps/s	$\color{#d91a1a}-3.58\%$
test_flatten_speed	0.1866ms	0.1011ms	9.8908 KOps/s	9.7920 KOps/s	$\color{#35bf28}+1.01\%$
test_unflatten_speed	0.7095ms	0.5126ms	1.9509 KOps/s	1.9054 KOps/s	$\color{#35bf28}+2.39\%$
test_common_ops	1.7894ms	0.7483ms	1.3364 KOps/s	1.2984 KOps/s	$\color{#35bf28}+2.93\%$
test_creation	45.2950μs	2.4877μs	401.9721 KOps/s	376.0471 KOps/s	$\textbf{\color{#35bf28}+6.89\%}$
test_creation_empty	32.8920μs	9.5150μs	105.0974 KOps/s	103.1581 KOps/s	$\color{#35bf28}+1.88\%$
test_creation_nested_1	38.3410μs	12.4624μs	80.2413 KOps/s	80.1200 KOps/s	$\color{#35bf28}+0.15\%$
test_creation_nested_2	37.7710μs	17.0233μs	58.7431 KOps/s	57.5631 KOps/s	$\color{#35bf28}+2.05\%$
test_clone	74.5700μs	13.4062μs	74.5924 KOps/s	73.3338 KOps/s	$\color{#35bf28}+1.72\%$
test_getitem[int]	1.2554ms	12.9086μs	77.4679 KOps/s	76.8288 KOps/s	$\color{#35bf28}+0.83\%$
test_getitem[slice_int]	0.1387ms	24.3230μs	41.1133 KOps/s	40.1741 KOps/s	$\color{#35bf28}+2.34\%$
test_getitem[range]	0.1724ms	48.8903μs	20.4540 KOps/s	20.7079 KOps/s	$\color{#d91a1a}-1.23\%$
test_getitem[tuple]	0.1335ms	19.9061μs	50.2358 KOps/s	47.6578 KOps/s	$\textbf{\color{#35bf28}+5.41\%}$
test_getitem[list]	0.1960ms	43.2736μs	23.1088 KOps/s	22.7659 KOps/s	$\color{#35bf28}+1.51\%$
test_setitem_dim[int]	57.5170μs	26.6233μs	37.5611 KOps/s	37.0612 KOps/s	$\color{#35bf28}+1.35\%$
test_setitem_dim[slice_int]	77.7250μs	52.0196μs	19.2235 KOps/s	18.9459 KOps/s	$\color{#35bf28}+1.47\%$
test_setitem_dim[range]	0.1231ms	75.4533μs	13.2532 KOps/s	13.2333 KOps/s	$\color{#35bf28}+0.15\%$
test_setitem_dim[tuple]	71.2730μs	40.6808μs	24.5816 KOps/s	23.5941 KOps/s	$\color{#35bf28}+4.19\%$
test_setitem	0.1217ms	19.2944μs	51.8285 KOps/s	51.0088 KOps/s	$\color{#35bf28}+1.61\%$
test_set	86.3020μs	18.2843μs	54.6916 KOps/s	53.7685 KOps/s	$\color{#35bf28}+1.72\%$
test_set_shared	3.6403ms	0.1728ms	5.7883 KOps/s	5.8163 KOps/s	$\color{#d91a1a}-0.48\%$
test_update	0.1489ms	20.1705μs	49.5773 KOps/s	48.6470 KOps/s	$\color{#35bf28}+1.91\%$
test_update_nested	99.6770μs	30.1542μs	33.1629 KOps/s	32.0276 KOps/s	$\color{#35bf28}+3.54\%$
test_update__nested	0.6457ms	33.4576μs	29.8885 KOps/s	28.5921 KOps/s	$\color{#35bf28}+4.53\%$
test_set_nested	89.0560μs	20.5440μs	48.6761 KOps/s	47.0010 KOps/s	$\color{#35bf28}+3.56\%$
test_set_nested_new	78.0660μs	25.1113μs	39.8227 KOps/s	38.4947 KOps/s	$\color{#35bf28}+3.45\%$
test_select	0.1525ms	41.1227μs	24.3175 KOps/s	23.6860 KOps/s	$\color{#35bf28}+2.67\%$
test_select_nested	0.1199ms	62.8898μs	15.9008 KOps/s	15.2421 KOps/s	$\color{#35bf28}+4.32\%$
test_exclude_nested	0.1781ms	81.3378μs	12.2944 KOps/s	11.8701 KOps/s	$\color{#35bf28}+3.57\%$
test_empty[True]	0.6171ms	0.4059ms	2.4634 KOps/s	2.3691 KOps/s	$\color{#35bf28}+3.98\%$
test_empty[False]	9.6980μs	1.3357μs	748.6649 KOps/s	720.4520 KOps/s	$\color{#35bf28}+3.92\%$
test_unbind_speed	1.0582ms	0.2793ms	3.5809 KOps/s	3.6639 KOps/s	$\color{#d91a1a}-2.26\%$
test_unbind_speed_stack0	0.4076ms	0.2694ms	3.7125 KOps/s	3.8906 KOps/s	$\color{#d91a1a}-4.58\%$
test_unbind_speed_stack1	0.1055s	0.7937ms	1.2599 KOps/s	1.4065 KOps/s	$\textbf{\color{#d91a1a}-10.42\%}$
test_split	1.7572ms	1.5753ms	634.7836 Ops/s	559.5763 Ops/s	$\textbf{\color{#35bf28}+13.44\%}$
test_chunk	0.1040s	1.8954ms	527.5966 Ops/s	560.5367 Ops/s	$\textbf{\color{#d91a1a}-5.88\%}$
test_consolidate_njt[False-None]	8.7183ms	8.2635ms	121.0135 Ops/s	120.4005 Ops/s	$\color{#35bf28}+0.51\%$
test_creation[device0]	0.4692ms	91.5584μs	10.9220 KOps/s	11.0487 KOps/s	$\color{#d91a1a}-1.15\%$
test_creation_from_tensor	3.4285ms	94.7100μs	10.5585 KOps/s	10.5379 KOps/s	$\color{#35bf28}+0.20\%$
test_add_one[memmap_tensor0]	0.1386ms	4.9219μs	203.1727 KOps/s	202.9119 KOps/s	$\color{#35bf28}+0.13\%$
test_contiguous[memmap_tensor0]	10.8910μs	0.5101μs	1.9605 MOps/s	1.9437 MOps/s	$\color{#35bf28}+0.87\%$
test_stack[memmap_tensor0]	39.9150μs	3.4439μs	290.3700 KOps/s	280.6249 KOps/s	$\color{#35bf28}+3.47\%$
test_memmaptd_index	0.8969ms	0.2332ms	4.2883 KOps/s	4.2834 KOps/s	$\color{#35bf28}+0.11\%$
test_memmaptd_index_astensor	0.5782ms	0.3210ms	3.1157 KOps/s	3.1266 KOps/s	$\color{#d91a1a}-0.35\%$
test_memmaptd_index_op	1.1240ms	0.5352ms	1.8685 KOps/s	1.8270 KOps/s	$\color{#35bf28}+2.27\%$
test_serialize_model	0.1309s	0.1183s	8.4560 Ops/s	8.5155 Ops/s	$\color{#d91a1a}-0.70\%$
test_serialize_model_pickle	0.4495s	0.4005s	2.4969 Ops/s	2.5922 Ops/s	$\color{#d91a1a}-3.68\%$
test_serialize_weights	0.1221s	0.1140s	8.7687 Ops/s	8.7251 Ops/s	$\color{#35bf28}+0.50\%$
test_serialize_weights_returnearly	0.2541s	0.1750s	5.7138 Ops/s	6.2655 Ops/s	$\textbf{\color{#d91a1a}-8.81\%}$
test_serialize_weights_pickle	0.6038s	0.4360s	2.2936 Ops/s	2.4338 Ops/s	$\textbf{\color{#d91a1a}-5.76\%}$
test_serialize_weights_filesystem	0.1498s	0.1399s	7.1483 Ops/s	7.0174 Ops/s	$\color{#35bf28}+1.86\%$
test_serialize_model_filesystem	0.1620s	0.1444s	6.9258 Ops/s	6.2658 Ops/s	$\textbf{\color{#35bf28}+10.53\%}$
test_reshape_pytree	64.8910μs	26.1525μs	38.2372 KOps/s	37.2711 KOps/s	$\color{#35bf28}+2.59\%$
test_reshape_td	0.1133ms	33.5616μs	29.7960 KOps/s	30.4184 KOps/s	$\color{#d91a1a}-2.05\%$
test_view_pytree	58.9500μs	26.0084μs	38.4491 KOps/s	37.2420 KOps/s	$\color{#35bf28}+3.24\%$
