diff --git "a/perf-df-bnb-1xT4.csv" "b/perf-df-bnb-1xT4.csv" --- "a/perf-df-bnb-1xT4.csv" +++ "b/perf-df-bnb-1xT4.csv" @@ -2980,7 +2980,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 162.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 58.12 MiB is free. Process 128567 has 14.68 GiB memory in use. Of the allocated memory 14.19 GiB is allocated by PyTorch, and 384.43 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 162.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 58.12 MiB is free. Process 134274 has 14.68 GiB memory in use. Of the allocated memory 14.19 GiB is allocated by PyTorch, and 384.43 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,stabilityai/stablelm-base-alpha-7b,stabilityai/stablelm-base-alpha-7b,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,,True,MB,1556.680704,6043.860992,0.0,5641.33888,5589.443072,s,1,15.3459052734375,15.3459052734375,0.0,15.3459052734375,15.3459052734375,15.3459052734375,15.3459052734375,[15.3459052734375],,kWh,0.00022218646855417317,2.450167138528771e-05,7.36295033479939e-05,0.00032031764328745475,,MB,1422.323712,7990.018048,0.0,7574.913024,6755.79136,s,10,9.935990173339844,0.9935990173339844,0.007965376884838362,0.9932402648925781,1.0041536987304687,1.0051293395996095,1.005909852294922,"[0.9806353149414062, 0.9870344848632813, 0.99700732421875, 0.9875632934570312, 0.9940713500976562, 1.0010421752929688, 0.9924091796875, 0.9861851806640625, 1.00610498046875, 1.0039368896484375]",tokens/s,257.64920811505715,kWh,2.8913279401896224e-05,3.1886690359871966e-06,1.9253500251271603e-05,5.135544868915503e-05,tokens/kWh,4984865.414174849,MB,1478.88128,7992.1152,0.0,7574.913024,6755.79392,s,10,50.50752978515625,5.050752978515625,0.016425835324374135,5.054232666015626,5.066164404296875,5.072623706054688,5.077791147460937,"[5.023517578125, 5.02657177734375, 5.0383994140625, 5.04711279296875, 5.0790830078125, 5.05138720703125, 5.057078125, 5.05864794921875, 5.0610029296875, 5.06472900390625]",tokens/s,12.473387684565635,kWh,0.00014834524946393877,1.636272579115718e-05,9.838745497252787e-05,0.0002630954302276238,tokens/kWh,239456.83870485294,,s,630,50.505484954833975,0.08016743643624442,0.0013153008877054171,0.08000113677978515,0.08106471481323242,0.08131376800537109,0.08881760345458985,"[0.09101910400390625, 0.08023741149902344, 0.07947593688964844, 0.07902457427978515, 0.07861885070800781, 0.0785567398071289, 0.07875199890136719, 0.07874380493164063, 0.07873036956787109, 0.07870553588867188, 0.07869222259521484, 0.07985273742675782, 0.08150931549072266, 0.08055971527099609, 0.08010793304443359, 0.08037366485595702, 0.07970121765136719, 0.07920524597167969, 0.07880908966064454, 0.07873900604248046, 0.07859394836425782, 0.07856492614746094, 0.0786884765625, 0.07973283386230469, 0.0810272674560547, 0.08046230316162109, 0.07994163513183594, 0.07926112365722657, 0.07932371520996094, 0.08060313415527344, 0.07986176300048828, 0.07942937469482422, 0.0789793243408203, 0.07874969482421874, 0.07869644927978516, 0.0787783660888672, 0.08027545928955078, 0.07985971069335937, 0.08106304168701171, 0.0807033920288086, 0.07982825469970703, 0.07935507202148437, 0.07927426910400391, 0.0797001953125, 0.07924972534179688, 0.07981439971923829, 0.07940502166748047, 0.07899772644042968, 0.07872201538085938, 0.07954867553710937, 0.07984703826904296, 0.07943580627441406, 0.07908601379394531, 0.07997657775878907, 0.07976080322265625, 0.07987088012695312, 0.07946607971191406, 0.07961641693115235, 0.08015068817138672, 0.07955244445800781, 0.08057231903076172, 0.08040866851806641, 0.07964399719238281, 0.08838150024414063, 0.080046142578125, 0.07936061096191406, 0.07977935791015625, 0.07931887817382813, 0.0790552978515625, 0.0790169906616211, 0.07868515014648438, 0.07842406463623047, 0.07847280120849609, 0.07849820709228515, 0.08004812622070312, 0.08060518646240235, 0.08051113891601562, 0.08035497283935547, 0.07981430053710938, 0.07917350769042969, 0.07890758514404297, 0.07995404815673827, 0.07944815826416016, 0.07895836639404297, 0.07871250915527343, 0.0786849594116211, 0.0799435806274414, 0.07944316864013672, 0.08019197082519532, 0.08005014038085938, 0.07976182556152343, 0.07922838592529297, 0.07927247619628906, 0.0799170913696289, 0.07940620422363281, 0.07906547546386719, 0.08013651275634766, 0.07941545867919922, 0.07887865447998046, 0.0790098876953125, 0.08005788421630859, 0.08003836822509766, 0.08012595367431641, 0.07944105529785156, 0.07943254089355468, 0.07969296264648437, 0.08006681823730469, 0.07938108825683594, 0.08071987152099609, 0.07997964477539063, 0.08062655639648438, 0.08019558715820313, 0.08058470153808593, 0.07998854064941406, 0.08036982727050782, 0.08000035095214844, 0.07943772888183594, 0.07928707122802735, 0.07933926391601563, 0.08014057922363281, 0.07986172485351563, 0.08045967864990235, 0.08000867462158204, 0.07951615905761719, 0.08017353820800781, 0.0795255355834961, 0.08819391632080079, 0.07993520355224609, 0.07941120147705077, 0.08018972778320313, 0.07993344116210938, 0.07935590362548828, 0.07932326507568359, 0.07904374694824219, 0.07907571411132812, 0.07904905700683594, 0.07983513641357422, 0.07975039672851562, 0.0803683853149414, 0.08028569793701172, 0.07991276550292968, 0.07981394958496094, 0.07949024200439453, 0.0791192626953125, 0.07994035339355468, 0.07917571258544921, 0.07914828491210937, 0.07960681915283203, 0.07929417419433593, 0.07983055877685546, 0.08056060791015625, 0.08081932830810547, 0.08035008239746094, 0.08073958587646485, 0.07998258972167968, 0.08046982574462891, 0.08048694610595702, 0.07954863739013672, 0.07930691528320312, 0.07928412628173828, 0.07928604888916016, 0.07964006042480469, 0.07964086151123047, 0.08004627227783204, 0.07990473937988281, 0.08027792358398438, 0.0799295654296875, 0.07970793914794921, 0.08023264312744141, 0.07956256103515626, 0.07954841613769531, 0.07951737976074219, 0.07946886444091797, 0.07955577850341797, 0.08020771026611329, 0.08088412475585938, 0.08025350189208984, 0.0796918716430664, 0.08016886138916016, 0.07964214324951172, 0.07967359924316406, 0.08011321258544922, 0.07938114929199219, 0.07985151672363282, 0.07980646514892578, 0.07941558074951172, 0.08015843200683594, 0.0806541748046875, 0.08135298919677734, 0.08861689758300781, 0.07989043426513671, 0.0792125473022461, 0.07961929321289063, 0.07988473510742188, 0.07950780487060546, 0.07910399627685546, 0.07906018829345703, 0.07985049438476563, 0.07922665405273438, 0.07917369842529297, 0.08001747131347656, 0.08047808074951172, 0.0812415008544922, 0.08045401763916016, 0.07979043579101562, 0.07934349060058593, 0.07922688293457031, 0.08020172882080077, 0.07956479644775391, 0.07941529846191406, 0.07920873260498047, 0.07935743713378907, 0.07992546844482422, 0.08011286163330078, 0.08121833801269532, 0.0805212173461914, 0.08036479949951172, 0.07994432067871093, 0.079334716796875, 0.08001747131347656, 0.08003852844238281, 0.07990300750732422, 0.07944905853271485, 0.07927859497070312, 0.07984780883789062, 0.08120294189453126, 0.08119484710693359, 0.08060723114013672, 0.08055248260498046, 0.07970012664794922, 0.07994115447998047, 0.08039647674560547, 0.08076697540283204, 0.08024269104003906, 0.07971593475341797, 0.08113967895507812, 0.08017113494873047, 0.07970211029052734, 0.08032892608642578, 0.08010956573486328, 0.07989453125, 0.08005996704101563, 0.07945670318603516, 0.07954637145996094, 0.07986930847167968, 0.08056508636474609, 0.08018883514404297, 0.0798194580078125, 0.08016476440429687, 0.07961373138427734, 0.08006982421875, 0.08049295806884765, 0.08980643463134766, 0.08163334655761718, 0.08063011169433594, 0.07916671752929688, 0.07942623901367188, 0.07946054077148437, 0.07949046325683594, 0.08123359680175782, 0.08172828674316407, 0.0821180191040039, 0.08163401794433593, 0.08069929504394531, 0.08173782348632813, 0.081082275390625, 0.0805090560913086, 0.08000192260742188, 0.08018351745605469, 0.07955465698242188, 0.0794525146484375, 0.07952623748779297, 0.08007881927490235, 0.07952413177490235, 0.0793268814086914, 0.08139170837402344, 0.08099366760253907, 0.081478271484375, 0.08075878143310547, 0.0799719009399414, 0.07936428833007812, 0.08047853088378906, 