[Feature] Module serialization

[vmoens]

No description provided.

[tcbegley]

LGTM

[github-actions]

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

Total Benchmarks: 47. Improved: $\large\color{#35bf28}1$. Worsened: $\large\color{#d91a1a}2$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_common_ops	1.2603ms	1.2215ms	818.6421 Ops/s	813.3087 Ops/s	$\color{#35bf28}+0.66\%$
test_creation	4.4181μs	4.1240μs	242.4854 KOps/s	236.0286 KOps/s	$\color{#35bf28}+2.74\%$
test_creation_empty	17.7773μs	16.5349μs	60.4783 KOps/s	61.3197 KOps/s	$\color{#d91a1a}-1.37\%$
test_creation_nested_1	29.7065μs	28.6878μs	34.8580 KOps/s	34.2405 KOps/s	$\color{#35bf28}+1.80\%$
test_creation_nested_2	30.5575μs	29.0596μs	34.4121 KOps/s	34.0938 KOps/s	$\color{#35bf28}+0.93\%$
test_clone	27.4525μs	26.3446μs	37.9584 KOps/s	37.9985 KOps/s	$\color{#d91a1a}-0.11\%$
test_getitem[int]	32.7147μs	32.1740μs	31.0810 KOps/s	31.2750 KOps/s	$\color{#d91a1a}-0.62\%$
test_getitem[slice_int]	74.5002μs	68.1020μs	14.6839 KOps/s	14.8964 KOps/s	$\color{#d91a1a}-1.43\%$
test_getitem[range]	70.5634μs	69.6000μs	14.3678 KOps/s	14.5754 KOps/s	$\color{#d91a1a}-1.42\%$
test_getitem[tuple]	63.2736μs	62.7013μs	15.9486 KOps/s	16.1464 KOps/s	$\color{#d91a1a}-1.22\%$
test_getitem[list]	62.1988μs	61.6060μs	16.2322 KOps/s	16.4203 KOps/s	$\color{#d91a1a}-1.15\%$
test_setitem_dim[int]	80.7010μs	45.6294μs	21.9157 KOps/s	22.6861 KOps/s	$\color{#d91a1a}-3.40\%$
test_setitem_dim[slice_int]	0.1858ms	81.8866μs	12.2120 KOps/s	12.3437 KOps/s	$\color{#d91a1a}-1.07\%$
test_setitem_dim[range]	0.1558ms	77.6173μs	12.8837 KOps/s	13.1778 KOps/s	$\color{#d91a1a}-2.23\%$
test_setitem_dim[tuple]	0.1657ms	74.9000μs	13.3511 KOps/s	13.6977 KOps/s	$\color{#d91a1a}-2.53\%$
test_setitem	39.6437μs	38.5731μs	25.9248 KOps/s	26.1786 KOps/s	$\color{#d91a1a}-0.97\%$
test_set	43.5957μs	38.2911μs	26.1157 KOps/s	26.9179 KOps/s	$\color{#d91a1a}-2.98\%$
test_set_shared	0.1781ms	0.1739ms	5.7521 KOps/s	5.6799 KOps/s	$\color{#35bf28}+1.27\%$
test_update	48.2738μs	47.2746μs	21.1530 KOps/s	21.2266 KOps/s	$\color{#d91a1a}-0.35\%$
test_update_nested	69.0952μs	68.0349μs	14.6983 KOps/s	14.8922 KOps/s	$\color{#d91a1a}-1.30\%$
test_set_nested	48.0908μs	46.9877μs	21.2822 KOps/s	21.3407 KOps/s	$\color{#d91a1a}-0.27\%$
test_set_nested_new	67.1882μs	65.3337μs	15.3060 KOps/s	15.3615 KOps/s	$\color{#d91a1a}-0.36\%$
