[Feature] TensorDict.softmax

[vmoens]

Stack from ghstack (oldest at bottom):

• [Feature] broadcast pointwise ops for tensor/tensordict mixed inputs #1166
• [Feature] TensorDict.clamp #1165
• [BugFix] Fix unitary ops for tensorclass #1164
• -> [Feature] TensorDict.softmax #1163
• [Feature] TensorDict.logsumexp #1162

[github-actions]

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

Total Benchmarks: 217. Improved: $\large\color{#35bf28}8$. Worsened: $\large\color{#d91a1a}43$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	41.5370μs	20.6825μs	48.3500 KOps/s	50.2283 KOps/s	$\color{#d91a1a}-3.74\%$
test_plain_set_stack_nested	46.6270μs	21.0241μs	47.5645 KOps/s	50.3000 KOps/s	$\textbf{\color{#d91a1a}-5.44\%}$
test_plain_set_nested_inplace	75.5000μs	22.8587μs	43.7470 KOps/s	45.6296 KOps/s	$\color{#d91a1a}-4.13\%$
test_plain_set_stack_nested_inplace	70.0400μs	22.7993μs	43.8609 KOps/s	45.4315 KOps/s	$\color{#d91a1a}-3.46\%$
test_items	34.6240μs	4.2400μs	235.8478 KOps/s	234.7238 KOps/s	$\color{#35bf28}+0.48\%$
test_items_nested	0.6321ms	0.4123ms	2.4253 KOps/s	2.4725 KOps/s	$\color{#d91a1a}-1.91\%$
test_items_nested_locked	0.7342ms	0.4094ms	2.4426 KOps/s	2.4712 KOps/s	$\color{#d91a1a}-1.16\%$
test_items_nested_leaf	0.1469ms	77.0068μs	12.9859 KOps/s	13.0968 KOps/s	$\color{#d91a1a}-0.85\%$
test_items_stack_nested	0.6686ms	0.4118ms	2.4286 KOps/s	2.4431 KOps/s	$\color{#d91a1a}-0.59\%$
test_items_stack_nested_leaf	0.1358ms	80.6572μs	12.3982 KOps/s	12.7020 KOps/s	$\color{#d91a1a}-2.39\%$
test_items_stack_nested_locked	0.8507ms	0.4113ms	2.4311 KOps/s	2.4437 KOps/s	$\color{#d91a1a}-0.51\%$
test_keys	31.1280μs	3.5769μs	279.5703 KOps/s	278.5926 KOps/s	$\color{#35bf28}+0.35\%$
test_keys_nested	0.3112ms	0.1683ms	5.9419 KOps/s	5.9631 KOps/s	$\color{#d91a1a}-0.36\%$
test_keys_nested_locked	0.8168ms	0.1740ms	5.7463 KOps/s	5.7824 KOps/s	$\color{#d91a1a}-0.62\%$
test_keys_nested_leaf	1.8908ms	0.1465ms	6.8239 KOps/s	6.8921 KOps/s	$\color{#d91a1a}-0.99\%$
test_keys_stack_nested	0.2594ms	0.1654ms	6.0472 KOps/s	6.0089 KOps/s	$\color{#35bf28}+0.64\%$
test_keys_stack_nested_leaf	0.2337ms	0.1437ms	6.9607 KOps/s	6.9153 KOps/s	$\color{#35bf28}+0.66\%$
test_keys_stack_nested_locked	0.2863ms	0.1719ms	5.8182 KOps/s	5.8255 KOps/s	$\color{#d91a1a}-0.13\%$
test_values	10.2336μs	1.0755μs	929.7863 KOps/s	965.1051 KOps/s	$\color{#d91a1a}-3.66\%$
test_values_nested	99.6960μs	63.3521μs	15.7848 KOps/s	16.0818 KOps/s	$\color{#d91a1a}-1.85\%$
test_values_nested_locked	0.1190ms	63.1521μs	15.8348 KOps/s	16.0108 KOps/s	$\color{#d91a1a}-1.10\%$
test_values_nested_leaf	0.1429ms	72.0608μs	13.8772 KOps/s	13.8986 KOps/s	$\color{#d91a1a}-0.15\%$
test_values_stack_nested	0.1576ms	63.1629μs	15.8321 KOps/s	14.8917 KOps/s	$\textbf{\color{#35bf28}+6.31\%}$
test_values_stack_nested_leaf	0.1381ms	72.0790μs	13.8737 KOps/s	13.6705 KOps/s	$\color{#35bf28}+1.49\%$
test_values_stack_nested_locked	0.1099ms	63.5926μs	15.7251 KOps/s	15.6537 KOps/s	$\color{#35bf28}+0.46\%$
test_membership	13.9060μs	0.8950μs	1.1173 MOps/s	1.1129 MOps/s	$\color{#35bf28}+0.40\%$
test_membership_nested	55.7250μs	2.9220μs	342.2286 KOps/s	344.5993 KOps/s	$\color{#d91a1a}-0.69\%$
test_membership_nested_leaf	44.5530μs	2.9653μs	337.2365 KOps/s	347.9678 KOps/s	$\color{#d91a1a}-3.08\%$
test_membership_stacked_nested	45.7550μs	2.9083μs	343.8493 KOps/s	345.5924 KOps/s	$\color{#d91a1a}-0.50\%$
test_membership_stacked_nested_leaf	23.7140μs	2.9143μs	343.1352 KOps/s	345.7949 KOps/s	$\color{#d91a1a}-0.77\%$
test_membership_nested_last	28.7130μs	4.3573μs	229.5011 KOps/s	232.4218 KOps/s	$\color{#d91a1a}-1.26\%$
test_membership_nested_leaf_last	35.3560μs	4.3806μs	228.2789 KOps/s	227.4231 KOps/s	$\color{#35bf28}+0.38\%$
test_membership_stacked_nested_last	33.9930μs	5.1236μs	195.1742 KOps/s	231.0654 KOps/s	$\textbf{\color{#d91a1a}-15.53\%}$
test_membership_stacked_nested_leaf_last	26.6300μs	5.1558μs	193.9568 KOps/s	229.8296 KOps/s	$\textbf{\color{#d91a1a}-15.61\%}$
test_nested_getleaf	39.9040μs	10.6623μs	93.7883 KOps/s	92.9834 KOps/s	$\color{#35bf28}+0.87\%$
test_nested_get	31.3180μs	10.1690μs	98.3380 KOps/s	97.1560 KOps/s	$\color{#35bf28}+1.22\%$
test_stacked_getleaf	45.8150μs	10.5876μs	94.4501 KOps/s	93.2720 KOps/s	$\color{#35bf28}+1.26\%$
test_stacked_get	40.9260μs	10.2551μs	97.5128 KOps/s	99.2313 KOps/s	$\color{#d91a1a}-1.73\%$
test_nested_getitemleaf	35.9670μs	11.5475μs	86.5990 KOps/s	89.4442 KOps/s	$\color{#d91a1a}-3.18\%$
test_nested_getitem	66.6640μs	10.0908μs	99.0997 KOps/s	95.4990 KOps/s	$\color{#35bf28}+3.77\%$
test_stacked_getitemleaf	35.8670μs	11.0900μs	90.1716 KOps/s	90.4617 KOps/s	$\color{#d91a1a}-0.32\%$
test_stacked_getitem	0.1015ms	10.3202μs	96.8973 KOps/s	96.2920 KOps/s	$\color{#35bf28}+0.63\%$
test_lock_nested	1.7318ms	0.4606ms	2.1709 KOps/s	2.2143 KOps/s	$\color{#d91a1a}-1.96\%$
test_lock_stack_nested	0.6989ms	0.4320ms	2.3146 KOps/s	2.3316 KOps/s	$\color{#d91a1a}-0.73\%$
test_unlock_nested	0.7824ms	0.3763ms	2.6577 KOps/s	2.6725 KOps/s	$\color{#d91a1a}-0.55\%$
test_unlock_stack_nested	0.5296ms	0.3447ms	2.9013 KOps/s	2.9212 KOps/s	$\color{#d91a1a}-0.68\%$
test_flatten_speed	0.2016ms	0.1025ms	9.7582 KOps/s	10.2545 KOps/s	$\color{#d91a1a}-4.84\%$
test_unflatten_speed	0.9683ms	0.5414ms	1.8472 KOps/s	1.8784 KOps/s	$\color{#d91a1a}-1.66\%$
test_common_ops	4.3316ms	0.8173ms	1.2235 KOps/s	1.3503 KOps/s	$\textbf{\color{#d91a1a}-9.39\%}$
test_creation	59.3110μs	2.5145μs	397.6939 KOps/s	398.8308 KOps/s	$\color{#d91a1a}-0.29\%$
test_creation_empty	34.6850μs	12.2506μs	81.6286 KOps/s	100.8409 KOps/s	$\textbf{\color{#d91a1a}-19.05\%}$
test_creation_nested_1	60.8940μs	15.0769μs	66.3268 KOps/s	78.6192 KOps/s	$\textbf{\color{#d91a1a}-15.64\%}$
test_creation_nested_2	50.2240μs	19.7710μs	50.5790 KOps/s	57.7699 KOps/s	$\textbf{\color{#d91a1a}-12.45\%}$
test_clone	0.2061ms	13.7638μs	72.6543 KOps/s	74.6216 KOps/s	$\color{#d91a1a}-2.64\%$
test_getitem[int]	0.7716ms	13.1608μs	75.9832 KOps/s	80.0537 KOps/s	$\textbf{\color{#d91a1a}-5.08\%}$