0.07991216278076171, 0.07946720123291015, 0.0798578872680664, 0.07941458892822266, 0.0800343017578125, 0.08058060455322266, 0.08145836639404297, 0.08202137756347656, 0.08131529235839843, 0.08111702728271485, 0.08182220458984375, 0.08098636627197266, 0.08082556915283202, 0.08108902740478516, 0.08005840301513673, 0.07953135681152344, 0.0807225570678711, 0.0801157455444336, 0.08117887878417969, 0.08111078643798827, 0.08129254150390625, 0.08105651092529297, 0.08080563354492187, 0.08039814758300781, 0.07946284484863281, 0.07953794860839844, 0.0800770263671875, 0.07957263946533204, 0.080056640625, 0.08035228729248046, 0.08024371337890625, 0.08042217254638671, 0.08052764892578125, 0.08914761352539062, 0.07993309020996094, 0.0793081283569336, 0.07904166412353515, 0.08020313262939453, 0.0798048324584961, 0.08090624237060547, 0.07981219482421875, 0.080552001953125, 0.07995423889160157, 0.07924098968505859, 0.08012210845947265, 0.08149587249755859, 0.08012009429931641, 0.07955446624755859, 0.07960108947753906, 0.07981705474853516, 0.07919586944580079, 0.07904009246826171, 0.07977871704101562, 0.07930614471435547, 0.07989513397216796, 0.07988755035400391, 0.08002642822265625, 0.08063571166992188, 0.08003398132324219, 0.0800186538696289, 0.07968233489990234, 0.07988428497314454, 0.07984742736816407, 0.0795832290649414, 0.07983939361572266, 0.07945814514160156, 0.07961804962158203, 0.08077043151855469, 0.07987468719482421, 0.07999282836914062, 0.07969324493408203, 0.0801819839477539, 0.08068899536132812, 0.07994342041015624, 0.08070937347412109, 0.080089599609375, 0.08057852935791016, 0.08002159881591797, 0.08061299133300781, 0.08009113311767578, 0.08076643371582032, 0.08004898834228516, 0.0799596176147461, 0.07983139038085937, 0.0801956787109375, 0.08031427001953124, 0.08045577239990234, 0.0802870101928711, 0.07985635375976563, 0.07979622650146484, 0.07998607635498046, 0.08008354949951171, 0.08058841705322266, 0.08119744110107421, 0.08018688201904296, 0.08005228424072265, 0.08889984130859375, 0.08007510375976562, 0.0791756820678711, 0.07892697906494141, 0.07963533020019531, 0.07987773132324219, 0.08021443176269531, 0.07990402984619141, 0.0792501449584961, 0.07921449279785156, 0.07922803497314453, 0.08078060913085937, 0.08116329956054688, 0.08062966156005859, 0.08071833801269532, 0.08000259399414063, 0.07985049438476563, 0.08006201934814453, 0.07959142303466797, 0.07991120147705078, 0.07927792358398437, 0.07936109161376953, 0.08002861022949219, 0.08016441345214843, 0.08124041748046874, 0.08090838623046875, 0.0811924819946289, 0.08025526428222657, 0.07978617858886719, 0.07928012847900391, 0.08019286346435547, 0.07981852722167969, 0.08007154846191407, 0.07979859161376954, 0.08005804443359375, 0.07980032348632812, 0.08055193328857421, 0.08065229034423828, 0.0804167709350586, 0.0800351333618164, 0.07941395568847656, 0.08022425842285157, 0.0798897933959961, 0.07940937805175781, 0.08122927856445313, 0.08075564575195313, 0.08000688171386719, 0.07943612670898438, 0.0811716766357422, 0.08049465942382812, 0.08080655670166016, 0.07999282836914062, 0.07964281463623046, 0.08008048248291015, 0.08057244873046875, 0.08007084655761719, 0.08023155212402344, 0.07986675262451172, 0.08112127685546874, 0.0808652801513672, 0.08010348510742188, 0.08084591674804688, 0.08062448120117187, 0.0890814437866211, 0.08012809753417968, 0.07933132934570312, 0.07917616271972656, 0.07975421142578125, 0.07926233673095703, 0.08042729949951172, 0.07996415710449219, 0.07964851379394532, 0.07934182739257813, 0.0790118408203125, 0.08203257751464844, 0.08133773040771484, 0.08031318664550781, 0.07990460968017578, 0.07970636749267578, 0.07986764526367188, 0.0792405776977539, 0.07963478088378906, 0.07934390258789062, 0.07921891021728515, 0.079351806640625, 0.07982498931884766, 0.08060486602783203, 0.08108646392822266, 0.08038323211669922, 0.08038188934326172, 0.07997702026367187, 0.08034732818603516, 0.07980857849121094, 0.07925350189208985, 0.07919641876220704, 0.08023190307617188, 0.07953846740722656, 0.07936204528808594, 0.08023375701904296, 0.08040521240234375, 0.08111309051513672, 0.08066252899169922, 0.07984537506103516, 0.08116838073730469, 0.0804947509765625, 0.08005001831054688, 0.0794051513671875, 0.07956185913085938, 0.07991990661621094, 0.08092476654052734, 0.08068643188476562, 0.08021590423583984, 0.08131190490722656, 0.08104812622070312, 0.0812298583984375, 0.080942626953125, 0.08001692962646484, 0.07971552276611328, 0.08004374694824219, 0.08082022094726563, 0.08019967651367188, 0.08016496276855468, 0.08042691040039063, 0.08086268615722657, 0.08152937316894532, 0.08034508514404297, 0.08889958190917968, 0.07996415710449219, 0.079325439453125, 0.07975910186767578, 0.07932444763183594, 0.08064688110351563, 0.07991295623779297, 0.08022160339355469, 0.07980912017822266, 0.08009228515625, 0.07950527954101562, 0.0801087646484375, 0.08105667114257813, 0.08081254577636719, 0.07980278778076172, 0.0794562225341797, 0.07935590362548828, 0.07996211242675781, 0.07999078369140625, 0.0798392333984375, 0.07975730895996094, 0.07948441314697266, 0.07937677001953125, 0.08144908905029297, 0.08074588775634765, 0.08117699432373048, 0.08014768218994141, 0.07983622741699219, 0.08010892486572266, 0.07982339477539062, 0.08040013122558594, 0.08022386932373046, 0.08027519989013672, 0.07994457244873047, 0.08038358306884766, 0.08057488250732422, 0.08008704376220703, 0.08108035278320312, 0.08101267242431641, 0.0800747528076172, 0.07958326721191407, 0.07941069030761719, 0.07994624328613281, 0.0803552017211914, 0.08002162933349609, 0.07982489776611328, 0.0802529296875, 0.07990681457519531, 0.0808051528930664, 0.08046428680419922, 0.08069344329833984, 0.08026329803466797, 0.07962214660644532, 0.08069734191894531, 0.08028921508789062, 0.07963897705078125, 0.08048652648925782, 0.08113970947265625, 0.08057651519775391, 0.08102706909179687, 0.08125440216064453, 0.08059635162353515, 0.08013011169433594, 0.08898976135253907, 0.0801115493774414, 0.0792795867919922, 0.07918572998046874, 0.07985814666748046, 0.08050918579101562, 0.08068736267089843, 0.07973862457275391, 0.08063180541992188, 0.08046198272705078, 0.07971004486083984, 0.08039119720458984, 0.08081833648681641, 0.0811374740600586, 0.08065945434570312, 0.07997030639648438, 0.07985539245605469, 0.07987152099609375, 0.07941600036621094, 0.07993276977539063, 0.07979689788818359, 0.08006451416015625, 0.08033014678955078, 0.08133193969726563, 0.08159295654296875, 0.08021836853027343, 0.0801495361328125, 0.079891357421875, 0.07938780975341797, 0.07993030548095703, 0.08033891296386719, 0.08016281890869141, 0.07925116729736328, 0.07979574584960937, 0.08030694580078125, 0.08029593658447266, 0.08065401458740235, 0.081244384765625, 0.08107977294921875, 0.08050956726074218, 0.08000621032714844, 0.08013670349121094, 0.08006291198730468, 0.08071372985839843, 0.07994931030273437, 0.0802636489868164, 0.07986300659179688, 0.08022713470458985, 0.08020716857910157, 0.08040278625488281, 0.08167868804931641, 0.08091420745849609, 0.07976700592041015, 0.0796272964477539, 0.08010089874267579, 0.08035939025878906, 0.08038626861572265, 0.08003584289550782, 0.08021401977539062, 0.08001519775390625, 0.08003936004638672, 0.080401123046875, 0.08161917114257812]",tokens/s,12.473892698256357,, @@ -3075,7 +3075,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 68.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 68.12 MiB is free. Process 77332 has 14.67 GiB memory in use. Of the allocated memory 14.27 GiB is allocated by PyTorch, and 293.36 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 68.