test_select	0.1047ms	0.1031ms	9.6968 KOps/s	9.7290 KOps/s	$\color{#d91a1a}-0.33\%$
test_creation[device0]	1.8891ms	0.5613ms	1.7815 KOps/s	1.8094 KOps/s	$\color{#d91a1a}-1.54\%$
test_creation_from_tensor	0.5709ms	0.4615ms	2.1670 KOps/s	2.1543 KOps/s	$\color{#35bf28}+0.59\%$
test_add_one[memmap_tensor0]	33.2036μs	30.0215μs	33.3095 KOps/s	33.0157 KOps/s	$\color{#35bf28}+0.89\%$
test_contiguous[memmap_tensor0]	8.4591μs	7.9682μs	125.4986 KOps/s	126.3449 KOps/s	$\color{#d91a1a}-0.67\%$
test_stack[memmap_tensor0]	0.1831ms	42.9207μs	23.2988 KOps/s	22.8255 KOps/s	$\color{#35bf28}+2.07\%$
test_reshape_pytree	38.5577μs	35.7967μs	27.9356 KOps/s	27.9865 KOps/s	$\color{#d91a1a}-0.18\%$
test_reshape_td	50.6929μs	48.2861μs	20.7099 KOps/s	20.5495 KOps/s	$\color{#35bf28}+0.78\%$
test_view_pytree	33.8366μs	32.7132μs	30.5687 KOps/s	29.8210 KOps/s	$\color{#35bf28}+2.51\%$
test_view_td	9.9452μs	8.9233μs	112.0656 KOps/s	112.1258 KOps/s	$\color{#d91a1a}-0.05\%$
test_unbind_pytree	37.5386μs	36.0784μs	27.7174 KOps/s	27.2725 KOps/s	$\color{#35bf28}+1.63\%$
test_unbind_td	0.1926ms	0.1895ms	5.2757 KOps/s	5.3567 KOps/s	$\color{#d91a1a}-1.51\%$
test_split_pytree	43.8508μs	41.8971μs	23.8680 KOps/s	24.0794 KOps/s	$\color{#d91a1a}-0.88\%$
test_split_td	0.1331ms	0.1161ms	8.6145 KOps/s	8.7571 KOps/s	$\color{#d91a1a}-1.63\%$
test_add_pytree	46.7198μs	44.9803μs	22.2320 KOps/s	22.0633 KOps/s	$\color{#35bf28}+0.76\%$
test_add_td	79.7974μs	77.7037μs	12.8694 KOps/s	13.1863 KOps/s	$\color{#d91a1a}-2.40\%$
test_distributed	73.6010μs	73.6010μs	13.5868 KOps/s	11.6549 KOps/s	$\textbf{\color{#35bf28}+16.58\%}$
test_tdmodule	71.0010μs	28.2020μs	35.4585 KOps/s	35.2193 KOps/s	$\color{#35bf28}+0.68\%$
test_tdmodule_dispatch	62.6280ms	66.9110μs	14.9452 KOps/s	16.2392 KOps/s	$\textbf{\color{#d91a1a}-7.97\%}$
test_tdseq	0.5931ms	39.6816μs	25.2006 KOps/s	25.7956 KOps/s	$\color{#d91a1a}-2.31\%$
test_tdseq_dispatch	0.1265ms	70.4221μs	14.2001 KOps/s	14.0889 KOps/s	$\color{#35bf28}+0.79\%$
test_instantiation_functorch	1.6438ms	1.5672ms	638.0701 Ops/s	644.5778 Ops/s	$\color{#d91a1a}-1.01\%$
test_instantiation_td	8.2483ms	1.2727ms	785.7163 Ops/s	828.2322 Ops/s	$\textbf{\color{#d91a1a}-5.13\%}$
test_exec_functorch	0.1842ms	0.1796ms	5.5667 KOps/s	5.4787 KOps/s	$\color{#35bf28}+1.61\%$
test_exec_td	0.3322ms	0.3274ms	3.0539 KOps/s	3.0480 KOps/s	$\color{#35bf28}+0.20\%$