test_view_td	79.1170μs	38.7501μs	25.8064 KOps/s	25.3651 KOps/s	$\color{#35bf28}+1.74\%$
test_unbind_pytree	69.7810μs	29.2032μs	34.2428 KOps/s	33.5226 KOps/s	$\color{#35bf28}+2.15\%$
test_unbind_td	0.2845ms	39.7299μs	25.1700 KOps/s	24.9380 KOps/s	$\color{#35bf28}+0.93\%$
test_split_pytree	71.8950μs	28.9900μs	34.4947 KOps/s	32.8617 KOps/s	$\color{#35bf28}+4.97\%$
test_split_td	0.5508ms	44.3035μs	22.5716 KOps/s	21.9001 KOps/s	$\color{#35bf28}+3.07\%$
test_add_pytree	77.5950μs	34.8440μs	28.6993 KOps/s	27.8110 KOps/s	$\color{#35bf28}+3.19\%$
test_add_td	0.1162ms	49.2040μs	20.3236 KOps/s	19.2069 KOps/s	$\textbf{\color{#35bf28}+5.81\%}$
test_compile_add_one_nested[tensordict-compile]	0.1474ms	62.3955μs	16.0268 KOps/s	15.3222 KOps/s	$\color{#35bf28}+4.60\%$
test_compile_add_one_nested[tensordict-eager]	0.3970ms	0.1751ms	5.7105 KOps/s	5.6996 KOps/s	$\color{#35bf28}+0.19\%$
test_compile_add_one_nested[pytree-compile]	0.1434ms	46.0805μs	21.7011 KOps/s	21.1912 KOps/s	$\color{#35bf28}+2.41\%$
test_compile_add_one_nested[pytree-eager]	0.2623ms	0.1196ms	8.3620 KOps/s	8.2963 KOps/s	$\color{#35bf28}+0.79\%$
test_compile_copy_nested[tensordict-compile]	0.1123ms	27.1519μs	36.8298 KOps/s	35.7910 KOps/s	$\color{#35bf28}+2.90\%$
test_compile_copy_nested[tensordict-eager]	0.1210ms	58.7892μs	17.0099 KOps/s	16.8959 KOps/s	$\color{#35bf28}+0.67\%$
test_compile_copy_nested[pytree-compile]	0.1504ms	79.4527μs	12.5861 KOps/s	12.4355 KOps/s	$\color{#35bf28}+1.21\%$
test_compile_copy_nested[pytree-eager]	0.1403ms	67.7769μs	14.7543 KOps/s	14.5783 KOps/s	$\color{#35bf28}+1.21\%$
test_compile_add_one_flat[tensordict-compile]	0.1833ms	0.1048ms	9.5459 KOps/s	9.3733 KOps/s	$\color{#35bf28}+1.84\%$
test_compile_add_one_flat[tensordict-eager]	0.4750ms	0.2139ms	4.6748 KOps/s	4.6081 KOps/s	$\color{#35bf28}+1.45\%$
test_compile_add_one_flat[tensorclass-compile]	0.1361ms	45.2505μs	22.0992 KOps/s	21.6125 KOps/s	$\color{#35bf28}+2.25\%$
test_compile_add_one_flat[tensorclass-eager]	0.5069ms	67.5564μs	14.8025 KOps/s	14.8204 KOps/s	$\color{#d91a1a}-0.12\%$
test_compile_add_one_flat[pytree-compile]	0.2002ms	0.1023ms	9.7715 KOps/s	9.6031 KOps/s	$\color{#35bf28}+1.75\%$
test_compile_add_one_flat[pytree-eager]	0.3221ms	0.2016ms	4.9602 KOps/s	5.0305 KOps/s	$\color{#d91a1a}-1.40\%$
test_compile_add_self_flat[tensordict-eager]	0.4692ms	0.2334ms	4.2838 KOps/s	4.2502 KOps/s	$\color{#35bf28}+0.79\%$
test_compile_add_self_flat[tensordict-compile]	0.1910ms	0.1050ms	9.5280 KOps/s	9.4442 KOps/s	$\color{#35bf28}+0.89\%$
test_compile_add_self_flat[tensorclass-eager]	0.1627ms	63.5928μs	15.7250 KOps/s	15.5060 KOps/s	$\color{#35bf28}+1.41\%$
test_compile_add_self_flat[tensorclass-compile]	0.1076ms	47.9943μs	20.8358 KOps/s	20.8957 KOps/s	$\color{#d91a1a}-0.29\%$
test_compile_add_self_flat[pytree-eager]	0.3140ms	0.1617ms	6.1841 KOps/s	6.3367 KOps/s	$\color{#d91a1a}-2.41\%$
test_compile_add_self_flat[pytree-compile]	0.2279ms	0.1032ms	9.6925 KOps/s	9.7482 KOps/s	$\color{#d91a1a}-0.57\%$
test_compile_copy_flat[tensordict-compile]	68.3280μs	21.8203μs	45.8290 KOps/s	41.7296 KOps/s	$\textbf{\color{#35bf28}+9.82\%}$
test_compile_copy_flat[tensordict-eager]	0.1484ms	67.2744μs	14.8645 KOps/s	15.0009 KOps/s	$\color{#d91a1a}-0.91\%$
test_compile_copy_flat[pytree-compile]	0.1972ms	79.7585μs	12.5378 KOps/s	12.5338 KOps/s	$\color{#35bf28}+0.03\%$
test_compile_copy_flat[pytree-eager]	0.1574ms	68.0473μs	14.6957 KOps/s	14.7513 KOps/s	$\color{#d91a1a}-0.38\%$
test_compile_assign_and_add[tensordict-compile]	0.3417ms	0.2091ms	4.7831 KOps/s	4.8609 KOps/s	$\color{#d91a1a}-1.60\%$
test_compile_assign_and_add[tensordict-eager]	2.4661ms	1.3445ms	743.7948 Ops/s	759.1004 Ops/s	$\color{#d91a1a}-2.02\%$
test_compile_assign_and_add[pytree-compile]	0.3110ms	0.2075ms	4.8188 KOps/s	4.9939 KOps/s	$\color{#d91a1a}-3.51\%$
test_compile_assign_and_add[pytree-eager]	1.3878ms	0.7796ms	1.2828 KOps/s	1.2755 KOps/s	$\color{#35bf28}+0.57\%$
test_compile_assign_and_add_stack[compile]	0.5299ms	0.4503ms	2.2209 KOps/s	2.2230 KOps/s	$\color{#d91a1a}-0.09\%$
test_compile_assign_and_add_stack[eager]	3.6112ms	2.5841ms	386.9759 Ops/s	391.7029 Ops/s	$\color{#d91a1a}-1.21\%$
test_compile_indexing[tensor-tensordict-compile]	90.9100μs	36.0378μs	27.7487 KOps/s	27.0949 KOps/s	$\color{#35bf28}+2.41\%$
test_compile_indexing[tensor-tensordict-eager]	0.5484ms	33.3181μs	30.0138 KOps/s	29.4987 KOps/s	$\color{#35bf28}+1.75\%$
test_compile_indexing[tensor-tensorclass-compile]	0.1011ms	29.0904μs	34.3757 KOps/s	32.0626 KOps/s	$\textbf{\color{#35bf28}+7.21\%}$
test_compile_indexing[tensor-tensorclass-eager]	69.7910μs	23.4624μs	42.6213 KOps/s	42.1796 KOps/s	$\color{#35bf28}+1.05\%$
test_compile_indexing[tensor-pytree-compile]	99.3960μs	30.3871μs	32.9087 KOps/s	32.1964 KOps/s	$\color{#35bf28}+2.21\%$
test_compile_indexing[tensor-pytree-eager]	91.5920μs	23.7806μs	42.0510 KOps/s	42.1952 KOps/s	$\color{#d91a1a}-0.34\%$
test_compile_indexing[slice-tensordict-compile]	0.1052ms	52.1408μs	19.1788 KOps/s	19.2238 KOps/s	$\color{#d91a1a}-0.23\%$
test_compile_indexing[slice-tensordict-eager]	0.5660ms	20.2335μs	49.4230 KOps/s	48.9278 KOps/s	$\color{#35bf28}+1.01\%$
test_compile_indexing[slice-tensorclass-compile]	0.1134ms	44.6351μs	22.4039 KOps/s	21.7814 KOps/s	$\color{#35bf28}+2.86\%$