test_getitem[slice_int]	0.1398ms	25.2190μs	39.6526 KOps/s	42.0850 KOps/s	$\textbf{\color{#d91a1a}-5.78\%}$
test_getitem[range]	0.1743ms	50.9847μs	19.6137 KOps/s	21.0974 KOps/s	$\textbf{\color{#d91a1a}-7.03\%}$
test_getitem[tuple]	0.1398ms	20.8360μs	47.9938 KOps/s	50.6781 KOps/s	$\textbf{\color{#d91a1a}-5.30\%}$
test_getitem[list]	0.1757ms	46.4644μs	21.5219 KOps/s	23.3545 KOps/s	$\textbf{\color{#d91a1a}-7.85\%}$
test_setitem_dim[int]	56.0840μs	27.2023μs	36.7616 KOps/s	41.3567 KOps/s	$\textbf{\color{#d91a1a}-11.11\%}$
test_setitem_dim[slice_int]	96.2100μs	55.4690μs	18.0281 KOps/s	19.3574 KOps/s	$\textbf{\color{#d91a1a}-6.87\%}$
test_setitem_dim[range]	0.1144ms	75.5158μs	13.2423 KOps/s	13.7425 KOps/s	$\color{#d91a1a}-3.64\%$
test_setitem_dim[tuple]	76.5330μs	43.6422μs	22.9136 KOps/s	24.5134 KOps/s	$\textbf{\color{#d91a1a}-6.53\%}$
test_setitem	82.5630μs	21.5303μs	46.4462 KOps/s	52.0931 KOps/s	$\textbf{\color{#d91a1a}-10.84\%}$
test_set	0.1518ms	21.2768μs	46.9995 KOps/s	53.5514 KOps/s	$\textbf{\color{#d91a1a}-12.23\%}$
test_set_shared	1.1785ms	0.1707ms	5.8587 KOps/s	6.0046 KOps/s	$\color{#d91a1a}-2.43\%$
test_update	0.3041ms	24.1677μs	41.3776 KOps/s	48.1336 KOps/s	$\textbf{\color{#d91a1a}-14.04\%}$
test_update_nested	0.3735ms	34.5055μs	28.9809 KOps/s	32.2272 KOps/s	$\textbf{\color{#d91a1a}-10.07\%}$
test_update__nested	0.5907ms	34.6472μs	28.8623 KOps/s	30.0819 KOps/s	$\color{#d91a1a}-4.05\%$
test_set_nested	0.2893ms	23.4473μs	42.6488 KOps/s	48.7508 KOps/s	$\textbf{\color{#d91a1a}-12.52\%}$
test_set_nested_new	97.2110μs	28.5926μs	34.9741 KOps/s	39.2617 KOps/s	$\textbf{\color{#d91a1a}-10.92\%}$
test_select	0.1053ms	45.1337μs	22.1564 KOps/s	24.3946 KOps/s	$\textbf{\color{#d91a1a}-9.17\%}$
test_select_nested	0.1229ms	62.9693μs	15.8808 KOps/s	16.1964 KOps/s	$\color{#d91a1a}-1.95\%$
test_exclude_nested	0.1558ms	83.3851μs	11.9925 KOps/s	12.4878 KOps/s	$\color{#d91a1a}-3.97\%$
test_empty[True]	0.7173ms	0.4244ms	2.3563 KOps/s	2.4443 KOps/s	$\color{#d91a1a}-3.60\%$
test_empty[False]	6.7400μs	1.3977μs	715.4814 KOps/s	724.0916 KOps/s	$\color{#d91a1a}-1.19\%$
test_unbind_speed	0.5209ms	0.2712ms	3.6875 KOps/s	3.7251 KOps/s	$\color{#d91a1a}-1.01\%$
test_unbind_speed_stack0	0.4450ms	0.2643ms	3.7834 KOps/s	3.7704 KOps/s	$\color{#35bf28}+0.35\%$
test_unbind_speed_stack1	99.9301ms	0.7836ms	1.2762 KOps/s	1.3639 KOps/s	$\textbf{\color{#d91a1a}-6.43\%}$
test_split	1.7887ms	1.6159ms	618.8355 Ops/s	572.1426 Ops/s	$\textbf{\color{#35bf28}+8.16\%}$
test_chunk	0.1065s	1.9526ms	512.1475 Ops/s	572.1762 Ops/s	$\textbf{\color{#d91a1a}-10.49\%}$
test_consolidate_njt[False-None]	8.6133ms	8.2197ms	121.6591 Ops/s	124.1283 Ops/s	$\color{#d91a1a}-1.99\%$
test_creation[device0]	4.5817ms	94.0892μs	10.6282 KOps/s	10.8214 KOps/s	$\color{#d91a1a}-1.79\%$
test_creation_from_tensor	0.2392ms	94.6658μs	10.5635 KOps/s	10.4551 KOps/s	$\color{#35bf28}+1.04\%$
test_add_one[memmap_tensor0]	0.2157ms	4.9457μs	202.1948 KOps/s	215.6424 KOps/s	$\textbf{\color{#d91a1a}-6.24\%}$
test_contiguous[memmap_tensor0]	16.2710μs	0.5116μs	1.9548 MOps/s	1.9567 MOps/s	$\color{#d91a1a}-0.10\%$
test_stack[memmap_tensor0]	51.4860μs	3.4080μs	293.4293 KOps/s	303.3288 KOps/s	$\color{#d91a1a}-3.26\%$
test_memmaptd_index	0.9423ms	0.2414ms	4.1428 KOps/s	4.1801 KOps/s	$\color{#d91a1a}-0.89\%$
test_memmaptd_index_astensor	0.7457ms	0.3327ms	3.0058 KOps/s	3.0785 KOps/s	$\color{#d91a1a}-2.36\%$
test_memmaptd_index_op	0.9747ms	0.6105ms	1.6379 KOps/s	1.8106 KOps/s	$\textbf{\color{#d91a1a}-9.54\%}$
test_serialize_model	0.1396s	0.1167s	8.5699 Ops/s	8.6516 Ops/s	$\color{#d91a1a}-0.95\%$
test_serialize_model_pickle	0.4443s	0.3881s	2.5767 Ops/s	2.5225 Ops/s	$\color{#35bf28}+2.15\%$
test_serialize_weights	0.1191s	0.1118s	8.9415 Ops/s	8.9459 Ops/s	$\color{#d91a1a}-0.05\%$
test_serialize_weights_returnearly	0.2569s	0.1713s	5.8385 Ops/s	6.3687 Ops/s	$\textbf{\color{#d91a1a}-8.33\%}$
test_serialize_weights_pickle	0.5637s	0.4202s	2.3798 Ops/s	2.5910 Ops/s	$\textbf{\color{#d91a1a}-8.15\%}$
test_serialize_weights_filesystem	0.1497s	0.1459s	6.8559 Ops/s	7.0825 Ops/s	$\color{#d91a1a}-3.20\%$
test_serialize_model_filesystem	0.1579s	0.1470s	6.8014 Ops/s	6.0522 Ops/s	$\textbf{\color{#35bf28}+12.38\%}$
test_reshape_pytree	61.2740μs	26.1822μs	38.1939 KOps/s	36.1869 KOps/s	$\textbf{\color{#35bf28}+5.55\%}$
test_reshape_td	71.1330μs	32.4472μs	30.8193 KOps/s	29.8905 KOps/s	$\color{#35bf28}+3.11\%$
test_view_pytree	80.2300μs	26.3375μs	37.9686 KOps/s	36.8598 KOps/s	$\color{#35bf28}+3.01\%$
test_view_td	89.6470μs	39.3617μs	25.4054 KOps/s	26.3126 KOps/s	$\color{#d91a1a}-3.45\%$
test_unbind_pytree	90.5980μs	29.9187μs	33.4239 KOps/s	33.7910 KOps/s	$\color{#d91a1a}-1.09\%$
test_unbind_td	0.3213ms	40.2202μs	24.8631 KOps/s	25.4624 KOps/s	$\color{#d91a1a}-2.35\%$
test_split_pytree	72.7350μs	29.8642μs	33.4849 KOps/s	33.5438 KOps/s	$\color{#d91a1a}-0.18\%$
test_split_td	0.5292ms	46.0176μs	21.7308 KOps/s	22.5075 KOps/s	$\color{#d91a1a}-3.45\%$
test_add_pytree	98.0220μs	36.5232μs	27.3799 KOps/s	28.2580 KOps/s	$\color{#d91a1a}-3.11\%$
test_add_td	0.1535ms	60.6723μs	16.4820 KOps/s	19.4438 KOps/s	$\textbf{\color{#d91a1a}-15.23\%}$
test_compile_add_one_nested[tensordict-compile]	0.1326ms	63.2784μs	15.8032 KOps/s	15.7512 KOps/s	$\color{#35bf28}+0.33\%$
test_compile_add_one_nested[tensordict-eager]	0.4057ms	0.1713ms	5.8363 KOps/s	5.8779 KOps/s	$\color{#d91a1a}-0.71\%$
test_compile_add_one_nested[pytree-compile]	0.1220ms	46.8053μs	21.3651 KOps/s	21.9480 KOps/s	$\color{#d91a1a}-2.66\%$
test_compile_add_one_nested[pytree-eager]	0.2588ms	0.1234ms	8.1044 KOps/s	8.3408 KOps/s	$\color{#d91a1a}-2.83\%$
test_compile_copy_nested[tensordict-compile]	60.5520μs	26.6251μs	37.5585 KOps/s	39.9828 KOps/s	$\textbf{\color{#d91a1a}-6.06\%}$
test_compile_copy_nested[tensordict-eager]	0.1404ms	59.6573μs	16.7624 KOps/s	16.8436 KOps/s	$\color{#d91a1a}-0.48\%$
test_compile_copy_nested[pytree-compile]	0.1841ms	79.2504μs	12.6182 KOps/s	12.4856 KOps/s	$\color{#35bf28}+1.06\%$