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 68.12 MiB is free. Process 66199 has 14.67 GiB memory in use. Of the allocated memory 14.27 GiB is allocated by PyTorch, and 293.36 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neo,EleutherAI/gpt-neo-125m,EleutherAI/gpt-neo-125m,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,897.257472,577.69984,0.0,182.452224,179.733504,s,1,7.877572265625,7.877572265625,0.0,7.877572265625,7.877572265625,7.877572265625,7.877572265625,[7.877572265625],,kWh,2.558793690833454e-05,2.8154384288212715e-06,9.265562967999963e-06,3.7668938305155774e-05,,MB,1205.055488,674.168832,0.0,266.338304,224.293376,s,10,0.26332313537597657,0.026332313537597657,0.000610555558624612,0.026153087615966797,0.0267448860168457,0.027288587379455564,0.027723548469543457,"[0.026606048583984375, 0.02783228874206543, 0.026089855194091797, 0.02617750358581543, 0.026577024459838866, 0.02662406349182129, 0.025579135894775392, 0.026088127136230467, 0.026128671646118165, 0.025620416641235353]",tokens/s,9721.895481552714,kWh,7.676794833879956e-07,8.462414545547406e-08,5.081871957008012e-07,1.3604908245442707e-06,tokens/kWh,188167384.43330067,MB,1238.601728,686.751744,0.0,278.921216,224.295936,s,10,11.490330200195315,1.1490330200195316,0.01679524657240238,1.1495841064453125,1.16569306640625,1.1701708740234376,1.1737531201171876,"[1.164697998046875, 1.1567578125, 1.174648681640625, 1.14178076171875, 1.16360595703125, 1.13744384765625, 1.12833935546875, 1.142410400390625, 1.119871826171875, 1.1607735595703126]",tokens/s,54.82871153600887,kWh,3.332297936244537e-05,3.6750928415303726e-06,1.3648851907698941e-05,5.064692411167469e-05,tokens/kWh,1243905.747584734,,s,630,11.485126411438001,0.018230359383234904,0.0005112548961969608,0.018171520233154298,0.01880431652069092,0.018962901496887206,0.019943947334289565,"[0.018439231872558595, 0.01866339111328125, 0.018562015533447267, 0.018786624908447267, 0.01859756851196289, 0.018696384429931642, 0.018680704116821288, 0.018768447875976563, 0.01888889694213867, 0.018825408935546874, 0.01873481559753418, 0.018884767532348634, 0.018837568283081054, 0.018847232818603517, 0.01878041648864746, 0.01879449653625488, 0.01886240005493164, 0.01865465545654297, 0.018614784240722656, 0.01889244842529297, 0.018651552200317383, 0.018642080307006835, 0.018993600845336914, 0.019159839630126952, 0.018760608673095702, 0.01876483154296875, 0.018767551422119142, 0.018751487731933594, 0.01871254348754883, 0.018669120788574217, 0.018997728347778322, 0.01865727996826172, 0.01872307205200195, 0.018747135162353514, 0.018733119964599608, 0.018628864288330077, 0.018670400619506835, 0.018710527420043945, 0.018918272018432616, 0.018571231842041017, 0.018412832260131837, 0.018470848083496094, 0.01838371276855469, 0.018323423385620118, 0.018042560577392577, 0.017882944107055664, 0.01795123291015625, 0.017831808090209962, 0.01785625648498535, 0.017836416244506836, 0.01785206413269043, 0.01781491279602051, 0.01791689682006836, 0.018007167816162108, 0.017929056167602538, 0.018064863204956056, 0.01795129585266113, 0.017974880218505858, 0.017905248641967773, 0.018010751724243164, 0.017913728713989257, 0.017920320510864257, 0.01792854309082031, 0.018087072372436522, 0.018312000274658204, 0.01863680076599121, 0.01884547233581543, 0.01850332832336426, 0.018425567626953125, 0.018295648574829102, 0.0181711368560791, 0.018076416015625, 0.018091552734375, 0.018323392868041993, 0.017973888397216798, 0.019459999084472657, 0.018114303588867186, 0.018006336212158202, 0.018280384063720703, 0.01801366424560547, 0.017959583282470704, 0.0182291202545166, 0.01833087921142578, 0.01811942481994629, 0.01926380729675293, 0.018413248062133788, 0.018160991668701172, 0.01817888069152832, 0.018222944259643555, 0.018433664321899416, 0.018229631423950194, 0.019092832565307617, 0.018766464233398436, 0.018614303588867186, 0.018473024368286132, 0.01861030387878418, 0.01933401679992676, 0.019438623428344726, 0.01968230438232422, 0.018762304306030275, 0.01846681594848633, 0.01818454360961914, 0.018422815322875978, 0.01789411163330078, 0.01792639923095703, 0.018190336227416993, 0.018378751754760742, 0.018219200134277344, 0.01799884796142578, 0.017906496047973633, 0.017922048568725587, 0.017897087097167967, 0.017891359329223634, 0.01800831985473633, 0.018032127380371094, 0.018299615859985352, 0.018300800323486327, 0.018249727249145507, 0.018130048751831055, 0.018365312576293945, 0.018343584060668945, 0.018393407821655272, 0.018468544006347655, 0.018223424911499024, 0.018124832153320312, 0.0180982723236084, 0.018126144409179687, 0.018141824722290038, 0.01901568031311035, 0.018462400436401367, 0.018114879608154298, 0.01816166305541992, 0.01819379234313965, 0.01808563232421875, 0.018104352951049806, 0.018232416152954102, 0.018239231109619142, 0.018585599899291993, 0.018135040283203126, 0.018251840591430663, 0.01834592056274414, 0.018346176147460938, 0.0183621768951416, 0.018786304473876952, 0.019564544677734375, 0.018759679794311524, 0.01868560028076172, 0.018627199172973632, 0.01886012840270996, 0.018640512466430663, 0.01861984062194824, 0.018639423370361327, 0.01872467231750488, 0.018883039474487304, 0.018880224227905272, 0.018796768188476563, 0.018868000030517577, 0.018724863052368163, 0.018819103240966795, 0.019064767837524414, 0.018784223556518556, 0.018669408798217775, 0.018666784286499025, 0.018643903732299804, 0.018905248641967774, 0.019111072540283203, 0.018805440902709962, 0.01882316780090332, 0.018759424209594727, 0.018876384735107422, 0.018817119598388672, 0.018800832748413085, 0.018894847869873048, 0.0187391357421875, 0.01873721694946289, 0.018755584716796874, 0.01872617530822754, 0.01858563232421875, 0.018837535858154297, 0.018987520217895508, 0.018685247421264647, 0.018786815643310546, 0.019511680603027343, 0.01848931121826172, 0.018337791442871093, 0.018595199584960937, 0.018342527389526367, 0.01827052879333496, 0.01834592056274414, 0.017897024154663085, 0.018120319366455077, 0.018475648880004882, 0.0180118408203125, 0.017993791580200195, 0.017995647430419922, 0.018128896713256838, 0.0182728328704834, 0.01823686408996582, 0.01834454345703125, 0.018230464935302733, 0.0181911678314209, 0.018081375122070312, 0.018038335800170897, 0.018010688781738282, 0.018067968368530272, 0.018048704147338866, 0.017905344009399415, 0.017852832794189453, 0.017952768325805665, 0.01782809638977051, 0.017860416412353516, 0.017888639450073244, 0.017857088088989257, 0.017838336944580077, 0.01788003158569336, 0.01787766456604004, 0.017918399810791016, 0.017970783233642578, 0.017916095733642577, 0.01792195129394531, 0.017905664443969727, 0.01795686340332031, 0.017955007553100585, 0.01814713668823242, 0.0182807674407959, 0.018402687072753908, 0.01845689582824707, 0.018367712020874023, 0.018315935134887697, 0.018139263153076172, 0.01802604866027832, 0.017954336166381837, 0.018262943267822265, 0.01809212875366211, 0.018159008026123045, 0.01830963134765625, 0.01858723258972168, 0.018257759094238282, 0.018180992126464842, 0.0182609920501709, 0.018119552612304687, 0.018253631591796875, 0.018354175567626953, 0.018612192153930663, 0.018640512466430663, 0.018073888778686525, 0.0180184326171875, 0.018083200454711915, 0.018041311264038087, 0.018000032424926756, 0.01803264045715332, 0.01838876724243164, 0.01776470375061035, 0.018001920700073244, 0.0177903995513916, 0.01799635124206543, 0.017934335708618163, 0.018020383834838866, 0.017950687408447265, 0.017817024230957032, 0.017854463577270507, 0.01785094451904297, 0.018098175048828127, 0.017834175109863282, 0.017806943893432618, 0.017901952743530274, 0.01787228775024414, 0.017864831924438475, 0.017879648208618162, 0.017778495788574218, 0.0177806396484375, 0.017905664443969727, 0.017751455307006836, 0.017988351821899413, 0.01821673583984375, 0.018307167053222655, 0.01840025520324707, 0.01846540832519531, 0.018581855773925782, 0.01848320007324219, 0.018308095932006836, 0.018222272872924803, 0.01930441665649414, 0.020285343170166014, 0.018415231704711914, 0.018433664321899416, 0.018299455642700194, 0.01858572769165039, 0.018651264190673828, 0.01947225570678711, 0.018766048431396485, 0.01884752082824707, 0.01882931137084961, 0.0188538875579834, 0.018677759170532226, 0.018625663757324218, 0.018721664428710937, 0.018924768447875977, 0.018773792266845703, 0.018637823104858398, 0.0186943359375, 0.018739007949829103, 0.01884160041809082, 0.01868185615539551, 0.019314176559448244, 0.020423168182373046, 0.018675167083740233, 0.018636863708496095, 0.01875119972229004, 0.018706880569458007, 0.01865532875061035, 0.018790624618530274, 0.01884172821044922, 0.0187674560546875, 0.018928192138671876, 0.018383007049560546, 0.01911180877685547, 0.0189833927154541, 