test_compile_indexing[slice-tensorclass-eager]	83.4850μs	18.6986μs	53.4799 KOps/s	52.9760 KOps/s	$\color{#35bf28}+0.95\%$
test_compile_indexing[slice-pytree-compile]	0.1049ms	45.7825μs	21.8424 KOps/s	21.6798 KOps/s	$\color{#35bf28}+0.75\%$
test_compile_indexing[slice-pytree-eager]	72.6360μs	18.6654μs	53.5751 KOps/s	52.9815 KOps/s	$\color{#35bf28}+1.12\%$
test_compile_indexing[int-tensordict-compile]	0.1097ms	53.3576μs	18.7415 KOps/s	18.5446 KOps/s	$\color{#35bf28}+1.06\%$
test_compile_indexing[int-tensordict-eager]	0.9487ms	19.8450μs	50.3905 KOps/s	49.5054 KOps/s	$\color{#35bf28}+1.79\%$
test_compile_indexing[int-tensorclass-compile]	0.1556ms	45.2566μs	22.0962 KOps/s	21.4267 KOps/s	$\color{#35bf28}+3.12\%$
test_compile_indexing[int-tensorclass-eager]	78.2360μs	18.4711μs	54.1386 KOps/s	52.9385 KOps/s	$\color{#35bf28}+2.27\%$
test_compile_indexing[int-pytree-compile]	0.1253ms	44.8537μs	22.2947 KOps/s	21.3760 KOps/s	$\color{#35bf28}+4.30\%$
test_compile_indexing[int-pytree-eager]	0.8249ms	18.4554μs	54.1848 KOps/s	53.3213 KOps/s	$\color{#35bf28}+1.62\%$
test_mod_add[eager]	99.3060μs	33.5517μs	29.8047 KOps/s	29.8641 KOps/s	$\color{#d91a1a}-0.20\%$
test_mod_add[compile]	0.1234ms	48.5003μs	20.6184 KOps/s	20.6306 KOps/s	$\color{#d91a1a}-0.06\%$
test_mod_add[compile-overhead]	0.1203ms	49.2114μs	20.3205 KOps/s	20.7275 KOps/s	$\color{#d91a1a}-1.96\%$
test_mod_wrap[eager]	0.3500ms	0.2207ms	4.5311 KOps/s	4.4463 KOps/s	$\color{#35bf28}+1.91\%$
test_mod_wrap[compile]	0.3997ms	0.2111ms	4.7381 KOps/s	4.5772 KOps/s	$\color{#35bf28}+3.52\%$
test_mod_wrap[compile-overhead]	0.3817ms	0.2070ms	4.8315 KOps/s	4.7781 KOps/s	$\color{#35bf28}+1.12\%$
test_mod_wrap_and_backward[eager]	20.7550ms	12.6583ms	78.9995 Ops/s	84.2877 Ops/s	$\textbf{\color{#d91a1a}-6.27\%}$
test_mod_wrap_and_backward[compile]	17.8635ms	12.6071ms	79.3203 Ops/s	87.4856 Ops/s	$\textbf{\color{#d91a1a}-9.33\%}$
test_mod_wrap_and_backward[compile-overhead]	16.8940ms	13.3364ms	74.9827 Ops/s	73.4562 Ops/s	$\color{#35bf28}+2.08\%$
test_seq_add[eager]	0.2425ms	0.1105ms	9.0523 KOps/s	8.7521 KOps/s	$\color{#35bf28}+3.43\%$
test_seq_add[compile]	0.1410ms	63.6004μs	15.7232 KOps/s	15.5645 KOps/s	$\color{#35bf28}+1.02\%$
test_seq_add[compile-overhead]	0.1279ms	60.6360μs	16.4919 KOps/s	16.0288 KOps/s	$\color{#35bf28}+2.89\%$
test_seq_wrap[eager]	0.5690ms	0.4356ms	2.2957 KOps/s	2.2328 KOps/s	$\color{#35bf28}+2.82\%$
test_seq_wrap[compile]	0.4583ms	0.2299ms	4.3496 KOps/s	4.2251 KOps/s	$\color{#35bf28}+2.95\%$
test_seq_wrap[compile-overhead]	0.4277ms	0.2270ms	4.4062 KOps/s	4.2615 KOps/s	$\color{#35bf28}+3.40\%$
test_func_call_runtime[False-eager]	0.8923ms	0.5482ms	1.8241 KOps/s	1.7428 KOps/s	$\color{#35bf28}+4.66\%$
test_func_call_runtime[False-compile]	0.5259ms	0.4222ms	2.3687 KOps/s	2.3252 KOps/s	$\color{#35bf28}+1.87\%$
test_func_call_runtime[False-compile-overhead]	0.8112ms	0.4288ms	2.3320 KOps/s	2.3469 KOps/s	$\color{#d91a1a}-0.64\%$
test_func_call_runtime[True-eager]	1.0889ms	0.7721ms	1.2952 KOps/s	1.2662 KOps/s	$\color{#35bf28}+2.29\%$
test_func_call_runtime[True-compile]	0.6059ms	0.4583ms	2.1821 KOps/s	2.1316 KOps/s	$\color{#35bf28}+2.37\%$
test_func_call_runtime[True-compile-overhead]	0.6477ms	0.4583ms	2.1822 KOps/s	2.1326 KOps/s	$\color{#35bf28}+2.33\%$
test_func_call_cm_runtime[False-eager]	1.2497ms	0.5431ms	1.8413 KOps/s	1.7516 KOps/s	$\textbf{\color{#35bf28}+5.13\%}$
test_func_call_cm_runtime[False-compile]	0.5414ms	0.4204ms	2.3786 KOps/s	2.3409 KOps/s	$\color{#35bf28}+1.61\%$
test_func_call_cm_runtime[False-compile-overhead]	0.5443ms	0.4209ms	2.3758 KOps/s	2.3275 KOps/s	$\color{#35bf28}+2.07\%$
test_func_call_cm_runtime[True-eager]	1.0508ms	0.9152ms	1.0927 KOps/s	1.0742 KOps/s	$\color{#35bf28}+1.72\%$
test_func_call_cm_runtime[True-compile]	0.5865ms	0.4867ms	2.0546 KOps/s	1.9897 KOps/s	$\color{#35bf28}+3.26\%$
test_func_call_cm_runtime[True-compile-overhead]	0.6536ms	0.4843ms	2.0650 KOps/s	1.9971 KOps/s	$\color{#35bf28}+3.40\%$
test_vmap_func_call_cm_runtime[eager]	2.4372ms	1.9095ms	523.6965 Ops/s	508.5175 Ops/s	$\color{#35bf28}+2.98\%$
test_vmap_func_call_cm_runtime[compile]	0.9127ms	0.5228ms	1.9129 KOps/s	1.9040 KOps/s	$\color{#35bf28}+0.47\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.8897ms	0.5294ms	1.8889 KOps/s	1.9032 KOps/s	$\color{#d91a1a}-0.75\%$
test_distributed	0.2703ms	0.1247ms	8.0221 KOps/s	7.7828 KOps/s	$\color{#35bf28}+3.07\%$
test_tdmodule	69.3200μs	24.9635μs	40.0585 KOps/s	39.2775 KOps/s	$\color{#35bf28}+1.99\%$
test_tdmodule_dispatch	77.9350μs	46.1653μs	21.6613 KOps/s	21.4891 KOps/s	$\color{#35bf28}+0.80\%$
test_tdseq	45.1050μs	27.1566μs	36.8235 KOps/s	35.1758 KOps/s	$\color{#35bf28}+4.68\%$
test_tdseq_dispatch	76.8740μs	49.8152μs	20.0742 KOps/s	18.8677 KOps/s	$\textbf{\color{#35bf28}+6.39\%}$
test_instantiation_functorch	2.6785ms	1.5525ms	644.1021 Ops/s	644.1201 Ops/s	$-0.00\%$
test_exec_functorch	0.3061ms	0.1820ms	5.4939 KOps/s	5.4136 KOps/s	$\color{#35bf28}+1.48\%$
test_exec_functional_call	0.3117ms	0.1701ms	5.8781 KOps/s	5.5181 KOps/s	$\textbf{\color{#35bf28}+6.52\%}$
test_exec_td_decorator	0.4731ms	0.2299ms	4.3505 KOps/s	4.2465 KOps/s	$\color{#35bf28}+2.45\%$
test_vmap_mlp_speed_decorator[True-True]	0.9658ms	0.6533ms	1.5307 KOps/s	1.5454 KOps/s	$\color{#d91a1a}-0.95\%$