test_compile_copy_nested[pytree-eager]	0.1217ms	66.8844μs	14.9512 KOps/s	14.4961 KOps/s	$\color{#35bf28}+3.14\%$
test_compile_add_one_flat[tensordict-compile]	0.2004ms	0.1057ms	9.4627 KOps/s	9.3111 KOps/s	$\color{#35bf28}+1.63\%$
test_compile_add_one_flat[tensordict-eager]	0.3856ms	0.2165ms	4.6196 KOps/s	4.6415 KOps/s	$\color{#d91a1a}-0.47\%$
test_compile_add_one_flat[tensorclass-compile]	0.1191ms	46.3217μs	21.5882 KOps/s	21.7262 KOps/s	$\color{#d91a1a}-0.64\%$
test_compile_add_one_flat[tensorclass-eager]	0.4782ms	66.2118μs	15.1031 KOps/s	15.6771 KOps/s	$\color{#d91a1a}-3.66\%$
test_compile_add_one_flat[pytree-compile]	0.1883ms	0.1049ms	9.5303 KOps/s	9.8052 KOps/s	$\color{#d91a1a}-2.80\%$
test_compile_add_one_flat[pytree-eager]	0.3137ms	0.2028ms	4.9315 KOps/s	4.9823 KOps/s	$\color{#d91a1a}-1.02\%$
test_compile_add_self_flat[tensordict-eager]	0.4047ms	0.2381ms	4.1997 KOps/s	4.2734 KOps/s	$\color{#d91a1a}-1.72\%$
test_compile_add_self_flat[tensordict-compile]	0.2192ms	0.1102ms	9.0738 KOps/s	9.3835 KOps/s	$\color{#d91a1a}-3.30\%$
test_compile_add_self_flat[tensorclass-eager]	0.2081ms	61.6018μs	16.2333 KOps/s	16.9484 KOps/s	$\color{#d91a1a}-4.22\%$
test_compile_add_self_flat[tensorclass-compile]	0.6302ms	46.3413μs	21.5790 KOps/s	21.9848 KOps/s	$\color{#d91a1a}-1.85\%$
test_compile_add_self_flat[pytree-eager]	0.5670ms	0.1582ms	6.3192 KOps/s	6.3745 KOps/s	$\color{#d91a1a}-0.87\%$
test_compile_add_self_flat[pytree-compile]	0.1819ms	0.1039ms	9.6263 KOps/s	9.7487 KOps/s	$\color{#d91a1a}-1.26\%$
test_compile_copy_flat[tensordict-compile]	66.2230μs	20.9059μs	47.8333 KOps/s	46.8017 KOps/s	$\color{#35bf28}+2.20\%$
test_compile_copy_flat[tensordict-eager]	0.1869ms	69.0696μs	14.4782 KOps/s	15.0720 KOps/s	$\color{#d91a1a}-3.94\%$
test_compile_copy_flat[pytree-compile]	0.1526ms	81.4036μs	12.2845 KOps/s	12.1129 KOps/s	$\color{#35bf28}+1.42\%$
test_compile_copy_flat[pytree-eager]	0.1275ms	69.4040μs	14.4084 KOps/s	14.2296 KOps/s	$\color{#35bf28}+1.26\%$
test_compile_assign_and_add[tensordict-compile]	0.2973ms	0.2077ms	4.8144 KOps/s	4.9115 KOps/s	$\color{#d91a1a}-1.98\%$
test_compile_assign_and_add[tensordict-eager]	1.7441ms	1.3551ms	737.9644 Ops/s	767.7678 Ops/s	$\color{#d91a1a}-3.88\%$
test_compile_assign_and_add[pytree-compile]	0.3445ms	0.2058ms	4.8582 KOps/s	5.0184 KOps/s	$\color{#d91a1a}-3.19\%$
test_compile_assign_and_add[pytree-eager]	1.0298ms	0.7781ms	1.2851 KOps/s	1.3159 KOps/s	$\color{#d91a1a}-2.34\%$
test_compile_assign_and_add_stack[compile]	0.5505ms	0.4539ms	2.2032 KOps/s	2.2073 KOps/s	$\color{#d91a1a}-0.18\%$
test_compile_assign_and_add_stack[eager]	3.4719ms	2.8530ms	350.5109 Ops/s	397.3091 Ops/s	$\textbf{\color{#d91a1a}-11.78\%}$
test_compile_indexing[tensor-tensordict-compile]	89.0550μs	35.7888μs	27.9417 KOps/s	27.6866 KOps/s	$\color{#35bf28}+0.92\%$
test_compile_indexing[tensor-tensordict-eager]	0.5046ms	34.8341μs	28.7075 KOps/s	30.7524 KOps/s	$\textbf{\color{#d91a1a}-6.65\%}$
test_compile_indexing[tensor-tensorclass-compile]	85.3390μs	28.8157μs	34.7034 KOps/s	33.6261 KOps/s	$\color{#35bf28}+3.20\%$
test_compile_indexing[tensor-tensorclass-eager]	63.1980μs	23.7396μs	42.1238 KOps/s	43.1039 KOps/s	$\color{#d91a1a}-2.27\%$
test_compile_indexing[tensor-pytree-compile]	86.5020μs	29.3176μs	34.1092 KOps/s	32.6862 KOps/s	$\color{#35bf28}+4.35\%$
test_compile_indexing[tensor-pytree-eager]	62.9070μs	23.6513μs	42.2810 KOps/s	43.0006 KOps/s	$\color{#d91a1a}-1.67\%$
test_compile_indexing[slice-tensordict-compile]	0.1144ms	51.6897μs	19.3462 KOps/s	19.2085 KOps/s	$\color{#35bf28}+0.72\%$
test_compile_indexing[slice-tensordict-eager]	0.6323ms	21.0795μs	47.4395 KOps/s	49.7339 KOps/s	$\color{#d91a1a}-4.61\%$
test_compile_indexing[slice-tensorclass-compile]	90.3680μs	44.7894μs	22.3267 KOps/s	22.5137 KOps/s	$\color{#d91a1a}-0.83\%$
test_compile_indexing[slice-tensorclass-eager]	55.8440μs	19.1388μs	52.2500 KOps/s	52.3170 KOps/s	$\color{#d91a1a}-0.13\%$
test_compile_indexing[slice-pytree-compile]	0.1075ms	45.5828μs	21.9381 KOps/s	22.3765 KOps/s	$\color{#d91a1a}-1.96\%$
test_compile_indexing[slice-pytree-eager]	53.6400μs	19.0920μs	52.3780 KOps/s	53.3801 KOps/s	$\color{#d91a1a}-1.88\%$
test_compile_indexing[int-tensordict-compile]	0.1181ms	53.2007μs	18.7967 KOps/s	19.1123 KOps/s	$\color{#d91a1a}-1.65\%$
test_compile_indexing[int-tensordict-eager]	1.0372ms	20.7369μs	48.2232 KOps/s	49.8400 KOps/s	$\color{#d91a1a}-3.24\%$
test_compile_indexing[int-tensorclass-compile]	97.2210μs	45.3591μs	22.0463 KOps/s	22.1314 KOps/s	$\color{#d91a1a}-0.38\%$
test_compile_indexing[int-tensorclass-eager]	50.2140μs	19.0929μs	52.3754 KOps/s	52.7099 KOps/s	$\color{#d91a1a}-0.63\%$
test_compile_indexing[int-pytree-compile]	0.1135ms	45.2173μs	22.1154 KOps/s	22.4261 KOps/s	$\color{#d91a1a}-1.39\%$
test_compile_indexing[int-pytree-eager]	70.3510μs	18.8099μs	53.1634 KOps/s	52.4611 KOps/s	$\color{#35bf28}+1.34\%$
test_mod_add[eager]	89.8070μs	35.6456μs	28.0540 KOps/s	31.1388 KOps/s	$\textbf{\color{#d91a1a}-9.91\%}$
test_mod_add[compile]	0.1077ms	47.4500μs	21.0748 KOps/s	20.6978 KOps/s	$\color{#35bf28}+1.82\%$
test_mod_add[compile-overhead]	0.1341ms	47.6796μs	20.9733 KOps/s	20.7871 KOps/s	$\color{#35bf28}+0.90\%$
test_mod_wrap[eager]	0.4558ms	0.2261ms	4.4222 KOps/s	4.5681 KOps/s	$\color{#d91a1a}-3.19\%$
test_mod_wrap[compile]	0.4436ms	0.2189ms	4.5680 KOps/s	4.8991 KOps/s	$\textbf{\color{#d91a1a}-6.76\%}$
test_mod_wrap[compile-overhead]	0.4186ms	0.2074ms	4.8207 KOps/s	4.8474 KOps/s	$\color{#d91a1a}-0.55\%$
test_mod_wrap_and_backward[eager]	18.5038ms	11.8347ms	84.4971 Ops/s	88.1358 Ops/s	$\color{#d91a1a}-4.13\%$
test_mod_wrap_and_backward[compile]	17.7854ms	12.6738ms	78.9031 Ops/s	79.9234 Ops/s	$\color{#d91a1a}-1.28\%$
test_mod_wrap_and_backward[compile-overhead]	18.1711ms	12.5378ms	79.7590 Ops/s	79.5108 Ops/s	$\color{#35bf28}+0.31\%$
test_seq_add[eager]	0.2514ms	0.1187ms	8.4215 KOps/s	8.9715 KOps/s	$\textbf{\color{#d91a1a}-6.13\%}$