0.018937856674194335, 0.01876291275024414, 0.018572128295898438, 0.01861222457885742, 0.018558143615722656, 0.01867804718017578, 0.018591327667236326, 0.018519296646118164, 0.018783935546875, 0.018552671432495116, 0.018382944107055665, 0.018494592666625977, 0.01835523223876953, 0.018136192321777343, 0.01818448066711426, 0.01776255989074707, 0.017764671325683594, 0.017987039566040038, 0.01781920051574707, 0.017826719284057616, 0.017722496032714842, 0.018014528274536132, 0.018021184921264647, 0.01789107131958008, 0.01817190361022949, 0.017748064041137695, 0.017639328002929687, 0.017708703994750975, 0.017740224838256834, 0.01771244812011719, 0.017648416519165037, 0.01762006378173828, 0.01767193603515625, 0.01765475273132324, 0.017624576568603514, 0.01774473571777344, 0.01795552062988281, 0.01800217628479004, 0.01906537628173828, 0.01839023971557617, 0.018117599487304688, 0.01790979194641113, 0.01779836845397949, 0.017554176330566405, 0.017498111724853514, 0.017541120529174805, 0.01750204849243164, 0.017576095581054687, 0.017473407745361328, 0.017583360671997072, 0.017730432510375975, 0.017459199905395507, 0.01755673599243164, 0.017492128372192384, 0.017465408325195313, 0.017443679809570314, 0.017476959228515623, 0.01744316864013672, 0.01854182434082031, 0.02026380729675293, 0.01785241508483887, 0.017913856506347657, 0.01781171226501465, 0.017888256072998047, 0.0178920955657959, 0.018091936111450196, 0.018077152252197266, 0.01778879928588867, 0.017787647247314454, 0.017837087631225587, 0.017953760147094728, 0.01787654495239258, 0.017920480728149415, 0.017924064636230468, 0.017985919952392578, 0.018091039657592775, 0.018255552291870116, 0.01819139289855957, 0.018390815734863283, 0.018151679992675782, 0.018032480239868164, 0.017912128448486327, 0.017702592849731445, 0.017741823196411134, 0.017956159591674806, 0.01765635108947754, 0.01760892868041992, 0.01790355110168457, 0.017641471862792968, 0.017598207473754884, 0.017674495697021484, 0.017769887924194337, 0.01776291275024414, 0.01880419158935547, 0.017959455490112304, 0.017896575927734373, 0.01782054328918457, 0.017795072555541993, 0.017899168014526366, 0.01798793601989746, 0.017823808670043944, 0.017712736129760744, 0.017670495986938477, 0.01768684768676758, 0.017811199188232422, 0.017799104690551758, 0.017862655639648437, 0.017958303451538087, 0.017893503189086914, 0.01786419105529785, 0.01778278350830078, 0.017738719940185547, 0.017628576278686522, 0.017582687377929687, 0.017566848754882812, 0.01768886375427246, 0.01773833656311035, 0.017812896728515625, 0.017921760559082033, 0.018150272369384764, 0.018296640396118165, 0.01841119956970215, 0.01855539131164551, 0.01799577522277832, 0.018405376434326173, 0.018394784927368166, 0.019083295822143555, 0.018364736557006836, 0.018322751998901366, 0.018330144882202148, 0.018512224197387694, 0.01842355155944824, 0.018923583984375, 0.01856208038330078, 0.018430944442749023, 0.018400543212890624, 0.01830790328979492, 0.01833683204650879, 0.018536895751953126, 0.01825833511352539, 0.018232608795166017, 0.01853308868408203, 0.018335039138793946, 0.01840358352661133, 0.018372671127319336, 0.01834432029724121, 0.018550048828125, 0.019481311798095702, 0.018677312850952147, 0.01840787124633789, 0.018368703842163086, 0.0183438720703125, 0.018482847213745116, 0.018645248413085937, 0.01839321517944336, 0.01832943916320801, 0.018191423416137695, 0.01813804817199707, 0.01804902458190918, 0.01799100875854492, 0.01777110481262207, 0.017710847854614256, 0.017543807983398437, 0.017491552352905275, 0.017647712707519532, 0.017475584030151366, 0.017506048202514647, 0.017506111145019532, 0.01751215934753418, 0.01749068832397461, 0.01767523193359375, 0.017627967834472656, 0.017670431137084962, 0.018237344741821288, 0.01777471923828125, 0.017827423095703124, 0.017889568328857422, 0.017712480545043947, 0.017781408309936523, 0.017743072509765624, 0.017806112289428713, 0.017590240478515626, 0.017645280838012697, 0.017813823699951173, 0.017699935913085937, 0.017898399353027342, 0.01773036766052246, 0.01773664093017578, 0.017696767807006835, 0.01762713623046875, 0.017508352279663086, 0.017661951065063478, 0.017679647445678712, 0.01758451271057129, 0.017727840423583986, 0.017946624755859376, 0.017541120529174805, 0.017841728210449218, 0.01805766487121582, 0.018256927490234377, 0.018207584381103516, 0.020897823333740233, 0.018595935821533204, 0.01809119987487793, 0.017761375427246092, 0.01776406478881836, 0.017780895233154296, 0.01760268783569336, 0.017775583267211913, 0.017532831192016603, 0.017599327087402344, 0.017719615936279298, 0.017599903106689452, 0.01758172798156738, 0.01758019256591797, 0.017506784439086914, 0.017579616546630858, 0.01758835220336914, 0.017493919372558595, 0.01833203125, 0.017549312591552735, 0.017485408782958983, 0.017561792373657226, 0.01763555145263672, 0.017530879974365234, 0.017729536056518554, 0.017982784271240233, 0.017863359451293945, 0.017735551834106446, 0.017756288528442382, 0.01785638427734375, 0.01765830421447754, 0.017761951446533204, 0.01769424057006836, 0.01776896095275879, 0.01765376091003418, 0.017602752685546875, 0.01746073532104492, 0.017531423568725585, 0.017544992446899416, 0.017481727600097655, 0.01754681587219238, 0.017637760162353515, 0.01797283172607422, 0.01764352035522461, 0.017650400161743164, 0.01763699150085449, 0.017601760864257812, 0.017681312561035157, 0.01739967918395996, 0.01804966354370117, 0.01767203140258789, 0.01787446403503418, 0.017766271591186523, 0.01776291275024414, 0.017822719573974608, 0.01904332733154297, 0.018176416397094726, 0.018220640182495116, 0.017909759521484374, 0.01788703918457031, 0.017811647415161135, 0.01784012794494629, 0.01788083267211914, 0.017852672576904295, 0.018157215118408204, 0.017708511352539064, 0.017875839233398437, 0.017954879760742188, 0.018229183197021485, 0.018266111373901366, 0.018427072525024416, 0.018518783569335937, 0.018513599395751954, 0.018534175872802733, 0.018430496215820314, 0.01847324752807617, 0.018398271560668946, 0.01836310386657715, 0.018354175567626953, 0.018531391143798828, 0.018539072036743164, 0.020104703903198243, 0.02005081558227539, 0.0187108154296875, 0.018590944290161133, 0.018887168884277345, 0.018315263748168945, 0.01843404769897461, 0.018321279525756837, 0.018527551651000975, 0.01856384086608887, 0.01854080009460449, 0.018419519424438476, 0.01838307189941406, 0.01862224006652832, 0.018694143295288086, 0.01859584045410156, 0.018616031646728516, 0.018655584335327147, 0.018589599609375, 0.01870649528503418, 0.018569536209106445, 0.018463903427124024, 0.018489824295043946, 0.01838697624206543, 0.018337791442871093, 0.018305088043212892, 0.018067136764526367, 0.018082048416137694, 0.021626752853393556, 0.019337343215942382]",tokens/s,54.85355384269742,, @@ -3160,7 +3160,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 422.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 328.12 MiB is free. Process 203468 has 14.42 GiB memory in use. Of the allocated memory 14.22 GiB is allocated by PyTorch, and 91.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 422.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 328.12 MiB is free. Process 209198 has 14.42 GiB memory in use. Of the allocated memory 14.22 GiB is allocated by PyTorch, and 91.