test_vmap_mlp_speed_decorator[True-False]	1.0666ms	0.6541ms	1.5288 KOps/s	1.5341 KOps/s	$\color{#d91a1a}-0.35\%$
test_vmap_mlp_speed_decorator[False-True]	0.8438ms	0.5383ms	1.8576 KOps/s	1.8715 KOps/s	$\color{#d91a1a}-0.75\%$
test_vmap_mlp_speed_decorator[False-False]	0.9680ms	0.5359ms	1.8660 KOps/s	1.8709 KOps/s	$\color{#d91a1a}-0.26\%$
test_to_module_speed[True]	1.6913ms	1.3340ms	749.6008 Ops/s	729.9320 Ops/s	$\color{#35bf28}+2.69\%$
test_to_module_speed[False]	1.8258ms	1.2992ms	769.6971 Ops/s	754.4421 Ops/s	$\color{#35bf28}+2.02\%$
test_tc_init	96.4900μs	45.8129μs	21.8279 KOps/s	22.4815 KOps/s	$\color{#d91a1a}-2.91\%$
test_tc_init_nested	0.1501ms	91.6613μs	10.9097 KOps/s	11.2932 KOps/s	$\color{#d91a1a}-3.40\%$
test_tc_first_layer_tensor	38.5010μs	1.5800μs	632.9303 KOps/s	630.9741 KOps/s	$\color{#35bf28}+0.31\%$
test_tc_first_layer_nontensor	27.0910μs	5.1402μs	194.5434 KOps/s	212.4574 KOps/s	$\textbf{\color{#d91a1a}-8.43\%}$
test_tc_second_layer_tensor	24.3560μs	2.9292μs	341.3922 KOps/s	347.6049 KOps/s	$\color{#d91a1a}-1.79\%$
test_tc_second_layer_nontensor	46.3770μs	6.2856μs	159.0948 KOps/s	166.5600 KOps/s	$\color{#d91a1a}-4.48\%$
test_unbind	0.2363s	13.4628ms	74.2787 Ops/s	77.5577 Ops/s	$\color{#d91a1a}-4.23\%$
test_full_like	18.0313ms	11.8828ms	84.1552 Ops/s	82.1803 Ops/s	$\color{#35bf28}+2.40\%$
test_zeros_like	11.2453ms	7.1377ms	140.1009 Ops/s	126.1127 Ops/s	$\textbf{\color{#35bf28}+11.09\%}$
test_ones_like	12.1094ms	7.9260ms	126.1663 Ops/s	130.3149 Ops/s	$\color{#d91a1a}-3.18\%$
test_clone	13.2835ms	9.4340ms	105.9993 Ops/s	107.5229 Ops/s	$\color{#d91a1a}-1.42\%$
test_squeeze	58.4590μs	11.8234μs	84.5777 KOps/s	81.5354 KOps/s	$\color{#35bf28}+3.73\%$
test_unsqueeze	0.2969ms	91.6152μs	10.9152 KOps/s	11.0472 KOps/s	$\color{#d91a1a}-1.19\%$
test_split	0.3974ms	0.1999ms	5.0030 KOps/s	5.0856 KOps/s	$\color{#d91a1a}-1.62\%$
test_permute	0.2795ms	0.2010ms	4.9754 KOps/s	5.0038 KOps/s	$\color{#d91a1a}-0.57\%$
test_stack	29.9623ms	24.9111ms	40.1427 Ops/s	40.3167 Ops/s	$\color{#d91a1a}-0.43\%$
test_cat	29.3022ms	24.5323ms	40.7626 Ops/s	40.9552 Ops/s	$\color{#d91a1a}-0.47\%$

[github-actions]

$\color{#D29922}\textsf{\Large⚠\kern{0.2cm}\normalsize Warning}$ Result of GPU Benchmark Tests

Total Benchmarks: 229. Improved: $\large\color{#35bf28}33$. Worsened: $\large\color{#d91a1a}6$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	36.7700μs	12.1604μs	82.2343 KOps/s	77.2928 KOps/s	$\textbf{\color{#35bf28}+6.39\%}$
test_plain_set_stack_nested	35.1810μs	12.3944μs	80.6819 KOps/s	75.6935 KOps/s	$\textbf{\color{#35bf28}+6.59\%}$
test_plain_set_nested_inplace	42.2810μs	13.2189μs	75.6493 KOps/s	70.9914 KOps/s	$\textbf{\color{#35bf28}+6.56\%}$
test_plain_set_stack_nested_inplace	50.1610μs	13.3467μs	74.9249 KOps/s	70.7601 KOps/s	$\textbf{\color{#35bf28}+5.89\%}$
test_items	30.2510μs	2.8966μs	345.2309 KOps/s	342.5848 KOps/s	$\color{#35bf28}+0.77\%$
test_items_nested	0.4402ms	0.3627ms	2.7567 KOps/s	2.7295 KOps/s	$\color{#35bf28}+1.00\%$
test_items_nested_locked	0.4509ms	0.3656ms	2.7355 KOps/s	2.7308 KOps/s	$\color{#35bf28}+0.17\%$
test_items_nested_leaf	87.2210μs	59.1260μs	16.9130 KOps/s	17.0656 KOps/s	$\color{#d91a1a}-0.89\%$
test_items_stack_nested	0.3990ms	0.3694ms	2.7070 KOps/s	2.6723 KOps/s	$\color{#35bf28}+1.30\%$
test_items_stack_nested_leaf	0.1242ms	60.7118μs	16.4713 KOps/s	16.4726 KOps/s	$-0.01\%$
test_items_stack_nested_locked	0.4415ms	0.3658ms	2.7334 KOps/s	2.7083 KOps/s	$\color{#35bf28}+0.93\%$
test_keys	41.8100μs	3.7139μs	269.2615 KOps/s	282.7091 KOps/s	$\color{#d91a1a}-4.76\%$
test_keys_nested	0.1269ms	87.9966μs	11.3641 KOps/s	11.2808 KOps/s	$\color{#35bf28}+0.74\%$
test_keys_nested_locked	0.8254ms	94.0174μs	10.6363 KOps/s	10.5535 KOps/s	$\color{#35bf28}+0.78\%$
test_keys_nested_leaf	0.1627ms	79.0679μs	12.6474 KOps/s	12.5033 KOps/s	$\color{#35bf28}+1.15\%$
test_keys_stack_nested	0.1137ms	89.6862μs	11.1500 KOps/s	11.0618 KOps/s	$\color{#35bf28}+0.80\%$
test_keys_stack_nested_leaf	0.1237ms	80.5663μs	12.4121 KOps/s	12.1894 KOps/s	$\color{#35bf28}+1.83\%$
test_keys_stack_nested_locked	0.1319ms	94.7110μs	10.5584 KOps/s	10.3595 KOps/s	$\color{#35bf28}+1.92\%$
test_values	5.5235μs	0.8525μs	1.1730 MOps/s	1.1747 MOps/s	$\color{#d91a1a}-0.14\%$
test_values_nested	0.1193ms	37.4863μs	26.6764 KOps/s	26.5271 KOps/s	$\color{#35bf28}+0.56\%$
test_values_nested_locked	74.8110μs	38.7042μs	25.8370 KOps/s	25.3997 KOps/s	$\color{#35bf28}+1.72\%$
test_values_nested_leaf	64.8510μs	41.6562μs	24.0060 KOps/s	23.7377 KOps/s	$\color{#35bf28}+1.13\%$
test_values_stack_nested	61.7910μs	38.0347μs	26.2918 KOps/s	26.0260 KOps/s	$\color{#35bf28}+1.02\%$
test_values_stack_nested_leaf	0.1061ms	42.6442μs	23.4499 KOps/s	23.3623 KOps/s	$\color{#35bf28}+0.37\%$
test_values_stack_nested_locked	68.9620μs	39.2769μs	25.4603 KOps/s	25.2499 KOps/s	$\color{#35bf28}+0.83\%$
test_membership	1.5646μs	0.5119μs	1.9534 MOps/s	1.9019 MOps/s	$\color{#35bf28}+2.71\%$
test_membership_nested	20.5805μs	2.0277μs	493.1772 KOps/s	497.2547 KOps/s	$\color{#d91a1a}-0.82\%$
test_membership_nested_leaf	21.8205μs	2.0286μs	492.9475 KOps/s	493.7240 KOps/s	$\color{#d91a1a}-0.16\%$
test_membership_stacked_nested	35.5300μs	2.0607μs	485.2644 KOps/s	480.2739 KOps/s	$\color{#35bf28}+1.04\%$