test_seq_add[compile]	0.1289ms	63.0348μs	15.8642 KOps/s	15.6342 KOps/s	$\color{#35bf28}+1.47\%$
test_seq_add[compile-overhead]	0.1305ms	60.3032μs	16.5829 KOps/s	16.2821 KOps/s	$\color{#35bf28}+1.85\%$
test_seq_wrap[eager]	0.6984ms	0.4460ms	2.2422 KOps/s	2.3753 KOps/s	$\textbf{\color{#d91a1a}-5.61\%}$
test_seq_wrap[compile]	0.3455ms	0.2292ms	4.3632 KOps/s	4.3847 KOps/s	$\color{#d91a1a}-0.49\%$
test_seq_wrap[compile-overhead]	0.3589ms	0.2279ms	4.3872 KOps/s	4.4201 KOps/s	$\color{#d91a1a}-0.74\%$
test_func_call_runtime[False-eager]	0.9787ms	0.5603ms	1.7849 KOps/s	1.8429 KOps/s	$\color{#d91a1a}-3.15\%$
test_func_call_runtime[False-compile]	0.8218ms	0.4336ms	2.3062 KOps/s	2.4051 KOps/s	$\color{#d91a1a}-4.11\%$
test_func_call_runtime[False-compile-overhead]	0.8054ms	0.4323ms	2.3132 KOps/s	2.3777 KOps/s	$\color{#d91a1a}-2.71\%$
test_func_call_runtime[True-eager]	1.2615ms	0.7692ms	1.3000 KOps/s	1.3289 KOps/s	$\color{#d91a1a}-2.17\%$
test_func_call_runtime[True-compile]	0.8367ms	0.4702ms	2.1266 KOps/s	2.1606 KOps/s	$\color{#d91a1a}-1.57\%$
test_func_call_runtime[True-compile-overhead]	0.5929ms	0.4700ms	2.1276 KOps/s	2.1689 KOps/s	$\color{#d91a1a}-1.90\%$
test_func_call_cm_runtime[False-eager]	0.9298ms	0.5527ms	1.8092 KOps/s	1.8556 KOps/s	$\color{#d91a1a}-2.50\%$
test_func_call_cm_runtime[False-compile]	0.5425ms	0.4285ms	2.3336 KOps/s	2.3830 KOps/s	$\color{#d91a1a}-2.07\%$
test_func_call_cm_runtime[False-compile-overhead]	0.5835ms	0.4284ms	2.3342 KOps/s	2.3900 KOps/s	$\color{#d91a1a}-2.34\%$
test_func_call_cm_runtime[True-eager]	1.4780ms	0.9118ms	1.0967 KOps/s	1.1167 KOps/s	$\color{#d91a1a}-1.79\%$
test_func_call_cm_runtime[True-compile]	0.8727ms	0.4945ms	2.0222 KOps/s	2.0512 KOps/s	$\color{#d91a1a}-1.42\%$
test_func_call_cm_runtime[True-compile-overhead]	0.8554ms	0.4972ms	2.0111 KOps/s	2.0674 KOps/s	$\color{#d91a1a}-2.72\%$
test_vmap_func_call_cm_runtime[eager]	2.8388ms	1.9272ms	518.8976 Ops/s	525.5861 Ops/s	$\color{#d91a1a}-1.27\%$
test_vmap_func_call_cm_runtime[compile]	0.9042ms	0.5249ms	1.9050 KOps/s	1.9304 KOps/s	$\color{#d91a1a}-1.32\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.6972ms	0.5236ms	1.9099 KOps/s	1.9402 KOps/s	$\color{#d91a1a}-1.56\%$
test_distributed	0.2557ms	0.1264ms	7.9085 KOps/s	7.8330 KOps/s	$\color{#35bf28}+0.96\%$
test_tdmodule	0.1221ms	26.7583μs	37.3715 KOps/s	39.4510 KOps/s	$\textbf{\color{#d91a1a}-5.27\%}$
test_tdmodule_dispatch	82.7050μs	48.7516μs	20.5121 KOps/s	21.7825 KOps/s	$\textbf{\color{#d91a1a}-5.83\%}$
test_tdseq	60.4130μs	29.2122μs	34.2322 KOps/s	35.4830 KOps/s	$\color{#d91a1a}-3.52\%$
test_tdseq_dispatch	88.5650μs	54.9177μs	18.2091 KOps/s	19.1863 KOps/s	$\textbf{\color{#d91a1a}-5.09\%}$
test_instantiation_functorch	2.0519ms	1.5848ms	630.9972 Ops/s	650.7353 Ops/s	$\color{#d91a1a}-3.03\%$
test_exec_functorch	0.2797ms	0.1822ms	5.4872 KOps/s	5.7314 KOps/s	$\color{#d91a1a}-4.26\%$
test_exec_functional_call	0.3120ms	0.1765ms	5.6661 KOps/s	5.9956 KOps/s	$\textbf{\color{#d91a1a}-5.50\%}$
test_exec_td_decorator	0.4668ms	0.2370ms	4.2200 KOps/s	4.3871 KOps/s	$\color{#d91a1a}-3.81\%$
test_vmap_mlp_speed_decorator[True-True]	1.2409ms	0.6694ms	1.4938 KOps/s	1.5313 KOps/s	$\color{#d91a1a}-2.45\%$
test_vmap_mlp_speed_decorator[True-False]	0.8693ms	0.6631ms	1.5080 KOps/s	1.5541 KOps/s	$\color{#d91a1a}-2.97\%$
test_vmap_mlp_speed_decorator[False-True]	0.7308ms	0.5354ms	1.8678 KOps/s	1.9094 KOps/s	$\color{#d91a1a}-2.18\%$
test_vmap_mlp_speed_decorator[False-False]	0.8585ms	0.5379ms	1.8592 KOps/s	1.9004 KOps/s	$\color{#d91a1a}-2.17\%$
test_to_module_speed[True]	1.5394ms	1.3433ms	744.4354 Ops/s	742.7790 Ops/s	$\color{#35bf28}+0.22\%$
test_to_module_speed[False]	1.5678ms	1.3196ms	757.8049 Ops/s	761.5808 Ops/s	$\color{#d91a1a}-0.50\%$
test_tc_init	0.1266ms	49.9162μs	20.0336 KOps/s	22.4835 KOps/s	$\textbf{\color{#d91a1a}-10.90\%}$
test_tc_init_nested	0.1535ms	95.7153μs	10.4477 KOps/s	11.6603 KOps/s	$\textbf{\color{#d91a1a}-10.40\%}$
test_tc_first_layer_tensor	16.1000μs	1.5962μs	626.4991 KOps/s	651.0546 KOps/s	$\color{#d91a1a}-3.77\%$
test_tc_first_layer_nontensor	36.9280μs	4.7918μs	208.6885 KOps/s	214.1258 KOps/s	$\color{#d91a1a}-2.54\%$
test_tc_second_layer_tensor	43.7710μs	2.9261μs	341.7542 KOps/s	349.9554 KOps/s	$\color{#d91a1a}-2.34\%$
test_tc_second_layer_nontensor	27.2510μs	6.1724μs	162.0106 KOps/s	167.0392 KOps/s	$\color{#d91a1a}-3.01\%$
test_unbind	0.2205s	15.0916ms	66.2622 Ops/s	78.2721 Ops/s	$\textbf{\color{#d91a1a}-15.34\%}$
test_full_like	9.9184ms	7.5308ms	132.7888 Ops/s	71.3174 Ops/s	$\textbf{\color{#35bf28}+86.19\%}$
test_zeros_like	3.8349ms	2.9953ms	333.8515 Ops/s	134.6580 Ops/s	$\textbf{\color{#35bf28}+147.93\%}$
test_ones_like	5.0114ms	3.5879ms	278.7138 Ops/s	122.6627 Ops/s	$\textbf{\color{#35bf28}+127.22\%}$
test_clone	7.5466ms	5.5375ms	180.5854 Ops/s	104.5356 Ops/s	$\textbf{\color{#35bf28}+72.75\%}$
test_squeeze	65.6720μs	12.2545μs	81.6029 KOps/s	84.2360 KOps/s	$\color{#d91a1a}-3.13\%$
test_unsqueeze	0.1679ms	94.1699μs	10.6191 KOps/s	10.8648 KOps/s	$\color{#d91a1a}-2.26\%$
test_split	0.5857ms	0.2013ms	4.9666 KOps/s	5.1904 KOps/s	$\color{#d91a1a}-4.31\%$
test_permute	0.4045ms	0.2141ms	4.6711 KOps/s	4.8359 KOps/s	$\color{#d91a1a}-3.41\%$
test_stack	31.5510ms	25.2450ms	39.6118 Ops/s	38.3302 Ops/s	$\color{#35bf28}+3.34\%$
test_cat	31.5497ms	24.9342ms	40.1055 Ops/s	39.7962 Ops/s	$\color{#35bf28}+0.78\%$

[github-actions]

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

Total Benchmarks: 229. Improved: $\large\color{#35bf28}62$. Worsened: $\large\color{#d91a1a}7$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	31.9200μs	11.5810μs	86.3487 KOps/s	78.0322 KOps/s	$\textbf{\color{#35bf28}+10.66\%}$
test_plain_set_stack_nested	32.1800μs	11.8552μs	84.3509 KOps/s	77.1647 KOps/s	$\textbf{\color{#35bf28}+9.31\%}$
test_plain_set_nested_inplace	34.3010μs	12.7964μs	78.1470 KOps/s	71.9067 KOps/s	$\textbf{\color{#35bf28}+8.68\%}$