[github-actions]

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

Total Benchmarks: 47. Improved: $\large\color{#35bf28}5$. Worsened: $\large\color{#d91a1a}23$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_common_ops	1.9173ms	1.6644ms	600.8341 Ops/s	603.2992 Ops/s	$\color{#d91a1a}-0.41\%$
test_creation	6.3230μs	4.2827μs	233.4959 KOps/s	262.6970 KOps/s	$\textbf{\color{#d91a1a}-11.12\%}$
test_creation_empty	27.6541μs	18.3744μs	54.4236 KOps/s	59.9746 KOps/s	$\textbf{\color{#d91a1a}-9.26\%}$
test_creation_nested_1	41.6882μs	32.2706μs	30.9879 KOps/s	32.7845 KOps/s	$\textbf{\color{#d91a1a}-5.48\%}$
test_creation_nested_2	57.4723μs	33.1891μs	30.1304 KOps/s	31.7060 KOps/s	$\color{#d91a1a}-4.97\%$
test_clone	0.1018ms	32.7212μs	30.5612 KOps/s	33.3143 KOps/s	$\textbf{\color{#d91a1a}-8.26\%}$
test_getitem[int]	59.6881μs	39.7961μs	25.1281 KOps/s	27.1144 KOps/s	$\textbf{\color{#d91a1a}-7.33\%}$
test_getitem[slice_int]	0.1063ms	84.6032μs	11.8199 KOps/s	12.9471 KOps/s	$\textbf{\color{#d91a1a}-8.71\%}$
test_getitem[range]	0.1382ms	0.1036ms	9.6540 KOps/s	9.7755 KOps/s	$\color{#d91a1a}-1.24\%$
test_getitem[tuple]	0.1031ms	72.9349μs	13.7109 KOps/s	14.5211 KOps/s	$\textbf{\color{#d91a1a}-5.58\%}$
test_getitem[list]	0.1361ms	93.5898μs	10.6849 KOps/s	11.2554 KOps/s	$\textbf{\color{#d91a1a}-5.07\%}$
test_setitem_dim[int]	0.8141ms	64.5644μs	15.4884 KOps/s	15.0905 KOps/s	$\color{#35bf28}+2.64\%$
test_setitem_dim[slice_int]	1.7302ms	0.1179ms	8.4808 KOps/s	8.8396 KOps/s	$\color{#d91a1a}-4.06\%$
test_setitem_dim[range]	5.5161ms	0.1280ms	7.8142 KOps/s	7.8211 KOps/s	$\color{#d91a1a}-0.09\%$
test_setitem_dim[tuple]	2.1152ms	0.1010ms	9.8969 KOps/s	10.2669 KOps/s	$\color{#d91a1a}-3.60\%$
test_setitem	69.6263μs	46.6061μs	21.4564 KOps/s	17.2886 KOps/s	$\textbf{\color{#35bf28}+24.11\%}$
test_set	55.6012μs	46.7110μs	21.4082 KOps/s	17.5640 KOps/s	$\textbf{\color{#35bf28}+21.89\%}$
test_set_shared	0.3534ms	0.3028ms	3.3022 KOps/s	3.5001 KOps/s	$\textbf{\color{#d91a1a}-5.65\%}$
test_update	0.1065ms	56.6897μs	17.6399 KOps/s	18.5396 KOps/s	$\color{#d91a1a}-4.85\%$
test_update_nested	89.3484μs	77.5480μs	12.8952 KOps/s	13.1294 KOps/s	$\color{#d91a1a}-1.78\%$
test_set_nested	0.1895ms	62.9980μs	15.8735 KOps/s	17.8049 KOps/s	$\textbf{\color{#d91a1a}-10.85\%}$
test_set_nested_new	0.1424ms	85.1343μs	11.7462 KOps/s	13.4448 KOps/s	$\textbf{\color{#d91a1a}-12.63\%}$