test_membership_stacked_nested_leaf	64.3910μs	2.0875μs	479.0488 KOps/s	483.7824 KOps/s	$\color{#d91a1a}-0.98\%$
test_membership_nested_last	35.2310μs	3.1296μs	319.5277 KOps/s	322.5093 KOps/s	$\color{#d91a1a}-0.92\%$
test_membership_nested_leaf_last	44.5310μs	3.1521μs	317.2441 KOps/s	328.3025 KOps/s	$\color{#d91a1a}-3.37\%$
test_membership_stacked_nested_last	60.4210μs	3.6421μs	274.5690 KOps/s	241.0404 KOps/s	$\textbf{\color{#35bf28}+13.91\%}$
test_membership_stacked_nested_leaf_last	34.9610μs	3.6383μs	274.8567 KOps/s	238.1124 KOps/s	$\textbf{\color{#35bf28}+15.43\%}$
test_nested_getleaf	34.1910μs	6.1620μs	162.2851 KOps/s	163.6503 KOps/s	$\color{#d91a1a}-0.83\%$
test_nested_get	41.5600μs	5.8230μs	171.7315 KOps/s	172.0856 KOps/s	$\color{#d91a1a}-0.21\%$
test_stacked_getleaf	32.1300μs	6.1384μs	162.9089 KOps/s	163.6275 KOps/s	$\color{#d91a1a}-0.44\%$
test_stacked_get	64.9610μs	5.8244μs	171.6925 KOps/s	172.3025 KOps/s	$\color{#d91a1a}-0.35\%$
test_nested_getitemleaf	38.0310μs	6.4414μs	155.2463 KOps/s	156.9865 KOps/s	$\color{#d91a1a}-1.11\%$
test_nested_getitem	26.9210μs	6.2150μs	160.9007 KOps/s	164.6507 KOps/s	$\color{#d91a1a}-2.28\%$
test_stacked_getitemleaf	29.0500μs	6.4562μs	154.8896 KOps/s	155.5357 KOps/s	$\color{#d91a1a}-0.42\%$
test_stacked_getitem	67.1820μs	6.1159μs	163.5095 KOps/s	164.1392 KOps/s	$\color{#d91a1a}-0.38\%$
test_lock_nested	0.7300ms	0.3727ms	2.6834 KOps/s	2.6181 KOps/s	$\color{#35bf28}+2.50\%$
test_lock_stack_nested	0.3839ms	0.3461ms	2.8897 KOps/s	2.8986 KOps/s	$\color{#d91a1a}-0.31\%$
test_unlock_nested	0.6389ms	0.3183ms	3.1419 KOps/s	3.1908 KOps/s	$\color{#d91a1a}-1.53\%$
test_unlock_stack_nested	0.3136ms	0.2847ms	3.5128 KOps/s	3.5395 KOps/s	$\color{#d91a1a}-0.76\%$
test_flatten_speed	0.1529ms	76.6164μs	13.0520 KOps/s	13.1478 KOps/s	$\color{#d91a1a}-0.73\%$
test_unflatten_speed	0.4321ms	0.3163ms	3.1611 KOps/s	3.0934 KOps/s	$\color{#35bf28}+2.19\%$
test_common_ops	1.6097ms	0.5993ms	1.6685 KOps/s	1.5647 KOps/s	$\textbf{\color{#35bf28}+6.63\%}$
test_creation	0.1061ms	1.7548μs	569.8675 KOps/s	564.8825 KOps/s	$\color{#35bf28}+0.88\%$
test_creation_empty	1.4018ms	8.4065μs	118.9549 KOps/s	100.5220 KOps/s	$\textbf{\color{#35bf28}+18.34\%}$
test_creation_nested_1	34.5010μs	10.0227μs	99.7735 KOps/s	85.5179 KOps/s	$\textbf{\color{#35bf28}+16.67\%}$
test_creation_nested_2	35.4000μs	12.8992μs	77.5240 KOps/s	70.4418 KOps/s	$\textbf{\color{#35bf28}+10.05\%}$
test_clone	0.1185ms	10.2422μs	97.6357 KOps/s	98.7538 KOps/s	$\color{#d91a1a}-1.13\%$
test_getitem[int]	1.1996ms	11.0342μs	90.6274 KOps/s	93.2100 KOps/s	$\color{#d91a1a}-2.77\%$
test_getitem[slice_int]	0.1138ms	21.2509μs	47.0568 KOps/s	47.7115 KOps/s	$\color{#d91a1a}-1.37\%$
test_getitem[range]	0.1308ms	36.9925μs	27.0325 KOps/s	26.7071 KOps/s	$\color{#35bf28}+1.22\%$
test_getitem[tuple]	0.1098ms	18.4658μs	54.1542 KOps/s	55.0269 KOps/s	$\color{#d91a1a}-1.59\%$
test_getitem[list]	0.3401ms	32.6327μs	30.6441 KOps/s	30.3330 KOps/s	$\color{#35bf28}+1.03\%$
test_setitem_dim[int]	39.7710μs	18.9138μs	52.8714 KOps/s	54.3547 KOps/s	$\color{#d91a1a}-2.73\%$
test_setitem_dim[slice_int]	60.4510μs	37.5831μs	26.6077 KOps/s	26.3350 KOps/s	$\color{#35bf28}+1.04\%$
test_setitem_dim[range]	81.2110μs	52.2510μs	19.1384 KOps/s	18.8594 KOps/s	$\color{#35bf28}+1.48\%$
test_setitem_dim[tuple]	53.1210μs	31.0818μs	32.1732 KOps/s	30.6896 KOps/s	$\color{#35bf28}+4.83\%$
test_setitem	0.1020ms	14.2883μs	69.9871 KOps/s	64.4163 KOps/s	$\textbf{\color{#35bf28}+8.65\%}$
test_set	0.1126ms	13.9216μs	71.8309 KOps/s	66.6573 KOps/s	$\textbf{\color{#35bf28}+7.76\%}$
test_set_shared	1.7834ms	0.1509ms	6.6277 KOps/s	6.5484 KOps/s	$\color{#35bf28}+1.21\%$
test_update	0.3057ms	17.1600μs	58.2750 KOps/s	52.9868 KOps/s	$\textbf{\color{#35bf28}+9.98\%}$
test_update_nested	0.1084ms	22.3174μs	44.8081 KOps/s	41.2323 KOps/s	$\textbf{\color{#35bf28}+8.67\%}$
test_update__nested	0.4617ms	24.6946μs	40.4947 KOps/s	40.6581 KOps/s	$\color{#d91a1a}-0.40\%$
test_set_nested	97.2510μs	15.3936μs	64.9621 KOps/s	61.2502 KOps/s	$\textbf{\color{#35bf28}+6.06\%}$
test_set_nested_new	0.1166ms	17.7337μs	56.3900 KOps/s	53.1964 KOps/s	$\textbf{\color{#35bf28}+6.00\%}$
test_select	0.2180ms	29.9161μs	33.4268 KOps/s	32.5586 KOps/s	$\color{#35bf28}+2.67\%$
test_select_nested	75.8010μs	43.4362μs	23.0223 KOps/s	22.7952 KOps/s	$\color{#35bf28}+1.00\%$
test_exclude_nested	0.1377ms	61.9355μs	16.1458 KOps/s	15.6175 KOps/s	$\color{#35bf28}+3.38\%$
test_empty[True]	0.3482ms	0.2935ms	3.4069 KOps/s	3.4094 KOps/s	$\color{#d91a1a}-0.08\%$
test_empty[False]	3.2061μs	0.8175μs	1.2232 MOps/s	1.2009 MOps/s	$\color{#35bf28}+1.86\%$
test_to	86.3410μs	56.7723μs	17.6142 KOps/s	17.9021 KOps/s	$\color{#d91a1a}-1.61\%$
test_to_nonblocking	83.1310μs	48.0174μs	20.8258 KOps/s	20.8741 KOps/s	$\color{#d91a1a}-0.23\%$
test_unbind_speed	1.5898ms	0.2401ms	4.1641 KOps/s	4.2174 KOps/s	$\color{#d91a1a}-1.26\%$
test_unbind_speed_stack0	0.3007ms	0.2384ms	4.1952 KOps/s	4.1174 KOps/s	$\color{#35bf28}+1.89\%$
test_unbind_speed_stack1	95.4108ms	0.6747ms	1.4822 KOps/s	1.4895 KOps/s	$\color{#d91a1a}-0.49\%$
test_split	95.8192ms	1.6130ms	619.9460 Ops/s	628.2366 Ops/s	$\color{#d91a1a}-1.32\%$
test_chunk	97.7415ms	1.6082ms	621.8095 Ops/s	633.5333 Ops/s	$\color{#d91a1a}-1.85\%$