test_plain_set_stack_nested_inplace	33.8910μs	12.6555μs	79.0172 KOps/s	71.6669 KOps/s	$\textbf{\color{#35bf28}+10.26\%}$
test_items	20.9600μs	2.8997μs	344.8632 KOps/s	344.2328 KOps/s	$\color{#35bf28}+0.18\%$
test_items_nested	0.4840ms	0.3628ms	2.7565 KOps/s	2.8037 KOps/s	$\color{#d91a1a}-1.68\%$
test_items_nested_locked	0.4016ms	0.3632ms	2.7532 KOps/s	2.7807 KOps/s	$\color{#d91a1a}-0.99\%$
test_items_nested_leaf	83.6310μs	59.1970μs	16.8927 KOps/s	16.9269 KOps/s	$\color{#d91a1a}-0.20\%$
test_items_stack_nested	0.3954ms	0.3674ms	2.7215 KOps/s	2.7477 KOps/s	$\color{#d91a1a}-0.95\%$
test_items_stack_nested_leaf	88.9210μs	61.5065μs	16.2584 KOps/s	16.6128 KOps/s	$\color{#d91a1a}-2.13\%$
test_items_stack_nested_locked	0.3970ms	0.3668ms	2.7265 KOps/s	2.7536 KOps/s	$\color{#d91a1a}-0.98\%$
test_keys	22.7700μs	3.4995μs	285.7533 KOps/s	289.4395 KOps/s	$\color{#d91a1a}-1.27\%$
test_keys_nested	0.1109ms	81.5377μs	12.2643 KOps/s	12.3048 KOps/s	$\color{#d91a1a}-0.33\%$
test_keys_nested_locked	0.8283ms	85.7745μs	11.6585 KOps/s	11.4837 KOps/s	$\color{#35bf28}+1.52\%$
test_keys_nested_leaf	2.4591ms	72.3896μs	13.8141 KOps/s	13.9386 KOps/s	$\color{#d91a1a}-0.89\%$
test_keys_stack_nested	0.1070ms	81.9407μs	12.2039 KOps/s	11.8830 KOps/s	$\color{#35bf28}+2.70\%$
test_keys_stack_nested_leaf	97.0620μs	73.3438μs	13.6344 KOps/s	13.3885 KOps/s	$\color{#35bf28}+1.84\%$
test_keys_stack_nested_locked	0.1337ms	87.8783μs	11.3794 KOps/s	11.1518 KOps/s	$\color{#35bf28}+2.04\%$
test_values	5.3985μs	0.8549μs	1.1697 MOps/s	1.1654 MOps/s	$\color{#35bf28}+0.37\%$
test_values_nested	54.2300μs	34.8351μs	28.7067 KOps/s	28.9358 KOps/s	$\color{#d91a1a}-0.79\%$
test_values_nested_locked	61.8410μs	36.0988μs	27.7017 KOps/s	27.5624 KOps/s	$\color{#35bf28}+0.51\%$
test_values_nested_leaf	67.0910μs	39.2313μs	25.4898 KOps/s	25.4928 KOps/s	$\color{#d91a1a}-0.01\%$
test_values_stack_nested	75.6910μs	34.9375μs	28.6225 KOps/s	28.3641 KOps/s	$\color{#35bf28}+0.91\%$
test_values_stack_nested_leaf	76.9410μs	39.5764μs	25.2676 KOps/s	25.1888 KOps/s	$\color{#35bf28}+0.31\%$
test_values_stack_nested_locked	67.8510μs	36.6601μs	27.2776 KOps/s	27.0618 KOps/s	$\color{#35bf28}+0.80\%$
test_membership	1.8215μs	0.5202μs	1.9223 MOps/s	1.9486 MOps/s	$\color{#d91a1a}-1.35\%$
test_membership_nested	29.8910μs	2.0965μs	476.9938 KOps/s	471.7162 KOps/s	$\color{#35bf28}+1.12\%$
test_membership_nested_leaf	15.9350μs	2.0096μs	497.6038 KOps/s	491.1910 KOps/s	$\color{#35bf28}+1.31\%$
test_membership_stacked_nested	47.4900μs	2.1027μs	475.5694 KOps/s	463.6338 KOps/s	$\color{#35bf28}+2.57\%$
test_membership_stacked_nested_leaf	31.3410μs	2.0822μs	480.2601 KOps/s	476.6002 KOps/s	$\color{#35bf28}+0.77\%$
test_membership_nested_last	32.4210μs	3.1054μs	322.0221 KOps/s	324.3384 KOps/s	$\color{#d91a1a}-0.71\%$
test_membership_nested_leaf_last	31.1900μs	3.0975μs	322.8447 KOps/s	315.9784 KOps/s	$\color{#35bf28}+2.17\%$
test_membership_stacked_nested_last	46.7910μs	3.6182μs	276.3821 KOps/s	320.4616 KOps/s	$\textbf{\color{#d91a1a}-13.76\%}$
test_membership_stacked_nested_leaf_last	26.8410μs	3.5629μs	280.6670 KOps/s	321.6560 KOps/s	$\textbf{\color{#d91a1a}-12.74\%}$
test_nested_getleaf	39.3500μs	6.0683μs	164.7898 KOps/s	163.3576 KOps/s	$\color{#35bf28}+0.88\%$
test_nested_get	39.1810μs	5.8815μs	170.0254 KOps/s	172.1438 KOps/s	$\color{#d91a1a}-1.23\%$
test_stacked_getleaf	33.3500μs	6.1348μs	163.0058 KOps/s	161.7016 KOps/s	$\color{#35bf28}+0.81\%$
test_stacked_get	35.4200μs	5.8548μs	170.7995 KOps/s	171.9199 KOps/s	$\color{#d91a1a}-0.65\%$
test_nested_getitemleaf	86.1210μs	6.3183μs	158.2700 KOps/s	159.9705 KOps/s	$\color{#d91a1a}-1.06\%$
test_nested_getitem	38.3800μs	5.8832μs	169.9748 KOps/s	168.5819 KOps/s	$\color{#35bf28}+0.83\%$
test_stacked_getitemleaf	27.5500μs	6.2313μs	160.4807 KOps/s	160.3193 KOps/s	$\color{#35bf28}+0.10\%$
test_stacked_getitem	33.0310μs	5.8979μs	169.5518 KOps/s	168.8157 KOps/s	$\color{#35bf28}+0.44\%$
test_lock_nested	2.3644ms	0.3788ms	2.6396 KOps/s	2.6742 KOps/s	$\color{#d91a1a}-1.29\%$
test_lock_stack_nested	0.3903ms	0.3476ms	2.8771 KOps/s	2.8604 KOps/s	$\color{#35bf28}+0.58\%$
test_unlock_nested	0.6712ms	0.3161ms	3.1638 KOps/s	3.1403 KOps/s	$\color{#35bf28}+0.75\%$
test_unlock_stack_nested	0.3427ms	0.2865ms	3.4902 KOps/s	3.4771 KOps/s	$\color{#35bf28}+0.38\%$
test_flatten_speed	0.1248ms	75.5681μs	13.2331 KOps/s	13.3965 KOps/s	$\color{#d91a1a}-1.22\%$
test_unflatten_speed	0.3631ms	0.3191ms	3.1339 KOps/s	3.0652 KOps/s	$\color{#35bf28}+2.24\%$
test_common_ops	1.5321ms	0.5902ms	1.6943 KOps/s	1.5735 KOps/s	$\textbf{\color{#35bf28}+7.68\%}$
test_creation	0.1272ms	1.7430μs	573.7153 KOps/s	572.0595 KOps/s	$\color{#35bf28}+0.29\%$
test_creation_empty	39.8800μs	7.0503μs	141.8386 KOps/s	106.8111 KOps/s	$\textbf{\color{#35bf28}+32.79\%}$
test_creation_nested_1	33.1300μs	8.7508μs	114.2751 KOps/s	90.1545 KOps/s	$\textbf{\color{#35bf28}+26.75\%}$
test_creation_nested_2	48.9010μs	11.4340μs	87.4588 KOps/s	71.6002 KOps/s	$\textbf{\color{#35bf28}+22.15\%}$
test_clone	37.7500μs	10.9849μs	91.0337 KOps/s	93.7749 KOps/s	$\color{#d91a1a}-2.92\%$
test_getitem[int]	1.2967ms	10.9822μs	91.0567 KOps/s	92.8675 KOps/s	$\color{#d91a1a}-1.95\%$
test_getitem[slice_int]	0.1043ms	21.5910μs	46.3155 KOps/s	47.5745 KOps/s	$\color{#d91a1a}-2.65\%$
test_getitem[range]	0.1258ms	37.3665μs	26.7619 KOps/s	26.9623 KOps/s	$\color{#d91a1a}-0.74\%$
test_getitem[tuple]	0.1162ms	18.3482μs	54.5014 KOps/s	55.1638 KOps/s	$\color{#d91a1a}-1.20\%$
test_getitem[list]	0.1310ms	33.8179μs	29.5702 KOps/s	31.0787 KOps/s	$\color{#d91a1a}-4.85\%$
test_setitem_dim[int]	52.2810μs	19.0133μs	52.5948 KOps/s	53.2200 KOps/s	$\color{#d91a1a}-1.17\%$
test_setitem_dim[slice_int]	65.2610μs	39.0678μs	25.5965 KOps/s	25.7278 KOps/s	$\color{#d91a1a}-0.51\%$
test_setitem_dim[range]	78.0710μs	53.1712μs	18.8072 KOps/s	18.8772 KOps/s	$\color{#d91a1a}-0.37\%$
test_setitem_dim[tuple]	52.6110μs	32.7586μs	30.5264 KOps/s	30.5054 KOps/s	$\color{#35bf28}+0.07\%$