test_select	0.1845ms	0.1317ms	7.5921 KOps/s	8.4194 KOps/s	$\textbf{\color{#d91a1a}-9.83\%}$
test_creation[device0]	2.1654ms	0.7214ms	1.3862 KOps/s	1.4945 KOps/s	$\textbf{\color{#d91a1a}-7.25\%}$
test_creation_from_tensor	0.7175ms	0.6073ms	1.6467 KOps/s	1.5263 KOps/s	$\textbf{\color{#35bf28}+7.89\%}$
test_add_one[memmap_tensor0]	0.1511ms	75.3254μs	13.2757 KOps/s	17.9368 KOps/s	$\textbf{\color{#d91a1a}-25.99\%}$
test_contiguous[memmap_tensor0]	25.4591μs	17.2545μs	57.9558 KOps/s	93.6142 KOps/s	$\textbf{\color{#d91a1a}-38.09\%}$
test_stack[memmap_tensor0]	0.2708ms	90.6403μs	11.0326 KOps/s	12.9500 KOps/s	$\textbf{\color{#d91a1a}-14.81\%}$
test_reshape_pytree	72.6583μs	48.6210μs	20.5673 KOps/s	29.0320 KOps/s	$\textbf{\color{#d91a1a}-29.16\%}$
test_reshape_td	81.9464μs	68.0021μs	14.7054 KOps/s	19.5041 KOps/s	$\textbf{\color{#d91a1a}-24.60\%}$
test_view_pytree	48.9712μs	41.2632μs	24.2347 KOps/s	31.4360 KOps/s	$\textbf{\color{#d91a1a}-22.91\%}$
test_view_td	13.4621μs	9.0195μs	110.8710 KOps/s	116.5495 KOps/s	$\color{#d91a1a}-4.87\%$
test_unbind_pytree	58.1233μs	47.7815μs	20.9286 KOps/s	26.3709 KOps/s	$\textbf{\color{#d91a1a}-20.64\%}$
test_unbind_td	0.3253ms	0.2505ms	3.9913 KOps/s	4.6843 KOps/s	$\textbf{\color{#d91a1a}-14.79\%}$
test_split_pytree	66.9933μs	45.8458μs	21.8123 KOps/s	22.9077 KOps/s	$\color{#d91a1a}-4.78\%$
test_split_td	0.2267ms	0.1533ms	6.5247 KOps/s	7.5998 KOps/s	$\textbf{\color{#d91a1a}-14.15\%}$
test_add_pytree	0.1643ms	58.2894μs	17.1558 KOps/s	17.7816 KOps/s	$\color{#d91a1a}-3.52\%$
test_add_td	0.1619ms	0.1137ms	8.7942 KOps/s	8.8182 KOps/s	$\color{#d91a1a}-0.27\%$
test_distributed	0.1449ms	0.1449ms	6.9013 KOps/s	6.4850 KOps/s	$\textbf{\color{#35bf28}+6.42\%}$
test_tdmodule	0.4578ms	38.1685μs	26.1996 KOps/s	25.8127 KOps/s	$\color{#35bf28}+1.50\%$
test_tdmodule_dispatch	9.1121ms	85.3388μs	11.7180 KOps/s	11.7699 KOps/s	$\color{#d91a1a}-0.44\%$
test_tdseq	1.1154ms	49.2962μs	20.2856 KOps/s	20.0009 KOps/s	$\color{#35bf28}+1.42\%$
test_tdseq_dispatch	2.4262ms	99.8428μs	10.0157 KOps/s	9.8725 KOps/s	$\color{#35bf28}+1.45\%$
test_instantiation_functorch	2.3683ms	1.8077ms	553.1868 Ops/s	522.5698 Ops/s	$\textbf{\color{#35bf28}+5.86\%}$
test_instantiation_td	1.8037ms	1.3746ms	727.4675 Ops/s	722.8812 Ops/s	$\color{#35bf28}+0.63\%$
test_exec_functorch	0.3869ms	0.2664ms	3.7533 KOps/s	3.9638 KOps/s	$\textbf{\color{#d91a1a}-5.31\%}$
test_exec_td	0.5607ms	0.4615ms	2.1669 KOps/s	2.1410 KOps/s	$\color{#35bf28}+1.21\%$