test_consolidate[False-None]	0.1003s	2.9850ms	335.0127 Ops/s	335.7523 Ops/s	$\color{#d91a1a}-0.22\%$
test_consolidate[default-None]	2.6830ms	1.7078ms	585.5547 Ops/s	585.5819 Ops/s	$-0.00\%$
test_consolidate[reduce-overhead-None]	1.8348ms	1.7604ms	568.0394 Ops/s	571.1071 Ops/s	$\color{#d91a1a}-0.54\%$
test_consolidate_njt[False-None]	7.0287ms	6.5763ms	152.0604 Ops/s	150.4412 Ops/s	$\color{#35bf28}+1.08\%$
test_to[False-False-None]	1.7675ms	1.6942ms	590.2374 Ops/s	584.9811 Ops/s	$\color{#35bf28}+0.90\%$
test_to[True-False-None]	1.5924ms	1.3354ms	748.8619 Ops/s	735.2609 Ops/s	$\color{#35bf28}+1.85\%$
test_to[within-False-None]	4.4522ms	4.1478ms	241.0901 Ops/s	238.5639 Ops/s	$\color{#35bf28}+1.06\%$
test_to[True-default-None]	5.6983ms	5.3044ms	188.5217 Ops/s	186.6576 Ops/s	$\color{#35bf28}+1.00\%$
test_to_njt[False-False-None]	7.1147ms	6.9354ms	144.1885 Ops/s	143.1087 Ops/s	$\color{#35bf28}+0.75\%$
test_to_njt[True-False-None]	5.7031ms	5.5007ms	181.7962 Ops/s	181.4035 Ops/s	$\color{#35bf28}+0.22\%$
test_to_njt[within-False-None]	12.4025ms	12.2847ms	81.4022 Ops/s	80.2353 Ops/s	$\color{#35bf28}+1.45\%$
test_creation[device0]	0.4664ms	80.5937μs	12.4079 KOps/s	12.3514 KOps/s	$\color{#35bf28}+0.46\%$
test_creation_from_tensor	0.6318ms	83.5315μs	11.9715 KOps/s	11.8743 KOps/s	$\color{#35bf28}+0.82\%$
test_add_one[memmap_tensor0]	0.4221ms	6.3898μs	156.4995 KOps/s	156.4370 KOps/s	$\color{#35bf28}+0.04\%$
test_contiguous[memmap_tensor0]	1.7776μs	0.4309μs	2.3209 MOps/s	2.3168 MOps/s	$\color{#35bf28}+0.18\%$
test_stack[memmap_tensor0]	20.3800μs	4.6408μs	215.4812 KOps/s	222.7217 KOps/s	$\color{#d91a1a}-3.25\%$
test_memmaptd_index	1.3680ms	0.2604ms	3.8398 KOps/s	3.9264 KOps/s	$\color{#d91a1a}-2.21\%$
test_memmaptd_index_astensor	0.5887ms	0.3243ms	3.0831 KOps/s	3.1749 KOps/s	$\color{#d91a1a}-2.89\%$
test_memmaptd_index_op	1.0458ms	0.5847ms	1.7103 KOps/s	1.6701 KOps/s	$\color{#35bf28}+2.41\%$
test_serialize_model	0.1326s	0.1317s	7.5927 Ops/s	7.6200 Ops/s	$\color{#d91a1a}-0.36\%$
test_serialize_model_pickle	1.3492s	1.2106s	0.8260 Ops/s	0.8259 Ops/s	$\color{#35bf28}+0.02\%$
test_serialize_weights	0.1321s	0.1304s	7.6689 Ops/s	7.6726 Ops/s	$\color{#d91a1a}-0.05\%$
test_serialize_weights_returnearly	0.3428s	53.8079ms	18.5846 Ops/s	14.4342 Ops/s	$\textbf{\color{#35bf28}+28.75\%}$
test_serialize_weights_pickle	1.3458s	1.1860s	0.8432 Ops/s	0.7953 Ops/s	$\textbf{\color{#35bf28}+6.03\%}$
test_reshape_pytree	62.1910μs	22.9149μs	43.6398 KOps/s	44.2175 KOps/s	$\color{#d91a1a}-1.31\%$
test_reshape_td	63.6010μs	26.5714μs	37.6345 KOps/s	36.0599 KOps/s	$\color{#35bf28}+4.37\%$
test_view_pytree	0.1207ms	22.1478μs	45.1513 KOps/s	45.1227 KOps/s	$\color{#35bf28}+0.06\%$
test_view_td	59.6910μs	31.2538μs	31.9961 KOps/s	30.8098 KOps/s	$\color{#35bf28}+3.85\%$
test_unbind_pytree	0.1522ms	28.3553μs	35.2667 KOps/s	35.2425 KOps/s	$\color{#35bf28}+0.07\%$
test_unbind_td	0.7515ms	37.0397μs	26.9981 KOps/s	27.3447 KOps/s	$\color{#d91a1a}-1.27\%$
test_split_pytree	57.0210μs	30.6438μs	32.6330 KOps/s	31.8864 KOps/s	$\color{#35bf28}+2.34\%$
test_split_td	1.0053ms	39.7555μs	25.1537 KOps/s	24.9813 KOps/s	$\color{#35bf28}+0.69\%$
test_add_pytree	61.8510μs	33.6762μs	29.6946 KOps/s	29.9606 KOps/s	$\color{#d91a1a}-0.89\%$
test_add_td	0.1802ms	47.6207μs	20.9993 KOps/s	19.3273 KOps/s	$\textbf{\color{#35bf28}+8.65\%}$
test_compile_add_one_nested[tensordict-compile]	0.1757ms	0.1218ms	8.2123 KOps/s	7.9854 KOps/s	$\color{#35bf28}+2.84\%$
test_compile_add_one_nested[tensordict-eager]	0.2843ms	0.1325ms	7.5452 KOps/s	7.3881 KOps/s	$\color{#35bf28}+2.13\%$
test_compile_add_one_nested[pytree-compile]	0.1472ms	98.6278μs	10.1391 KOps/s	10.3441 KOps/s	$\color{#d91a1a}-1.98\%$
test_compile_add_one_nested[pytree-eager]	1.6285ms	0.1505ms	6.6465 KOps/s	6.8249 KOps/s	$\color{#d91a1a}-2.61\%$
test_compile_copy_nested[tensordict-compile]	96.7610μs	22.3815μs	44.6798 KOps/s	44.6272 KOps/s	$\color{#35bf28}+0.12\%$
test_compile_copy_nested[tensordict-eager]	82.9920μs	29.6450μs	33.7325 KOps/s	33.0964 KOps/s	$\color{#35bf28}+1.92\%$
test_compile_copy_nested[pytree-compile]	0.4085ms	65.1945μs	15.3387 KOps/s	15.0402 KOps/s	$\color{#35bf28}+1.98\%$
test_compile_copy_nested[pytree-eager]	83.6310μs	49.7392μs	20.1049 KOps/s	19.8329 KOps/s	$\color{#35bf28}+1.37\%$
test_compile_add_one_flat[tensordict-compile]	0.1824ms	0.1420ms	7.0419 KOps/s	6.9144 KOps/s	$\color{#35bf28}+1.84\%$
test_compile_add_one_flat[tensordict-eager]	0.3106ms	0.2165ms	4.6191 KOps/s	4.6053 KOps/s	$\color{#35bf28}+0.30\%$
test_compile_add_one_flat[tensorclass-compile]	0.1854ms	99.1465μs	10.0861 KOps/s	10.1742 KOps/s	$\color{#d91a1a}-0.87\%$
test_compile_add_one_flat[tensorclass-eager]	0.1150ms	56.3841μs	17.7355 KOps/s	17.8347 KOps/s	$\color{#d91a1a}-0.56\%$
test_compile_add_one_flat[pytree-compile]	0.1738ms	0.1354ms	7.3844 KOps/s	7.3848 KOps/s	$-0.01\%$
test_compile_add_one_flat[pytree-eager]	0.5729ms	0.4789ms	2.0880 KOps/s	2.1430 KOps/s	$\color{#d91a1a}-2.57\%$
test_compile_add_self_flat[tensordict-eager]	0.3816ms	0.2608ms	3.8348 KOps/s	3.8593 KOps/s	$\color{#d91a1a}-0.63\%$
test_compile_add_self_flat[tensordict-compile]	0.2022ms	0.1497ms	6.6784 KOps/s	7.0389 KOps/s	$\textbf{\color{#d91a1a}-5.12\%}$