test_setitem	39.8100μs	14.9014μs	67.1076 KOps/s	62.4946 KOps/s	$\textbf{\color{#35bf28}+7.38\%}$
test_set	0.1102ms	14.2981μs	69.9396 KOps/s	64.8961 KOps/s	$\textbf{\color{#35bf28}+7.77\%}$
test_set_shared	1.4656ms	0.1544ms	6.4768 KOps/s	6.5125 KOps/s	$\color{#d91a1a}-0.55\%$
test_update	0.4809ms	16.4460μs	60.8052 KOps/s	52.5554 KOps/s	$\textbf{\color{#35bf28}+15.70\%}$
test_update_nested	0.1116ms	21.7484μs	45.9805 KOps/s	39.6300 KOps/s	$\textbf{\color{#35bf28}+16.02\%}$
test_update__nested	0.5117ms	26.3716μs	37.9195 KOps/s	38.5446 KOps/s	$\color{#d91a1a}-1.62\%$
test_set_nested	0.1053ms	15.6984μs	63.7008 KOps/s	59.6106 KOps/s	$\textbf{\color{#35bf28}+6.86\%}$
test_set_nested_new	0.1134ms	19.4429μs	51.4327 KOps/s	52.4635 KOps/s	$\color{#d91a1a}-1.96\%$
test_select	0.2117ms	31.6616μs	31.5840 KOps/s	31.1518 KOps/s	$\color{#35bf28}+1.39\%$
test_select_nested	76.9510μs	43.8946μs	22.7819 KOps/s	22.1929 KOps/s	$\color{#35bf28}+2.65\%$
test_exclude_nested	95.7210μs	64.4857μs	15.5073 KOps/s	15.2444 KOps/s	$\color{#35bf28}+1.72\%$
test_empty[True]	0.3634ms	0.2886ms	3.4656 KOps/s	3.4423 KOps/s	$\color{#35bf28}+0.68\%$
test_empty[False]	3.6050μs	0.8794μs	1.1371 MOps/s	1.1268 MOps/s	$\color{#35bf28}+0.91\%$
test_to	85.9710μs	55.5945μs	17.9874 KOps/s	17.8097 KOps/s	$\color{#35bf28}+1.00\%$
test_to_nonblocking	0.1001ms	47.8488μs	20.8992 KOps/s	20.5748 KOps/s	$\color{#35bf28}+1.58\%$
test_unbind_speed	1.3167ms	0.2413ms	4.1447 KOps/s	4.0966 KOps/s	$\color{#35bf28}+1.17\%$
test_unbind_speed_stack0	0.2985ms	0.2388ms	4.1873 KOps/s	4.1201 KOps/s	$\color{#35bf28}+1.63\%$
test_unbind_speed_stack1	92.7464ms	0.6705ms	1.4915 KOps/s	1.4725 KOps/s	$\color{#35bf28}+1.29\%$
test_split	0.1023s	1.6410ms	609.3768 Ops/s	580.4213 Ops/s	$\color{#35bf28}+4.99\%$
test_chunk	94.5760ms	1.6440ms	608.2829 Ops/s	689.6312 Ops/s	$\textbf{\color{#d91a1a}-11.80\%}$
test_consolidate[False-None]	97.0657ms	2.9420ms	339.9036 Ops/s	338.7293 Ops/s	$\color{#35bf28}+0.35\%$
test_consolidate[default-None]	1.8557ms	1.6606ms	602.1899 Ops/s	587.1979 Ops/s	$\color{#35bf28}+2.55\%$
test_consolidate[reduce-overhead-None]	1.8157ms	1.6654ms	600.4476 Ops/s	582.3687 Ops/s	$\color{#35bf28}+3.10\%$
test_consolidate_njt[False-None]	7.0018ms	6.4374ms	155.3426 Ops/s	152.2004 Ops/s	$\color{#35bf28}+2.06\%$
test_to[False-False-None]	1.7558ms	1.6903ms	591.6038 Ops/s	575.6510 Ops/s	$\color{#35bf28}+2.77\%$
test_to[True-False-None]	1.5757ms	1.2922ms	773.8556 Ops/s	769.0263 Ops/s	$\color{#35bf28}+0.63\%$
test_to[within-False-None]	4.1775ms	4.0587ms	246.3849 Ops/s	241.0604 Ops/s	$\color{#35bf28}+2.21\%$
test_to[True-default-None]	5.4672ms	5.2765ms	189.5183 Ops/s	187.6885 Ops/s	$\color{#35bf28}+0.97\%$
test_to_njt[False-False-None]	7.3878ms	7.1939ms	139.0063 Ops/s	142.5187 Ops/s	$\color{#d91a1a}-2.46\%$
test_to_njt[True-False-None]	5.7487ms	5.4789ms	182.5194 Ops/s	178.3925 Ops/s	$\color{#35bf28}+2.31\%$
test_to_njt[within-False-None]	12.2645ms	12.1485ms	82.3147 Ops/s	72.2224 Ops/s	$\textbf{\color{#35bf28}+13.97\%}$
test_creation[device0]	0.5495ms	81.1264μs	12.3265 KOps/s	11.2823 KOps/s	$\textbf{\color{#35bf28}+9.25\%}$
test_creation_from_tensor	0.5017ms	86.4161μs	11.5719 KOps/s	11.0408 KOps/s	$\color{#35bf28}+4.81\%$
test_add_one[memmap_tensor0]	0.3691ms	6.8015μs	147.0270 KOps/s	142.9026 KOps/s	$\color{#35bf28}+2.89\%$
test_contiguous[memmap_tensor0]	2.1380μs	0.4167μs	2.4000 MOps/s	2.4434 MOps/s	$\color{#d91a1a}-1.78\%$
test_stack[memmap_tensor0]	38.5510μs	4.3727μs	228.6923 KOps/s	231.5721 KOps/s	$\color{#d91a1a}-1.24\%$
test_memmaptd_index	1.7255ms	0.2484ms	4.0251 KOps/s	3.9610 KOps/s	$\color{#35bf28}+1.62\%$
test_memmaptd_index_astensor	0.5828ms	0.3094ms	3.2325 KOps/s	3.1156 KOps/s	$\color{#35bf28}+3.75\%$
test_memmaptd_index_op	1.0012ms	0.5616ms	1.7805 KOps/s	1.6743 KOps/s	$\textbf{\color{#35bf28}+6.34\%}$
test_serialize_model	0.1321s	0.1314s	7.6118 Ops/s	7.5890 Ops/s	$\color{#35bf28}+0.30\%$
test_serialize_model_pickle	1.3460s	1.1876s	0.8420 Ops/s	0.8209 Ops/s	$\color{#35bf28}+2.58\%$
test_serialize_weights	0.1325s	0.1310s	7.6312 Ops/s	7.6661 Ops/s	$\color{#d91a1a}-0.46\%$
test_serialize_weights_returnearly	0.3336s	63.4441ms	15.7619 Ops/s	14.4756 Ops/s	$\textbf{\color{#35bf28}+8.89\%}$
test_serialize_weights_pickle	1.3739s	1.2163s	0.8221 Ops/s	0.8224 Ops/s	$\color{#d91a1a}-0.03\%$
test_reshape_pytree	51.9500μs	22.3578μs	44.7271 KOps/s	44.5452 KOps/s	$\color{#35bf28}+0.41\%$
test_reshape_td	59.6710μs	26.9967μs	37.0415 KOps/s	35.7033 KOps/s	$\color{#35bf28}+3.75\%$
test_view_pytree	50.1110μs	21.6009μs	46.2944 KOps/s	44.9342 KOps/s	$\color{#35bf28}+3.03\%$
test_view_td	59.8610μs	30.3092μs	32.9933 KOps/s	29.2340 KOps/s	$\textbf{\color{#35bf28}+12.86\%}$
test_unbind_pytree	50.2710μs	27.8523μs	35.9036 KOps/s	35.2426 KOps/s	$\color{#35bf28}+1.88\%$
test_unbind_td	0.6800ms	37.0917μs	26.9602 KOps/s	26.6820 KOps/s	$\color{#35bf28}+1.04\%$
test_split_pytree	84.6310μs	29.8697μs	33.4788 KOps/s	32.6548 KOps/s	$\color{#35bf28}+2.52\%$
test_split_td	0.8649ms	37.5477μs	26.6328 KOps/s	25.4148 KOps/s	$\color{#35bf28}+4.79\%$
test_add_pytree	67.4510μs	35.3247μs	28.3088 KOps/s	29.1597 KOps/s	$\color{#d91a1a}-2.92\%$
test_add_td	80.0600μs	49.3906μs	20.2468 KOps/s	18.9396 KOps/s	$\textbf{\color{#35bf28}+6.90\%}$
test_compile_add_one_nested[tensordict-compile]	0.1742ms	0.1199ms	8.3376 KOps/s	8.1252 KOps/s	$\color{#35bf28}+2.61\%$
test_compile_add_one_nested[tensordict-eager]	0.2254ms	0.1309ms	7.6415 KOps/s	7.5766 KOps/s	$\color{#35bf28}+0.86\%$
test_compile_add_one_nested[pytree-compile]	0.1315ms	96.1276μs	10.4028 KOps/s	10.1436 KOps/s	$\color{#35bf28}+2.56\%$
test_compile_add_one_nested[pytree-eager]	0.3379ms	0.1480ms	6.7550 KOps/s	6.4166 KOps/s	$\textbf{\color{#35bf28}+5.27\%}$
test_compile_copy_nested[tensordict-compile]	52.4910μs	22.5080μs	44.4286 KOps/s	43.7646 KOps/s	$\color{#35bf28}+1.52\%$
test_compile_copy_nested[tensordict-eager]	69.2300μs	29.6788μs	33.6940 KOps/s	33.2408 KOps/s	$\color{#35bf28}+1.36\%$