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neo,EleutherAI/gpt-neo-1.3B,EleutherAI/gpt-neo-1.3B,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,1047.486464,1393.491968,0.0,998.244352,995.459072,s,1,9.1995439453125,9.1995439453125,0.0,9.1995439453125,9.1995439453125,9.1995439453125,9.1995439453125,[9.1995439453125],,kWh,6.138048507500286e-05,6.759885228847262e-06,2.232835119599791e-05,9.046872149984803e-05,,MB,1316.298752,1619.984384,0.0,1212.153856,1174.001664,s,10,2.065621322631836,0.2065621322631836,0.001057923379171417,0.20703142547607423,0.20757976837158204,0.20789478073120116,0.20814679061889646,"[0.20546205139160156, 0.2072406005859375, 0.207006591796875, 0.2082097930908203, 0.20497737121582033, 0.20498719787597655, 0.2070726776123047, 0.20705625915527343, 0.20609901428222657, 0.207509765625]",tokens/s,1239.3365482586466,kWh,6.249252624379373e-06,6.890893309256835e-07,4.169678276638249e-06,1.1108020231943305e-05,tokens/kWh,23046411.030457206,MB,1338.712064,1619.984384,0.0,1212.153856,1174.004224,s,10,22.765936279296877,2.2765936279296874,0.010787243324537987,2.280182861328125,2.28728349609375,2.2916496826171873,2.2951426318359376,"[2.286313232421875, 2.257575927734375, 2.296015869140625, 2.26831201171875, 2.279771240234375, 2.28084423828125, 2.280594482421875, 2.27190283203125, 2.263395263671875, 2.281211181640625]",tokens/s,27.67292292621042,kWh,6.532558710478703e-05,7.2054362435162614e-06,3.535490889856203e-05,0.00010788593224686533,tokens/kWh,583950.0914340061,,s,630,22.761943065643333,0.03613006835816398,0.0007618873217216914,0.03608678245544433,0.0365433364868164,0.036807624435424804,0.038606347618103026,"[0.03631513595581055, 0.03638886260986328, 0.03630489730834961, 0.03641753768920898, 0.03643932723999024, 0.03671318435668945, 0.03632534408569336, 0.03666128158569336, 0.03630819320678711, 0.03644054412841797, 0.03646905517578125, 0.03628851318359375, 0.03624284744262695, 0.03624406433105469, 0.036173824310302735, 0.03623321533203125, 0.036154464721679686, 0.0362259521484375, 0.03627798461914063, 0.03643830490112305, 0.036062782287597656, 0.035994049072265624, 0.036268096923828125, 0.03680006408691406, 0.0360816650390625, 0.03593801498413086, 0.03613056182861328, 0.03600476837158203, 0.03606118392944336, 0.03620249557495117, 0.036171775817871094, 0.03632467269897461, 0.03680716705322266, 0.03615094375610352, 0.03638457489013672, 0.03647151947021485, 0.03633148956298828, 0.03638070297241211, 0.036269855499267575, 0.03617337417602539, 0.036276897430419924, 0.036345855712890625, 0.036216705322265626, 0.03619782257080078, 0.03608000183105469, 0.03618572616577148, 0.03661606216430664, 0.03612713623046875, 0.03635036849975586, 0.03609395217895508, 0.03621683120727539, 0.03639839935302734, 0.036049633026123046, 0.0360807991027832, 0.03611334228515625, 0.036359870910644534, 0.0361064338684082, 0.03632332611083984, 0.03639449691772461, 0.03618790435791015, 0.03641622543334961, 0.036286495208740235, 0.03637452697753906, 0.036089855194091795, 0.03622611236572266, 0.03609417724609375, 0.03588927841186523, 0.03541183853149414, 0.035952865600585936, 0.036131328582763675, 0.03614672088623047, 0.036176353454589846, 0.03618815994262695, 0.03607932662963867, 0.03604246520996094, 0.0361060791015625, 0.036063968658447264, 0.036073471069335936, 0.03620364761352539, 0.03607846450805664, 0.036116481781005856, 0.03645183944702148, 0.036114944458007815, 0.03611580657958984, 0.03637315368652344, 0.036345855712890625, 0.03615913772583008, 0.036267616271972655, 0.036692737579345706, 0.03616358566284179, 0.03586867141723633, 0.03549507141113281, 0.035915775299072264, 0.03579375839233399, 0.03560451126098633, 0.03560444641113281, 0.03557891082763672, 0.035503070831298826, 0.035604480743408204, 0.03596492767333984, 0.03577974319458008, 0.03558281707763672, 0.03577446365356445, 0.03547663879394531, 0.03570979309082031, 0.035563518524169925, 0.03548364639282227, 0.03538739013671875, 0.03550931167602539, 0.035275104522705075, 0.03582419204711914, 0.03569052886962891, 0.035741504669189454, 0.03582585525512695, 0.035827713012695314, 0.03561593627929688, 0.035447616577148434, 0.03531478500366211, 0.03510774230957031, 0.03537510299682617, 0.03563430404663086, 0.035447486877441405, 0.035483230590820314, 0.03567196655273437, 0.03564358520507813, 0.03540838241577148, 0.036004417419433596, 0.035833854675292966, 0.03548160171508789, 0.035438591003417966, 0.03548364639282227, 0.035315711975097655, 0.03712575912475586, 0.03573183822631836, 0.03578019332885742, 0.036105758666992185, 0.03603340911865234, 0.03621478271484375, 0.03589529418945313, 0.03587891387939453, 0.035776512145996094, 0.038744224548339846, 0.043791969299316405, 0.0459975357055664, 0.03665795135498047, 0.03726889419555664, 0.03670412826538086, 0.037017887115478515, 0.03653023910522461, 0.0365428466796875, 0.0364620475769043, 0.036416030883789065, 0.03625369644165039, 0.03636627197265625, 0.03639302444458008, 0.03640444946289063, 0.0364095344543457, 0.03645091247558594, 0.0363392333984375, 0.036831710815429686, 0.036397056579589845, 0.03664896011352539, 0.03649331283569336, 0.03610009765625, 0.035910816192626954, 0.035939167022705075, 0.035846080780029294, 0.03612377548217773, 0.03573356628417969, 0.035773311614990234, 0.03599788665771484, 0.03612854385375976, 0.03589532852172852, 0.0359444465637207, 0.03590348815917969, 0.03618406295776367, 0.0360263671875, 0.03596083068847656, 0.03633273696899414, 0.035928768157958986, 0.03603263854980469, 0.0359659538269043, 0.03593667221069336, 0.03576073455810547, 0.03557331085205078, 0.03573907089233398, 0.03591993713378906, 0.035851200103759764, 0.03592547225952149, 0.03603308868408203, 0.03596083068847656, 0.03570217514038086, 0.03559513473510742, 0.03530902481079102, 0.03533849716186523, 0.035514240264892576, 0.036141151428222655, 0.036136768341064454, 0.03607686233520508, 0.036037151336669924, 0.03591551971435547, 0.03571161651611328, 0.03590758514404297, 0.03583590316772461, 0.03603046417236328, 0.036050945281982424, 0.03593625640869141, 0.03629040145874023, 0.03593977737426758, 0.036444896697998046, 0.036087806701660154, 0.03611795043945312, 0.03611091232299805, 0.0362630729675293, 0.0359637451171875, 0.03611603164672852, 0.03605686569213867, 0.03604665756225586, 0.03621897506713867, 0.03638502502441406, 0.03612416076660156, 0.0359444465637207, 0.03610726547241211, 0.0363675537109375, 0.03611936187744141, 0.036797534942626955, 0.036170654296875, 0.03596822357177734, 0.03600259017944336, 0.035985408782958986, 0.0361085433959961, 0.03600950241088867, 0.03635398483276367, 0.03862748718261719, 0.03642176055908203, 0.035934207916259765, 0.03549184036254883, 0.035489791870117186, 0.03578265762329102, 0.03605500793457031, 0.036081695556640626, 0.03584342575073242, 0.03549801635742188, 0.035299457550048825, 0.035197311401367185, 0.03527692794799805, 0.035186622619628904, 0.035397022247314454, 0.03545353698730469, 0.03559596633911133, 0.03563484954833984, 0.03836928176879883, 0.03598332977294922, 0.03587376022338867, 0.035811328887939455, 0.035837310791015625, 0.03582169723510742, 0.03607807922363281, 0.03552175903320313, 0.03606403350830078, 0.03562905502319336, 0.0360079345703125, 0.03608575820922852, 0.03604044723510742, 0.03588940811157226, 0.036046527862548826, 0.036182334899902344, 0.03597721481323242, 0.03600352096557617, 0.03594249725341797, 0.03581267166137695, 0.035754913330078124, 0.0355404167175293, 0.03668751907348633, 0.035715744018554686, 0.035975425720214844, 0.036018177032470705, 0.036208641052246096, 0.03609600067138672, 0.035905025482177735, 0.03604326248168945, 0.036239360809326174, 0.036308609008789065, 0.03619190216064453, 0.03636912155151367, 0.036675582885742186, 0.03707411193847656, 0.03697747039794922, 0.03709952163696289, 0.03701929473876953, 0.036439937591552736, 0.03645196914672852, 0.036380992889404294, 0.03628441619873047, 0.03621532821655273, 0.03634902572631836, 0.03626691055297852, 0.03624905776977539, 0.03640169525146485, 0.036239360809326174, 0.036224414825439456, 0.036799072265625, 0.03633135986328125, 0.036369728088378905, 0.036363105773925784, 0.036359169006347655, 0.036365310668945314, 0.036428958892822265, 0.03614761734008789, 0.03631507110595703, 0.03589580917358398, 0.03579904174804688, 0.03594649505615234, 0.03587481689453125, 0.03630617523193359, 0.03628630447387695, 0.03569084930419922, 0.03552854537963867, 0.03561033630371094, 0.04100960159301758, 0.035587425231933596, 0.03542287826538086, 0.03538236618041992, 0.03623004913330078, 0.03603235244750977, 0.03594460678100586, 0.03570073699951172, 0.03851878356933594, 0.03585971069335937, 0.035526496887207035, 0.03513817596435547, 0.03518288040161133, 0.03566947174072266, 0.035435039520263674, 0.035604480743408204, 0.03538905715942383, 0.03527923202514648, 0.03544678497314453, 0.035794944763183595, 0.03634918212890625, 0.03610265731811523, 0.036398944854736326, 0.03721420669555664, 0.03613737487792969, 0.035872608184814456, 0.036019935607910156, 0.03630329513549805, 0.03643734359741211, 