test_compile_add_self_flat[tensorclass-eager]	0.1696ms	70.0694μs	14.2716 KOps/s	14.5317 KOps/s	$\color{#d91a1a}-1.79\%$
test_compile_add_self_flat[tensorclass-compile]	0.1510ms	0.1064ms	9.4023 KOps/s	10.1703 KOps/s	$\textbf{\color{#d91a1a}-7.55\%}$
test_compile_add_self_flat[pytree-eager]	0.5538ms	0.4089ms	2.4455 KOps/s	2.4698 KOps/s	$\color{#d91a1a}-0.98\%$
test_compile_add_self_flat[pytree-compile]	0.1756ms	0.1369ms	7.3058 KOps/s	7.4310 KOps/s	$\color{#d91a1a}-1.69\%$
test_compile_copy_flat[tensordict-compile]	79.6710μs	19.2872μs	51.8479 KOps/s	55.0637 KOps/s	$\textbf{\color{#d91a1a}-5.84\%}$
test_compile_copy_flat[tensordict-eager]	96.5020μs	30.8022μs	32.4652 KOps/s	31.6246 KOps/s	$\color{#35bf28}+2.66\%$
test_compile_copy_flat[pytree-compile]	0.1053ms	70.4879μs	14.1868 KOps/s	13.9578 KOps/s	$\color{#35bf28}+1.64\%$
test_compile_copy_flat[pytree-eager]	0.1373ms	52.1730μs	19.1670 KOps/s	19.2130 KOps/s	$\color{#d91a1a}-0.24\%$
test_compile_assign_and_add[tensordict-compile]	1.6349ms	0.3935ms	2.5414 KOps/s	2.1919 KOps/s	$\textbf{\color{#35bf28}+15.94\%}$
test_compile_assign_and_add[tensordict-eager]	2.6887ms	2.6134ms	382.6433 Ops/s	383.4553 Ops/s	$\color{#d91a1a}-0.21\%$
test_compile_assign_and_add[pytree-compile]	1.5948ms	0.4353ms	2.2975 KOps/s	2.1218 KOps/s	$\textbf{\color{#35bf28}+8.28\%}$
test_compile_assign_and_add[pytree-eager]	2.7371ms	2.6327ms	379.8401 Ops/s	381.5021 Ops/s	$\color{#d91a1a}-0.44\%$
test_compile_indexing[tensor-tensordict-compile]	0.1635ms	0.1117ms	8.9493 KOps/s	8.4632 KOps/s	$\textbf{\color{#35bf28}+5.74\%}$
test_compile_indexing[tensor-tensordict-eager]	0.5511ms	77.9309μs	12.8319 KOps/s	12.3661 KOps/s	$\color{#35bf28}+3.77\%$
test_compile_indexing[tensor-tensorclass-compile]	0.1705ms	0.1046ms	9.5589 KOps/s	9.5692 KOps/s	$\color{#d91a1a}-0.11\%$
test_compile_indexing[tensor-tensorclass-eager]	0.1106ms	67.2966μs	14.8596 KOps/s	14.4228 KOps/s	$\color{#35bf28}+3.03\%$
test_compile_indexing[tensor-pytree-compile]	0.1521ms	0.1056ms	9.4674 KOps/s	9.5142 KOps/s	$\color{#d91a1a}-0.49\%$
test_compile_indexing[tensor-pytree-eager]	0.1162ms	66.9258μs	14.9419 KOps/s	14.8972 KOps/s	$\color{#35bf28}+0.30\%$
test_compile_indexing[slice-tensordict-compile]	0.2335ms	0.1031ms	9.7005 KOps/s	9.9924 KOps/s	$\color{#d91a1a}-2.92\%$
test_compile_indexing[slice-tensordict-eager]	0.1677ms	17.4747μs	57.2255 KOps/s	57.5534 KOps/s	$\color{#d91a1a}-0.57\%$
test_compile_indexing[slice-tensorclass-compile]	0.1355ms	96.0504μs	10.4112 KOps/s	10.2783 KOps/s	$\color{#35bf28}+1.29\%$
test_compile_indexing[slice-tensorclass-eager]	57.0610μs	16.0511μs	62.3011 KOps/s	63.4290 KOps/s	$\color{#d91a1a}-1.78\%$
test_compile_indexing[slice-pytree-compile]	0.1443ms	96.5170μs	10.3609 KOps/s	10.0928 KOps/s	$\color{#35bf28}+2.66\%$
test_compile_indexing[slice-pytree-eager]	0.1150ms	15.8779μs	62.9808 KOps/s	63.5312 KOps/s	$\color{#d91a1a}-0.87\%$
test_compile_indexing[int-tensordict-compile]	0.1440ms	0.1001ms	9.9858 KOps/s	9.4654 KOps/s	$\textbf{\color{#35bf28}+5.50\%}$
test_compile_indexing[int-tensordict-eager]	0.5718ms	17.2797μs	57.8714 KOps/s	59.1036 KOps/s	$\color{#d91a1a}-2.08\%$
test_compile_indexing[int-tensorclass-compile]	0.1798ms	96.8586μs	10.3243 KOps/s	10.2614 KOps/s	$\color{#35bf28}+0.61\%$
test_compile_indexing[int-tensorclass-eager]	43.7710μs	15.8422μs	63.1225 KOps/s	64.0035 KOps/s	$\color{#d91a1a}-1.38\%$
test_compile_indexing[int-pytree-compile]	0.1395ms	96.6823μs	10.3432 KOps/s	10.2593 KOps/s	$\color{#35bf28}+0.82\%$
test_compile_indexing[int-pytree-eager]	47.9610μs	15.7370μs	63.5445 KOps/s	63.1236 KOps/s	$\color{#35bf28}+0.67\%$
test_mod_add[eager]	0.1300ms	38.4616μs	25.9999 KOps/s	25.6554 KOps/s	$\color{#35bf28}+1.34\%$
test_mod_add[compile]	0.1268ms	82.4111μs	12.1343 KOps/s	11.7979 KOps/s	$\color{#35bf28}+2.85\%$
test_mod_add[compile-overhead]	0.3182ms	0.1674ms	5.9742 KOps/s	5.6789 KOps/s	$\textbf{\color{#35bf28}+5.20\%}$
test_mod_wrap[eager]	0.3306ms	0.2521ms	3.9660 KOps/s	3.8181 KOps/s	$\color{#35bf28}+3.87\%$
test_mod_wrap[compile]	0.3374ms	0.2843ms	3.5177 KOps/s	3.4913 KOps/s	$\color{#35bf28}+0.75\%$
test_mod_wrap[compile-overhead]	7.0129ms	3.7234ms	268.5737 Ops/s	271.5757 Ops/s	$\color{#d91a1a}-1.11\%$
test_mod_wrap_and_backward[eager]	1.4668ms	1.3280ms	753.0186 Ops/s	709.5403 Ops/s	$\textbf{\color{#35bf28}+6.13\%}$
test_mod_wrap_and_backward[compile]	1.3594ms	1.2742ms	784.7818 Ops/s	730.8724 Ops/s	$\textbf{\color{#35bf28}+7.38\%}$
test_mod_wrap_and_backward[compile-overhead]	1.3589ms	0.9206ms	1.0862 KOps/s	907.8000 Ops/s	$\textbf{\color{#35bf28}+19.65\%}$
test_seq_add[eager]	0.2486ms	0.1165ms	8.5874 KOps/s	8.4640 KOps/s	$\color{#35bf28}+1.46\%$
test_seq_add[compile]	0.1356ms	88.8900μs	11.2499 KOps/s	11.1793 KOps/s	$\color{#35bf28}+0.63\%$
test_seq_add[compile-overhead]	0.1760ms	0.1298ms	7.7049 KOps/s	7.7251 KOps/s	$\color{#d91a1a}-0.26\%$
test_seq_wrap[eager]	0.4769ms	0.4117ms	2.4291 KOps/s	2.3596 KOps/s	$\color{#35bf28}+2.95\%$
test_seq_wrap[compile]	0.3514ms	0.3019ms	3.3120 KOps/s	3.3144 KOps/s	$\color{#d91a1a}-0.07\%$
test_seq_wrap[compile-overhead]	0.3075ms	0.2245ms	4.4539 KOps/s	4.4037 KOps/s	$\color{#35bf28}+1.14\%$
test_func_call_runtime[False-eager]	0.7886ms	0.7146ms	1.3995 KOps/s	1.3596 KOps/s	$\color{#35bf28}+2.93\%$