test_compile_copy_nested[pytree-compile]	0.4040ms	65.1314μs	15.3536 KOps/s	15.0758 KOps/s	$\color{#35bf28}+1.84\%$
test_compile_copy_nested[pytree-eager]	81.4810μs	49.9144μs	20.0343 KOps/s	19.9567 KOps/s	$\color{#35bf28}+0.39\%$
test_compile_add_one_flat[tensordict-compile]	0.1813ms	0.1393ms	7.1808 KOps/s	6.9355 KOps/s	$\color{#35bf28}+3.54\%$
test_compile_add_one_flat[tensordict-eager]	0.3023ms	0.2128ms	4.6990 KOps/s	4.6472 KOps/s	$\color{#35bf28}+1.12\%$
test_compile_add_one_flat[tensorclass-compile]	0.1363ms	97.3162μs	10.2758 KOps/s	9.6295 KOps/s	$\textbf{\color{#35bf28}+6.71\%}$
test_compile_add_one_flat[tensorclass-eager]	0.1087ms	53.4577μs	18.7064 KOps/s	18.2958 KOps/s	$\color{#35bf28}+2.24\%$
test_compile_add_one_flat[pytree-compile]	0.1706ms	0.1338ms	7.4749 KOps/s	7.3360 KOps/s	$\color{#35bf28}+1.89\%$
test_compile_add_one_flat[pytree-eager]	0.5219ms	0.4769ms	2.0967 KOps/s	2.0564 KOps/s	$\color{#35bf28}+1.96\%$
test_compile_add_self_flat[tensordict-eager]	0.3650ms	0.2544ms	3.9309 KOps/s	3.8534 KOps/s	$\color{#35bf28}+2.01\%$
test_compile_add_self_flat[tensordict-compile]	0.1849ms	0.1407ms	7.1077 KOps/s	7.0782 KOps/s	$\color{#35bf28}+0.42\%$
test_compile_add_self_flat[tensorclass-eager]	0.1432ms	64.3637μs	15.5367 KOps/s	15.4052 KOps/s	$\color{#35bf28}+0.85\%$
test_compile_add_self_flat[tensorclass-compile]	0.1378ms	97.3720μs	10.2699 KOps/s	10.0357 KOps/s	$\color{#35bf28}+2.33\%$
test_compile_add_self_flat[pytree-eager]	0.4483ms	0.3965ms	2.5222 KOps/s	2.2499 KOps/s	$\textbf{\color{#35bf28}+12.10\%}$
test_compile_add_self_flat[pytree-compile]	0.1897ms	0.1332ms	7.5076 KOps/s	7.3430 KOps/s	$\color{#35bf28}+2.24\%$
test_compile_copy_flat[tensordict-compile]	65.2400μs	18.6865μs	53.5147 KOps/s	53.8026 KOps/s	$\color{#d91a1a}-0.54\%$
test_compile_copy_flat[tensordict-eager]	62.4600μs	31.2137μs	32.0372 KOps/s	31.3977 KOps/s	$\color{#35bf28}+2.04\%$
test_compile_copy_flat[pytree-compile]	98.6310μs	71.5256μs	13.9810 KOps/s	14.0690 KOps/s	$\color{#d91a1a}-0.63\%$
test_compile_copy_flat[pytree-eager]	83.8610μs	52.1710μs	19.1677 KOps/s	19.0547 KOps/s	$\color{#35bf28}+0.59\%$
test_compile_assign_and_add[tensordict-compile]	1.6039ms	0.3871ms	2.5834 KOps/s	2.2439 KOps/s	$\textbf{\color{#35bf28}+15.13\%}$
test_compile_assign_and_add[tensordict-eager]	2.7966ms	2.6388ms	378.9656 Ops/s	375.7283 Ops/s	$\color{#35bf28}+0.86\%$
test_compile_assign_and_add[pytree-compile]	1.5617ms	0.3783ms	2.6435 KOps/s	2.2959 KOps/s	$\textbf{\color{#35bf28}+15.14\%}$
test_compile_assign_and_add[pytree-eager]	3.0672ms	2.6315ms	380.0088 Ops/s	380.8688 Ops/s	$\color{#d91a1a}-0.23\%$
test_compile_indexing[tensor-tensordict-compile]	0.5408ms	0.1135ms	8.8119 KOps/s	8.4167 KOps/s	$\color{#35bf28}+4.70\%$
test_compile_indexing[tensor-tensordict-eager]	0.5571ms	79.3188μs	12.6074 KOps/s	12.1770 KOps/s	$\color{#35bf28}+3.53\%$
test_compile_indexing[tensor-tensorclass-compile]	0.1570ms	0.1091ms	9.1626 KOps/s	9.1922 KOps/s	$\color{#d91a1a}-0.32\%$
test_compile_indexing[tensor-tensorclass-eager]	0.4904ms	70.9064μs	14.1031 KOps/s	13.5226 KOps/s	$\color{#35bf28}+4.29\%$
test_compile_indexing[tensor-pytree-compile]	0.5252ms	0.1121ms	8.9203 KOps/s	8.5657 KOps/s	$\color{#35bf28}+4.14\%$
test_compile_indexing[tensor-pytree-eager]	0.4921ms	71.1533μs	14.0542 KOps/s	13.2496 KOps/s	$\textbf{\color{#35bf28}+6.07\%}$
test_compile_indexing[slice-tensordict-compile]	0.1419ms	0.1023ms	9.7749 KOps/s	9.0016 KOps/s	$\textbf{\color{#35bf28}+8.59\%}$
test_compile_indexing[slice-tensordict-eager]	0.4433ms	17.4918μs	57.1695 KOps/s	42.6618 KOps/s	$\textbf{\color{#35bf28}+34.01\%}$
test_compile_indexing[slice-tensorclass-compile]	0.5276ms	97.2176μs	10.2862 KOps/s	9.2987 KOps/s	$\textbf{\color{#35bf28}+10.62\%}$
test_compile_indexing[slice-tensorclass-eager]	0.4219ms	16.2723μs	61.4540 KOps/s	56.3722 KOps/s	$\textbf{\color{#35bf28}+9.01\%}$
test_compile_indexing[slice-pytree-compile]	0.5147ms	98.0239μs	10.2016 KOps/s	9.2245 KOps/s	$\textbf{\color{#35bf28}+10.59\%}$
test_compile_indexing[slice-pytree-eager]	50.8900μs	16.0014μs	62.4943 KOps/s	56.3723 KOps/s	$\textbf{\color{#35bf28}+10.86\%}$
test_compile_indexing[int-tensordict-compile]	0.5183ms	0.1028ms	9.7264 KOps/s	8.8503 KOps/s	$\textbf{\color{#35bf28}+9.90\%}$
test_compile_indexing[int-tensordict-eager]	0.5705ms	16.9816μs	58.8872 KOps/s	50.9253 KOps/s	$\textbf{\color{#35bf28}+15.63\%}$
test_compile_indexing[int-tensorclass-compile]	0.5053ms	97.9179μs	10.2126 KOps/s	9.2642 KOps/s	$\textbf{\color{#35bf28}+10.24\%}$
test_compile_indexing[int-tensorclass-eager]	50.8610μs	15.9733μs	62.6044 KOps/s	55.4241 KOps/s	$\textbf{\color{#35bf28}+12.96\%}$
test_compile_indexing[int-pytree-compile]	0.5065ms	97.7447μs	10.2307 KOps/s	9.2579 KOps/s	$\textbf{\color{#35bf28}+10.51\%}$
test_compile_indexing[int-pytree-eager]	49.3310μs	15.8674μs	63.0221 KOps/s	55.8574 KOps/s	$\textbf{\color{#35bf28}+12.83\%}$
test_mod_add[eager]	83.9610μs	37.1592μs	26.9112 KOps/s	22.6949 KOps/s	$\textbf{\color{#35bf28}+18.58\%}$
test_mod_add[compile]	0.1314ms	79.1119μs	12.6403 KOps/s	11.1728 KOps/s	$\textbf{\color{#35bf28}+13.13\%}$
test_mod_add[compile-overhead]	0.3185ms	0.1680ms	5.9537 KOps/s	5.7415 KOps/s	$\color{#35bf28}+3.70\%$
test_mod_wrap[eager]	0.3485ms	0.2541ms	3.9362 KOps/s	3.5982 KOps/s	$\textbf{\color{#35bf28}+9.39\%}$
test_mod_wrap[compile]	0.3583ms	0.2825ms	3.5403 KOps/s	3.3914 KOps/s	$\color{#35bf28}+4.39\%$
test_mod_wrap[compile-overhead]	7.1035ms	3.7376ms	267.5484 Ops/s	284.0270 Ops/s	$\textbf{\color{#d91a1a}-5.80\%}$
test_mod_wrap_and_backward[eager]	1.4628ms	1.3637ms	733.3218 Ops/s	685.6287 Ops/s	$\textbf{\color{#35bf28}+6.96\%}$
test_mod_wrap_and_backward[compile]	1.4035ms	1.2697ms	787.5975 Ops/s	731.0145 Ops/s	$\textbf{\color{#35bf28}+7.74\%}$
test_mod_wrap_and_backward[compile-overhead]	1.3929ms	0.9451ms	1.0580 KOps/s	917.4492 Ops/s	$\textbf{\color{#35bf28}+15.32\%}$
test_seq_add[eager]	0.1737ms	0.1144ms	8.7427 KOps/s	7.9733 KOps/s	$\textbf{\color{#35bf28}+9.65\%}$