0.03562972640991211, 0.03586492919921875, 0.03560524749755859, 0.03623004913330078, 0.0363765754699707, 0.03624755096435547, 0.03659161758422851, 0.03575603103637695, 0.03544416046142578, 0.03628268814086914, 0.03634991836547852, 0.036260128021240234, 0.03649331283569336, 0.03633724975585938, 0.036647327423095705, 0.03644163131713867, 0.03692303848266602, 0.0366569938659668, 0.03649212646484375, 0.03632326507568359, 0.03637267303466797, 0.03637990570068359, 0.03635891342163086, 0.03686604690551758, 0.036517887115478515, 0.03700735855102539, 0.03647283172607422, 0.036547744750976566, 0.036413345336914066, 0.0363570556640625, 0.03855459213256836, 0.036674560546875, 0.036335166931152345, 0.036018623352050784, 0.035694591522216795, 0.03568025588989258, 0.0360079345703125, 0.03698908615112305, 0.03581353759765625, 0.03591446304321289, 0.036236255645751954, 0.036239360809326174, 0.03592192077636719, 0.036044769287109375, 0.03585436630249023, 0.03557580947875977, 0.03546908950805664, 0.035727039337158206, 0.04241667175292969, 0.03923932647705078, 0.03579702377319336, 0.03566828918457031, 0.03592752075195312, 0.03591222381591797, 0.03601747131347656, 0.035754688262939455, 0.03562656021118164, 0.03557011032104492, 0.03562496185302735, 0.035624286651611325, 0.035830432891845704, 0.03587071990966797, 0.035880863189697264, 0.03617728042602539, 0.03585004806518555, 0.036109214782714845, 0.0357498893737793, 0.035703968048095704, 0.0362402229309082, 0.03618201446533203, 0.03595017623901367, 0.035665599822998044, 0.03557449722290039, 0.03554508972167969, 0.03590566253662109, 0.03616748809814453, 0.035856063842773435, 0.03563763046264649, 0.03566947174072266, 0.035846687316894534, 0.0364738883972168, 0.036293598175048826, 0.036238624572753904, 0.03631135940551758, 0.036729248046875, 0.03618764877319336, 0.0363524169921875, 0.03630259323120117, 0.036331871032714846, 0.036274177551269535, 0.036173473358154296, 0.03651824188232422, 0.03658863830566406, 0.036259166717529295, 0.03645708847045898, 0.03639292907714844, 0.036344894409179686, 0.035871742248535156, 0.03554099273681641, 0.03529523086547852, 0.03530137634277344, 0.035619998931884764, 0.03618697738647461, 0.03595468902587891, 0.036224384307861325, 0.03605295944213867, 0.03575260925292969, 0.03583590316772461, 0.035676063537597655, 0.035468929290771486, 0.03545471954345703, 0.03521200180053711, 0.035398815155029295, 0.036025150299072266, 0.0357540168762207, 0.036066753387451175, 0.03582534408569336, 0.03592057418823242, 0.03614739227294922, 0.03609190368652344, 0.03606902313232422, 0.03583158493041992, 0.03618649673461914, 0.03573574447631836, 0.03577967834472656, 0.03621571350097656, 0.03596492767333984, 0.035764320373535156, 0.03631504058837891, 0.036435966491699216, 0.0372262077331543, 0.03676803207397461, 0.036029792785644534, 0.03587676620483399, 0.035832576751708985, 0.03578060913085938, 0.036122623443603515, 0.035899391174316404, 0.03649846267700195, 0.03775968170166016, 0.036117889404296874, 0.03606780624389649, 0.03690335845947266, 0.03605692672729492, 0.03600751876831055, 0.03616128158569336, 0.03617670440673828, 0.036083713531494144, 0.03609132766723633, 0.03614777755737305, 0.036261886596679685, 0.03626598358154297, 0.036222976684570314, 0.03626355361938476, 0.03615071868896484, 0.03620345687866211, 0.03631382369995117, 0.03618953704833985, 0.03646716690063476, 0.03631052780151367, 0.036324031829833986, 0.03648716735839844, 0.036190208435058595, 0.0365588493347168, 0.03680799865722656, 0.036307647705078126, 0.03648876953125, 0.036630977630615236, 0.03636633682250977, 0.03693772888183594, 0.036206592559814454, 0.03641753768920898, 0.03631039810180664, 0.03772684860229492, 0.03576617431640625, 0.035458656311035154, 0.03545753479003906, 0.03533824157714844, 0.03544863891601562, 0.03528278350830078, 0.03553519821166992, 0.035683551788330076, 0.03586947250366211, 0.03572447967529297, 0.03551110458374023, 0.035422206878662106, 0.03552249526977539, 0.035272640228271486, 0.035461246490478514, 0.03572323226928711, 0.036388607025146485, 0.036030750274658206, 0.036111679077148434, 0.03584451293945313, 0.03639865493774414, 0.03620060729980469, 0.036041278839111325, 0.03581542587280274, 0.035672065734863284, 0.035432449340820314, 0.03598745727539063, 0.036111934661865235, 0.03669561767578125, 0.035289985656738285, 0.035278335571289066, 0.03565824127197265, 0.03575807952880859, 0.03544582366943359, 0.03556857681274414, 0.03561417770385742, 0.03551286315917969, 0.03536076736450195, 0.03554099273681641, 0.03572531127929687, 0.03574915313720703, 0.035482303619384765, 0.03573276901245117, 0.03549078369140625, 0.035643199920654296, 0.03655267333984375, 0.03637152099609375, 0.0363570556640625, 0.03666329574584961, 0.03645587158203125, 0.03635843276977539, 0.03630313491821289, 0.03613654327392578, 0.036192798614501955, 0.036124542236328126, 0.036534271240234374, 0.0364150390625, 0.03649491119384766, 0.03656179046630859, 0.036296703338623046, 0.036347904205322266, 0.03629171371459961, 0.03660070419311524, 0.036474815368652345, 0.036467937469482424, 0.035936351776123046, 0.036082176208496096, 0.036319488525390624, 0.03579084777832031, 0.0355676155090332, 0.03625574493408203, 0.036536319732666016, 0.03630080032348633, 0.036446208953857424, 0.03794851303100586, 0.035977886199951174, 0.03591932678222656, 0.03552259063720703, 0.035584766387939455, 0.03611033630371094, 0.0355810546875, 0.03575616073608399, 0.03627494430541992, 0.0366196174621582, 0.036238304138183595, 0.03626361465454102, 0.03598057556152344, 0.03578339385986328, 0.035741119384765624, 0.03582624053955078, 0.035706241607666014, 0.03554982376098633, 0.03544063949584961, 0.03544073486328125, 0.035727264404296875, 0.03570406341552734, 0.03633023834228516, 0.03594623947143555, 0.03628249740600586, 0.0366429443359375, 0.03623321533203125, 0.036459617614746094, 0.03705948638916016, 0.03685171127319336, 0.03671244812011719, 0.03612089538574219, 0.03596681594848633, 0.03631497573852539]",tokens/s,27.67777769161179,, @@ -3209,7 +3209,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 28.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 4.12 MiB is free. Process 178182 has 14.73 GiB memory in use. Of the allocated memory 14.49 GiB is allocated by PyTorch, and 137.12 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 28.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 4.12 MiB is free. Process 184047 has 14.73 GiB memory in use. Of the allocated memory 14.49 GiB is allocated by PyTorch, and 137.12 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen,Qwen/Qwen-72B,Qwen/Qwen-72B,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -3291,7 +3291,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 98.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 80.12 MiB is free. Process 124593 has 14.66 GiB memory in use. Of the allocated memory 14.13 GiB is allocated by PyTorch, and 428.13 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 98.