test_func_call_runtime[False-compile]	0.8074ms	0.7540ms	1.3262 KOps/s	1.3451 KOps/s	$\color{#d91a1a}-1.40\%$
test_func_call_runtime[False-compile-overhead]	0.4240ms	0.3661ms	2.7314 KOps/s	2.7109 KOps/s	$\color{#35bf28}+0.75\%$
test_func_call_runtime[True-eager]	0.9428ms	0.8697ms	1.1498 KOps/s	1.1456 KOps/s	$\color{#35bf28}+0.37\%$
test_func_call_runtime[True-compile]	0.8289ms	0.7747ms	1.2908 KOps/s	1.3083 KOps/s	$\color{#d91a1a}-1.34\%$
test_func_call_runtime[True-compile-overhead]	0.4478ms	0.3870ms	2.5843 KOps/s	2.5818 KOps/s	$\color{#35bf28}+0.10\%$
test_func_call_cm_runtime[False-eager]	0.8085ms	0.7127ms	1.4030 KOps/s	1.4136 KOps/s	$\color{#d91a1a}-0.75\%$
test_func_call_cm_runtime[False-compile]	0.8052ms	0.7570ms	1.3210 KOps/s	1.3252 KOps/s	$\color{#d91a1a}-0.32\%$
test_func_call_cm_runtime[False-compile-overhead]	0.4376ms	0.3676ms	2.7206 KOps/s	2.7130 KOps/s	$\color{#35bf28}+0.28\%$
test_func_call_cm_runtime[True-eager]	1.0664ms	0.9818ms	1.0185 KOps/s	999.9982 Ops/s	$\color{#35bf28}+1.85\%$
test_func_call_cm_runtime[True-compile]	0.9481ms	0.8013ms	1.2479 KOps/s	1.2568 KOps/s	$\color{#d91a1a}-0.71\%$
test_func_call_cm_runtime[True-compile-overhead]	0.5597ms	0.4115ms	2.4303 KOps/s	2.4129 KOps/s	$\color{#35bf28}+0.72\%$
test_vmap_func_call_cm_runtime[eager]	2.4839ms	2.0239ms	494.0973 Ops/s	492.1497 Ops/s	$\color{#35bf28}+0.40\%$
test_vmap_func_call_cm_runtime[compile]	0.9538ms	0.8186ms	1.2216 KOps/s	1.2297 KOps/s	$\color{#d91a1a}-0.66\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.5218ms	0.4140ms	2.4153 KOps/s	2.4009 KOps/s	$\color{#35bf28}+0.60\%$
test_distributed	0.6819ms	0.1624ms	6.1583 KOps/s	8.5343 KOps/s	$\textbf{\color{#d91a1a}-27.84\%}$
test_tdmodule	0.1675ms	20.7283μs	48.2433 KOps/s	47.4886 KOps/s	$\color{#35bf28}+1.59\%$
test_tdmodule_dispatch	70.0910μs	36.3810μs	27.4869 KOps/s	26.4004 KOps/s	$\color{#35bf28}+4.12\%$
test_tdseq	32.6110μs	20.8580μs	47.9433 KOps/s	45.4092 KOps/s	$\textbf{\color{#35bf28}+5.58\%}$
test_tdseq_dispatch	59.7210μs	38.7125μs	25.8315 KOps/s	24.2046 KOps/s	$\textbf{\color{#35bf28}+6.72\%}$
test_instantiation_functorch	1.6062ms	1.5420ms	648.5210 Ops/s	639.9389 Ops/s	$\color{#35bf28}+1.34\%$
test_exec_functorch	0.2084ms	0.1408ms	7.1024 KOps/s	7.0790 KOps/s	$\color{#35bf28}+0.33\%$
test_exec_functional_call	0.2012ms	0.1314ms	7.6075 KOps/s	7.4408 KOps/s	$\color{#35bf28}+2.24\%$
test_exec_td_decorator	0.3685ms	0.1787ms	5.5945 KOps/s	5.4593 KOps/s	$\color{#35bf28}+2.48\%$
test_vmap_mlp_speed_decorator[True-True]	0.8233ms	0.6674ms	1.4983 KOps/s	1.4942 KOps/s	$\color{#35bf28}+0.27\%$
test_vmap_mlp_speed_decorator[True-False]	0.8346ms	0.6701ms	1.4922 KOps/s	1.5016 KOps/s	$\color{#d91a1a}-0.62\%$
test_vmap_mlp_speed_decorator[False-True]	0.7217ms	0.5821ms	1.7180 KOps/s	1.7452 KOps/s	$\color{#d91a1a}-1.55\%$
test_vmap_mlp_speed_decorator[False-False]	0.7486ms	0.5881ms	1.7005 KOps/s	1.7470 KOps/s	$\color{#d91a1a}-2.66\%$
test_vmap_transformer_speed_decorator[True-True]	19.2999ms	18.6852ms	53.5182 Ops/s	53.9437 Ops/s	$\color{#d91a1a}-0.79\%$
test_vmap_transformer_speed_decorator[True-False]	19.3566ms	18.7070ms	53.4558 Ops/s	53.9514 Ops/s	$\color{#d91a1a}-0.92\%$
test_vmap_transformer_speed_decorator[False-True]	18.8087ms	18.6279ms	53.6830 Ops/s	54.5233 Ops/s	$\color{#d91a1a}-1.54\%$
test_vmap_transformer_speed_decorator[False-False]	18.7740ms	18.6011ms	53.7602 Ops/s	54.3385 Ops/s	$\color{#d91a1a}-1.06\%$
test_to_module_speed[True]	1.0653ms	0.9736ms	1.0272 KOps/s	1.0364 KOps/s	$\color{#d91a1a}-0.89\%$
test_to_module_speed[False]	1.3530ms	0.9653ms	1.0359 KOps/s	1.0441 KOps/s	$\color{#d91a1a}-0.78\%$
test_tc_init	76.8120μs	35.2068μs	28.4036 KOps/s	26.6669 KOps/s	$\textbf{\color{#35bf28}+6.51\%}$
test_tc_init_nested	0.2397ms	69.5196μs	14.3844 KOps/s	13.2001 KOps/s	$\textbf{\color{#35bf28}+8.97\%}$
test_tc_first_layer_tensor	13.4973μs	0.7143μs	1.4000 MOps/s	1.2071 MOps/s	$\textbf{\color{#35bf28}+15.98\%}$
test_tc_first_layer_nontensor	67.1910μs	2.3070μs	433.4695 KOps/s	434.9652 KOps/s	$\color{#d91a1a}-0.34\%$
test_tc_second_layer_tensor	47.7973μs	1.4370μs	695.9178 KOps/s	694.3641 KOps/s	$\color{#35bf28}+0.22\%$
test_tc_second_layer_nontensor	36.2310μs	3.0287μs	330.1780 KOps/s	331.9329 KOps/s	$\color{#d91a1a}-0.53\%$
test_unbind	0.2300s	10.2275ms	97.7758 Ops/s	140.4178 Ops/s	$\textbf{\color{#d91a1a}-30.37\%}$
test_full_like	10.9462ms	10.0337ms	99.6638 Ops/s	103.9180 Ops/s	$\color{#d91a1a}-4.09\%$
test_zeros_like	9.3303ms	7.3134ms	136.7358 Ops/s	138.9037 Ops/s	$\color{#d91a1a}-1.56\%$
test_ones_like	5.0610ms	4.3868ms	227.9588 Ops/s	227.4273 Ops/s	$\color{#35bf28}+0.23\%$
test_clone	12.1982ms	9.4718ms	105.5766 Ops/s	147.4653 Ops/s	$\textbf{\color{#d91a1a}-28.41\%}$
test_squeeze	54.5110μs	9.8328μs	101.7006 KOps/s	103.3615 KOps/s	$\color{#d91a1a}-1.61\%$
test_unsqueeze	0.1935ms	73.2570μs	13.6506 KOps/s	13.3855 KOps/s	$\color{#35bf28}+1.98\%$
test_split	0.3924ms	0.1598ms	6.2561 KOps/s	5.8446 KOps/s	$\textbf{\color{#35bf28}+7.04\%}$
test_permute	0.3209ms	0.1802ms	5.5504 KOps/s	5.5609 KOps/s	$\color{#d91a1a}-0.19\%$
test_stack	52.0102ms	50.5134ms	19.7967 Ops/s	19.5183 Ops/s	$\color{#35bf28}+1.43\%$
test_cat	51.8713ms	51.3347ms	19.4800 Ops/s	19.5330 Ops/s	$\color{#d91a1a}-0.27\%$