test_seq_add[compile]	0.2239ms	88.2074μs	11.3369 KOps/s	10.6016 KOps/s	$\textbf{\color{#35bf28}+6.94\%}$
test_seq_add[compile-overhead]	0.2220ms	0.1308ms	7.6478 KOps/s	7.4455 KOps/s	$\color{#35bf28}+2.72\%$
test_seq_wrap[eager]	0.4757ms	0.4141ms	2.4150 KOps/s	2.2568 KOps/s	$\textbf{\color{#35bf28}+7.01\%}$
test_seq_wrap[compile]	0.3512ms	0.2982ms	3.3532 KOps/s	3.1383 KOps/s	$\textbf{\color{#35bf28}+6.85\%}$
test_seq_wrap[compile-overhead]	0.3031ms	0.2252ms	4.4396 KOps/s	4.4258 KOps/s	$\color{#35bf28}+0.31\%$
test_func_call_runtime[False-eager]	0.8095ms	0.7347ms	1.3612 KOps/s	1.2808 KOps/s	$\textbf{\color{#35bf28}+6.27\%}$
test_func_call_runtime[False-compile]	0.9993ms	0.7399ms	1.3515 KOps/s	1.2427 KOps/s	$\textbf{\color{#35bf28}+8.75\%}$
test_func_call_runtime[False-compile-overhead]	0.4137ms	0.3639ms	2.7481 KOps/s	2.6134 KOps/s	$\textbf{\color{#35bf28}+5.16\%}$
test_func_call_runtime[True-eager]	0.9577ms	0.8982ms	1.1134 KOps/s	1.0099 KOps/s	$\textbf{\color{#35bf28}+10.25\%}$
test_func_call_runtime[True-compile]	0.8223ms	0.7602ms	1.3154 KOps/s	1.3178 KOps/s	$\color{#d91a1a}-0.18\%$
test_func_call_runtime[True-compile-overhead]	0.4333ms	0.3861ms	2.5903 KOps/s	2.5933 KOps/s	$\color{#d91a1a}-0.12\%$
test_func_call_cm_runtime[False-eager]	0.7937ms	0.7291ms	1.3715 KOps/s	1.3189 KOps/s	$\color{#35bf28}+3.99\%$
test_func_call_cm_runtime[False-compile]	0.7917ms	0.7450ms	1.3422 KOps/s	1.3419 KOps/s	$\color{#35bf28}+0.03\%$
test_func_call_cm_runtime[False-compile-overhead]	0.4233ms	0.3664ms	2.7293 KOps/s	2.7303 KOps/s	$\color{#d91a1a}-0.04\%$
test_func_call_cm_runtime[True-eager]	1.1019ms	1.0022ms	997.8537 Ops/s	981.4942 Ops/s	$\color{#35bf28}+1.67\%$
test_func_call_cm_runtime[True-compile]	0.8486ms	0.7868ms	1.2710 KOps/s	1.2595 KOps/s	$\color{#35bf28}+0.91\%$
test_func_call_cm_runtime[True-compile-overhead]	0.4643ms	0.4096ms	2.4417 KOps/s	2.4079 KOps/s	$\color{#35bf28}+1.40\%$
test_vmap_func_call_cm_runtime[eager]	2.5653ms	2.0881ms	478.9064 Ops/s	471.5418 Ops/s	$\color{#35bf28}+1.56\%$
test_vmap_func_call_cm_runtime[compile]	0.9146ms	0.8147ms	1.2274 KOps/s	1.2337 KOps/s	$\color{#d91a1a}-0.51\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.4895ms	0.4137ms	2.4171 KOps/s	2.4154 KOps/s	$\color{#35bf28}+0.07\%$
test_distributed	3.2015ms	0.1859ms	5.3778 KOps/s	8.3218 KOps/s	$\textbf{\color{#d91a1a}-35.38\%}$
test_tdmodule	43.9300μs	19.2564μs	51.9307 KOps/s	47.7854 KOps/s	$\textbf{\color{#35bf28}+8.67\%}$
test_tdmodule_dispatch	58.5310μs	34.3470μs	29.1147 KOps/s	25.6213 KOps/s	$\textbf{\color{#35bf28}+13.63\%}$
test_tdseq	38.9710μs	20.2790μs	49.3121 KOps/s	45.5595 KOps/s	$\textbf{\color{#35bf28}+8.24\%}$
test_tdseq_dispatch	62.0210μs	37.4576μs	26.6969 KOps/s	24.6543 KOps/s	$\textbf{\color{#35bf28}+8.28\%}$
test_instantiation_functorch	1.6488ms	1.5751ms	634.8923 Ops/s	636.5040 Ops/s	$\color{#d91a1a}-0.25\%$
test_exec_functorch	0.2058ms	0.1481ms	6.7522 KOps/s	6.8860 KOps/s	$\color{#d91a1a}-1.94\%$
test_exec_functional_call	0.1855ms	0.1394ms	7.1745 KOps/s	7.1819 KOps/s	$\color{#d91a1a}-0.10\%$
test_exec_td_decorator	0.3850ms	0.1862ms	5.3712 KOps/s	5.3137 KOps/s	$\color{#35bf28}+1.08\%$
test_vmap_mlp_speed_decorator[True-True]	0.7547ms	0.6865ms	1.4567 KOps/s	1.3735 KOps/s	$\textbf{\color{#35bf28}+6.06\%}$
test_vmap_mlp_speed_decorator[True-False]	0.8248ms	0.6827ms	1.4647 KOps/s	1.3662 KOps/s	$\textbf{\color{#35bf28}+7.21\%}$
test_vmap_mlp_speed_decorator[False-True]	0.7093ms	0.5978ms	1.6729 KOps/s	1.5735 KOps/s	$\textbf{\color{#35bf28}+6.31\%}$
test_vmap_mlp_speed_decorator[False-False]	0.7109ms	0.5958ms	1.6785 KOps/s	1.5676 KOps/s	$\textbf{\color{#35bf28}+7.07\%}$
test_vmap_transformer_speed_decorator[True-True]	19.3633ms	19.1812ms	52.1343 Ops/s	51.4574 Ops/s	$\color{#35bf28}+1.32\%$
test_vmap_transformer_speed_decorator[True-False]	19.9269ms	19.2612ms	51.9180 Ops/s	51.2986 Ops/s	$\color{#35bf28}+1.21\%$
test_vmap_transformer_speed_decorator[False-True]	19.2654ms	19.1301ms	52.2736 Ops/s	51.4826 Ops/s	$\color{#35bf28}+1.54\%$
test_vmap_transformer_speed_decorator[False-False]	19.3817ms	19.0823ms	52.4045 Ops/s	51.7839 Ops/s	$\color{#35bf28}+1.20\%$
test_to_module_speed[True]	1.0657ms	0.9748ms	1.0258 KOps/s	1.0112 KOps/s	$\color{#35bf28}+1.45\%$
test_to_module_speed[False]	1.3397ms	0.9604ms	1.0412 KOps/s	1.0360 KOps/s	$\color{#35bf28}+0.50\%$
test_tc_init	61.0910μs	35.2047μs	28.4053 KOps/s	26.6634 KOps/s	$\textbf{\color{#35bf28}+6.53\%}$
test_tc_init_nested	0.1128ms	69.5846μs	14.3710 KOps/s	13.4726 KOps/s	$\textbf{\color{#35bf28}+6.67\%}$
test_tc_first_layer_tensor	20.7500μs	0.8281μs	1.2076 MOps/s	1.2239 MOps/s	$\color{#d91a1a}-1.33\%$
test_tc_first_layer_nontensor	27.8310μs	2.3162μs	431.7339 KOps/s	432.8706 KOps/s	$\color{#d91a1a}-0.26\%$
test_tc_second_layer_tensor	9.1653μs	1.4424μs	693.2900 KOps/s	682.9727 KOps/s	$\color{#35bf28}+1.51\%$
test_tc_second_layer_nontensor	25.9700μs	3.0829μs	324.3676 KOps/s	323.9778 KOps/s	$\color{#35bf28}+0.12\%$
test_unbind	0.2430s	10.3176ms	96.9216 Ops/s	141.7124 Ops/s	$\textbf{\color{#d91a1a}-31.61\%}$
test_full_like	12.1733ms	9.2502ms	108.1055 Ops/s	106.6248 Ops/s	$\color{#35bf28}+1.39\%$
test_zeros_like	6.0543ms	4.3367ms	230.5900 Ops/s	230.3083 Ops/s	$\color{#35bf28}+0.12\%$
test_ones_like	5.0743ms	4.4411ms	225.1701 Ops/s	225.5189 Ops/s	$\color{#d91a1a}-0.15\%$
test_clone	11.4594ms	9.2436ms	108.1829 Ops/s	153.6019 Ops/s	$\textbf{\color{#d91a1a}-29.57\%}$
test_squeeze	62.8310μs	9.8033μs	102.0069 KOps/s	103.6975 KOps/s	$\color{#d91a1a}-1.63\%$
test_unsqueeze	0.1233ms	74.1847μs	13.4799 KOps/s	13.4128 KOps/s	$\color{#35bf28}+0.50\%$
test_split	0.3816ms	0.1629ms	6.1382 KOps/s	5.9653 KOps/s	$\color{#35bf28}+2.90\%$
test_permute	0.2197ms	0.1759ms	5.6855 KOps/s	5.3312 KOps/s	$\textbf{\color{#35bf28}+6.65\%}$
test_stack	51.3354ms	51.0684ms	19.5816 Ops/s	19.5028 Ops/s	$\color{#35bf28}+0.40\%$
test_cat	52.8748ms	51.7185ms	19.3354 Ops/s	19.2463 Ops/s	$\color{#35bf28}+0.46\%$