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 80.12 MiB is free. Process 130310 has 14.66 GiB memory in use. Of the allocated memory 14.13 GiB is allocated by PyTorch, and 428.13 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2_moe,Qwen/Qwen1.5-MoE-A2.7B,Qwen/Qwen1.5-MoE-A2.7B,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,1078.571008,9784.19712,0.0,9388.949504,9304.608768,s,1,33.56804296875,33.56804296875,0.0,33.56804296875,33.56804296875,33.56804296875,33.56804296875,[33.56804296875],,kWh,0.0007570145398666909,8.349687089604282e-05,0.00027338716315400413,0.0011138985739167378,,MB,1442.779136,10197.336064,0.0,9789.505536,9597.896704,s,10,8.024159423828126,0.8024159423828126,0.0037759130374792073,0.8023334350585938,0.8048954711914063,0.8080676086425781,0.8106053186035157,"[0.8041905517578125, 0.803518798828125, 0.7991083984375, 0.81123974609375, 0.8016114501953125, 0.8036441650390626, 0.8002913818359375, 0.7958880615234375, 0.8020947265625, 0.8025721435546875]",tokens/s,319.036532648885,kWh,2.3606744060574178e-05,2.603420795459561e-06,1.4281101168460964e-05,4.0491266024494706e-05,tokens/kWh,6322351.092829152,MB,1471.410176,10197.336064,0.0,9789.505536,9597.899264,s,10,376.05490625,37.605490625,0.1423211389298986,37.6388125,37.76045625,37.78117734375,37.79775421875,"[37.7558515625, 37.45809375, 37.6530859375, 37.8018984375, 37.680734375, 37.5471328125, 37.35794921875, 37.436078125, 37.73954296875, 37.6245390625]",tokens/s,1.6752872772817333,kWh,0.001104775240299009,0.00012186504322589423,0.0004200425518457404,0.0016466828353706438,tokens/kWh,38258.73364728408,,s,630,376.05202130126935,0.5969079703194754,0.00458039882502152,0.5970689392089843,0.6014270935058594,0.6045695404052734,0.6133359204101563,"[0.6004793090820313, 0.59568603515625, 0.5983150024414062, 0.5971712646484375, 0.5964830322265625, 0.5946448364257813, 0.6069822998046875, 0.603926513671875, 0.5973463134765625, 0.6034421997070313, 0.5990032348632812, 0.600175048828125, 0.600375732421875, 0.601417724609375, 0.5993401489257812, 0.598215576171875, 0.5988536376953125, 0.6002418212890624, 0.5966739501953126, 0.5974447631835937, 0.5969735107421875, 0.5988904418945312, 0.5986395263671875, 0.6071869506835937, 0.60400390625, 0.5985402221679688, 0.5982113647460937, 0.5988589477539062, 0.597866943359375, 0.5977251586914063, 0.602064453125, 0.597414306640625, 0.5975443115234375, 0.59811083984375, 0.5987000122070313, 0.5957073974609375, 0.5968655395507813, 0.59815283203125, 0.596900146484375, 0.5984912109375, 0.60626123046875, 0.6044815063476563, 0.600195068359375, 0.5990802001953125, 0.5992078857421875, 0.598090087890625, 0.5971746215820313, 0.6007172241210937, 0.5994456787109375, 0.5960455932617188, 0.59719921875, 0.5961911010742188, 0.5984378662109375, 0.60060986328125, 0.5966236572265625, 0.5973423461914062, 0.603420166015625, 0.605086669921875, 0.6004592895507812, 0.5994905395507812, 0.5986683959960938, 0.5983241577148437, 0.5999329223632812, 0.5987039794921875, 0.6004655151367188, 0.595375732421875, 0.596560302734375, 0.5953699951171875, 0.594818603515625, 0.5980304565429687, 0.5979303588867187, 0.5962640380859375, 0.6027191162109375, 0.5994303588867187, 0.59736962890625, 0.59692236328125, 0.5960028076171875, 0.591583251953125, 0.5895291137695312, 0.5947446899414063, 0.5901292724609375, 0.5896566162109375, 0.5879984130859375, 0.5889564819335937, 0.5868004150390626, 0.5895543823242188, 0.5913001098632813, 0.5909202880859376, 0.5916732177734375, 0.5997586059570312, 0.5964392700195312, 0.5928029174804688, 0.5908571166992187, 0.5948597412109375, 0.5899429931640625, 0.5903914184570312, 0.5945098266601563, 0.59052392578125, 0.5895480346679688, 0.590845947265625, 0.59912939453125, 0.597419921875, 0.5991373901367187, 0.60061669921875, 0.60052685546875, 0.6007255249023438, 0.6045711059570312, 0.6095562133789062, 0.5999685668945313, 0.5974201049804687, 0.5981220092773437, 0.5915159912109375, 0.5926892700195312, 0.5878067016601562, 0.5925120239257813, 0.5952144165039063, 0.58997607421875, 0.5930389404296875, 0.5917305297851563, 0.5916942749023437, 0.588666748046875, 0.5904390258789063, 0.5925928955078125, 0.6007597045898437, 0.5906499633789063, 0.5920354614257812, 0.59857275390625, 0.5967245483398438, 0.6008688354492188, 0.5964656372070313, 0.5970862426757813, 0.5971251220703125, 0.59636328125, 0.596336669921875, 0.5975347290039063, 0.6004019165039063, 0.59886181640625, 0.599341064453125, 0.6138900756835938, 0.6055706176757812, 0.5977914428710938, 0.5973641967773438, 0.5989378051757812, 0.597885009765625, 0.596038818359375, 0.5986885986328125, 0.59691748046875, 0.5977874145507812, 0.5956680908203125, 0.59686181640625, 0.5982411499023438, 0.5929468383789063, 0.5946557006835937, 0.59449365234375, 0.5981058959960938, 0.606013427734375, 0.5958430786132812, 0.59447705078125, 0.5954662475585938, 0.5960722045898438, 0.5968630981445312, 0.59592919921875, 0.59464501953125, 0.5944805297851562, 0.6029273071289063, 0.60008642578125, 0.59781787109375, 0.5953638305664063, 0.5948907470703125, 0.595585205078125, 0.5950830688476563, 0.597391357421875, 0.6060230102539063, 0.5968836669921875, 0.5955005493164063, 0.5992354736328125, 0.5962216186523438, 0.599525390625, 0.6015245361328125, 0.5963460693359375, 0.5948171997070313, 0.595548828125, 0.5993286743164062, 0.5966597900390626, 0.595879638671875, 0.5945802001953125, 0.5940714721679687, 0.5949031372070313, 0.5992950439453125, 0.5966018676757813, 0.5981992797851563, 0.5979851684570312, 0.59512255859375, 0.597765625, 0.5947230834960937, 0.5960410766601563, 0.5953389282226562, 0.5958194580078126, 0.5947756958007813, 0.5951156005859375, 0.6132478637695312, 0.6152252807617188, 0.6057738037109375, 0.6098370361328125, 0.6183765869140625, 0.6193934936523438, 0.6091259765625, 0.6064729614257812, 0.603739013671875, 0.6000545043945312, 0.6011077270507813, 0.5995889282226563, 0.5968084716796875, 0.59753466796875, 0.5952074584960938, 0.59660986328125, 0.5948067626953125, 0.5954007568359375, 0.5988023071289063, 0.5995245971679688, 0.6012404174804687, 0.603479248046875, 0.597592529296875, 0.5988560180664062, 0.5971099853515625, 0.598666015625, 0.5959555053710938, 0.6005142822265624, 0.5991405029296875, 0.599218505859375, 0.5973622436523438, 0.5999540405273438, 0.5991586303710937, 0.5985156860351563, 0.597485595703125, 0.598877685546875, 0.5990526123046875, 0.6079464111328124, 0.5997179565429688, 0.59822119140625, 0.6002606811523438, 0.6012161865234374, 0.598807373046875, 0.5991383056640625, 0.5984088134765625, 0.5982332763671875, 0.596343017578125, 0.5982119140625, 0.598999755859375, 0.5978890380859375, 0.5959164428710938, 0.5959904174804688, 0.6044815063476563, 0.6069657592773438, 0.6004791259765625, 0.5995720825195312, 0.5961649780273437, 0.598435791015625, 0.5975311889648437, 0.5995951538085937, 0.5956484985351562, 0.6011754760742187, 0.5952840576171875, 0.5979202880859374, 0.5974712524414062, 0.5964451293945312, 0.6007623901367187, 0.59650634765625, 0.5972930297851563, 0.6051753540039062, 0.600068359375, 0.6002447509765625, 0.59643701171875, 0.5965925903320313, 0.59482470703125, 0.5983768920898438, 0.596989990234375, 0.5959901733398437, 0.5953162841796875, 0.5955325317382812, 0.59702685546875, 0.5968765869140625, 0.5952971801757813, 0.5957722778320312, 0.5971724853515625, 0.60042919921875, 0.6080061645507813, 0.5949615478515625, 0.5953770141601562, 0.5987880859375, 0.5982637939453125, 0.5975543212890625, 0.5996019897460938, 0.595951416015625, 0.5963368530273437, 0.5959393310546875, 0.594935791015625, 0.5952304077148437, 0.597755859375, 0.604116943359375, 0.5960966796875, 0.5985327758789063, 0.6015114135742188, 0.6076749267578125, 0.5980487670898438, 0.6006435546875, 0.5979259033203125, 0.595861328125, 0.5966128540039063, 0.5995729370117188, 0.5943173217773438, 0.597823486328125, 0.5940162353515624, 0.5958881225585938, 0.5932373657226563, 0.6018928833007813, 0.5966554565429687, 0.5999664306640625, 0.6031705322265625, 0.5981074829101563, 0.5961983032226562, 0.5982576904296875, 0.59667578125, 0.5979472045898437, 0.5966356201171875, 0.5988740844726562, 0.5950750732421874, 0.59464501953125, 0.5959188232421875, 0.597173828125, 0.6030330810546874, 0.59887060546875, 0.5971211547851563, 0.5971533203125, 0.6003432006835937, 0.6133718872070313, 0.595173583984375, 0.5976878662109375, 0.5946596069335938, 0.5977293090820313, 0.590903076171875, 0.5955565185546875, 0.5908944091796875, 0.5900397338867187, 0.5872164916992187, 0.5898387451171875, 0.588695068359375, 0.5992227783203125, 0.5991546630859375, 0.6000045776367188, 0.5997272338867188, 0.6043222045898438, 0.6048569946289063, 0.5994698486328125, 0.6001105346679687, 0.5976450805664062, 0.597142333984375, 0.5974384765625, 0.5970759887695313, 0.5905541381835937, 0.593681396484375, 0.5903206176757813, 0.5895465087890625, 0.5894717407226563, 0.589731689453125, 0.5914810180664063, 0.5898090209960938, 0.5908190307617187, 0.59707421875, 0.5995111694335937, 0.5907072143554688, 0.5919744262695312, 0.58979736328125, 0.594145263671875, 0.5923717041015625, 0.59804296875, 0.5930797119140625, 0.593076171875, 0.5949239501953125, 0.592190673828125, 0.592826416015625, 0.5938594970703125, 0.5946546630859375, 0.5976929321289063, 0.6159707641601563, 0.595388427734375, 0.5930552978515625, 0.596017578125, 0.5924225463867188, 0.5903302001953125, 0.5951565551757813, 0.5902269287109375, 0.5919303588867187, 0.5909688110351563, 0.5929512939453125, 0.5906862182617187, 0.5879285888671875, 0.591690673828125, 0.5901394653320312, 0.592201171875, 0.5956367797851563, 0.6000435180664062, 0.5915443115234374, 0.5951054077148438, 0.591812744140625, 0.5918026123046874, 0.58849072265625, 0.595726318359375, 0.5919313354492187, 0.5972828369140625, 0.5893836669921875, 0.593006591796875, 0.59108740234375, 0.590979248046875, 0.5911411743164062, 0.5916771240234375, 0.5907025756835937, 0.59791357421875, 0.5995593872070313, 0.5906926879882812, 0.5898677368164063, 0.593170166015625, 0.5920870361328125, 0.589391845703125, 0.5958204956054688, 0.590388427734375, 0.5910286254882813, 0.5901746826171875, 0.5900157470703125, 0.5913583374023438, 0.5913128662109375, 0.590477783203125, 0.5902556762695312, 0.5918931884765625, 0.5957220458984375, 0.600690673828125, 0.590388671875, 0.59058642578125, 0.5907967529296875, 0.5913981323242188, 0.59211083984375, 0.5902474365234375, 0.587992919921875, 0.5887446899414063, 0.5863717041015625, 0.5877821655273437, 0.5867615966796875, 0.5886040649414063, 0.5885358276367187, 0.5941903076171875, 0.5941514282226562, 0.5992528686523437, 0.590245361328125, 0.587977294921875, 0.58908984375, 0.5893883056640625, 0.5897506713867188, 0.58909521484375, 0.587362060546875, 0.5910056762695313, 0.5893345336914062, 0.5890919189453125, 0.5894024658203125, 0.58707568359375, 0.588322509765625, 0.5886635131835938, 0.5929758911132812, 0.5966889038085937, 0.5970636596679687, 0.5937377319335938, 0.5945128784179687, 0.5954426879882813, 0.59519970703125, 0.6000829467773438, 0.5966046752929688, 0.5942167358398438, 0.5963052978515625, 0.5975029907226562, 0.5953101196289062, 0.5963593139648438, 0.5953899536132813, 0.594767578125, 0.592407470703125, 0.5954293823242187, 0.60524951171875, 0.600276123046875, 0.5980428466796875, 0.5945698852539063, 0.5956915283203125, 0.5968141479492187, 0.599828125, 0.596664306640625, 0.59563134765625, 0.5962001953125, 0.595198974609375, 0.5977835693359375, 0.5982815551757813, 0.5990254516601563, 0.5986903076171874, 0.6004895629882813, 0.60535888671875, 0.6024540405273437, 0.6028369750976562, 0.598262939453125, 0.598118408203125, 0.59943115234375, 0.5952921752929687, 0.598362060546875, 0.5996151733398437, 0.5961767578125, 0.5982315063476562, 0.5971517333984375, 0.5977986450195313, 0.5980674438476562, 0.5999363403320312, 0.6010046997070313, 0.6022569580078125, 0.6052492065429688, 0.5988505859375, 0.5998998413085938, 0.5965784912109375, 0.5971251220703125, 0.599250732421875, 0.5977110595703125, 0.59770263671875, 0.598725830078125, 0.6026717529296876, 0.5997733764648437, 0.60202392578125, 0.5994004516601562, 0.5970841674804688, 0.59827197265625, 0.6030213012695312, 0.61406005859375, 0.6005678100585937, 0.5983573608398437, 0.6053334350585937, 0.59907666015625, 0.5979268188476563, 0.6012926025390625, 0.5956641845703124, 0.5959152221679688, 0.597344482421875, 0.596909423828125, 0.5984203491210938, 0.5971619873046875, 0.5981572875976563, 0.5977435913085938, 0.597190673828125, 0.60233935546875, 0.604567626953125, 0.5984596557617188, 0.5972875366210938, 0.5958082275390625, 0.5963690795898438, 0.5995562133789063, 0.5963605346679688, 0.5962469482421875, 0.595840576171875, 0.5970195922851562, 0.5972930297851563, 0.5991724243164063, 0.5969491577148438, 0.5959049682617188, 0.5978124389648437, 0.6027620849609375, 0.603598876953125, 0.6010101928710937, 0.595933349609375, 0.5975206909179688, 0.594540283203125, 0.5983009643554688, 0.5924844360351562, 0.5947349853515626, 0.5926602172851563, 0.5940177612304688, 0.59410693359375, 0.5937256469726563, 0.595998779296875, 0.5968008422851563, 0.5963701171875, 0.6013972778320312, 0.6001024780273437, 0.5975904541015625, 0.59519384765625, 0.5953491821289062, 0.595202392578125, 0.5945548706054687, 0.5977111206054687, 0.5951444702148437, 0.5948168334960937, 0.5969163818359375, 0.5929103393554688, 0.5976002807617188, 0.5947371215820313, 0.5980469360351562, 0.5974894409179687, 0.599140380859375, 0.60188671875, 0.6043787231445312, 0.6006676025390625, 0.59879931640625, 0.5957635498046875, 0.5973761596679688, 0.5968289184570312, 0.5975631713867188, 0.5970902099609375, 0.5986216430664062, 0.5950140380859374, 0.5960767822265625, 0.5955543212890625, 0.597099609375, 0.596447265625, 0.5972276000976563, 0.5998764038085938, 0.60377294921875, 0.605216796875, 0.5977272338867188, 0.5971763305664063, 0.5968251342773437, 0.596970458984375, 0.599923828125, 0.5974495849609375, 0.5984010620117187, 0.5963182373046875, 0.5974935913085937, 0.595628173828125, 0.595881103515625, 0.5988564453125, 0.5982308959960938]",tokens/s,1.6753001295405434,, @@ -3390,7 +3390,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 110.12 MiB is free. Process 163178 has 14.63 GiB memory in use. Of the allocated memory 14.47 GiB is allocated by PyTorch, and 51.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 110.12 MiB is free. Process 169134 has 14.63 GiB memory in use. Of the allocated memory 14.47 GiB is allocated by PyTorch, and 51.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,Deci/DeciCoder-1b,Deci/DeciCoder-1b,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -3478,7 +3478,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 448.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 436.12 MiB is free. Process 169647 has 14.31 GiB memory in use. Of the allocated memory 14.14 GiB is allocated by PyTorch, and 61.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 448.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 436.12 MiB is free. Process 175598 has 14.31 GiB memory in use. Of the allocated memory 14.14 GiB is allocated by PyTorch, and 61.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,togethercomputer/RedPajama-INCITE-Base-7B-v0.1,togethercomputer/RedPajama-INCITE-Base-7B-v0.1,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,,True,MB,1043.406848,5037.228032,0.0,4634.70592,4621.451776,s,1,14.5569150390625,14.5569150390625,0.0,14.5569150390625,14.5569150390625,14.5569150390625,14.5569150390625,[14.5569150390625],,kWh,0.00020359975043334846,2.24512509358292e-05,6.792283211599914e-05,0.0002939738334851768,,MB,1344.303104,5911.740416,0.0,5496.635392,5337.002496,s,10,9.430091003417969,0.9430091003417967,0.004016338338506861,0.9434315490722656,0.9475170654296875,0.9486765625,0.9496041601562499,"[0.9350957641601563, 0.9436987915039062, 0.9439705200195313, 0.9393595581054688, 0.943164306640625, 0.94009912109375, 0.9416194458007813, 0.9498360595703125, 0.9472593994140625, 0.945988037109375]",tokens/s,271.4713992762233,kWh,2.7483738558712364e-05,3.0309794922599e-06,1.8280494422364728e-05,4.8795212473336996e-05,tokens/kWh,5246416.339305526,MB,1401.081856,5913.837568,0.0,5496.635392,5337.005056,s,10,50.873770507812495,5.0873770507812495,0.0128487359401144,5.08795947265625,5.09986904296875,5.105783642578125,5.110515322265625,"[5.08163134765625, 5.09122900390625, 5.06926220703125, 5.071123046875, 5.07480419921875, 5.095923828125, 5.08468994140625, 5.09485400390625, 5.1116982421875, 5.0985546875]",tokens/s,12.383591656593515,kWh,0.0001488463528554562,1.641833582762712e-05,9.600945812063467e-05,0.0002612741468037179,tokens/kWh,241126.03857177158,,s,630,50.871336517333994,0.08074815320211744,0.0012559569604315007,0.08051854705810546,0.08123368911743165,0.08146673469543457,0.08965459014892578,"[0.09022844696044922, 0.08232227325439453, 0.08083660888671874, 0.08053097534179687, 0.0812077407836914, 0.08055197143554688, 0.08018329620361328, 0.0799477767944336, 0.08024285125732422, 0.08000809478759766, 0.08000198364257813, 0.080216064453125, 0.08036351776123046, 0.08029933166503907, 0.08015737915039063, 0.08152857971191406, 0.08069760131835937, 0.08047821044921875, 0.08031378936767578, 0.08035334777832032, 0.08033071899414063, 0.08127875518798829, 0.08084352111816406, 0.08048844909667968, 0.08046562957763671, 0.08022188568115235, 0.08030473327636718, 0.08030413055419922, 0.08024473571777344, 0.0802606430053711, 0.08023401641845704, 0.08011856079101562, 0.0803165740966797, 0.08030729675292969, 0.08030223846435547, 0.08045849609375, 0.08044089508056641, 0.08047660827636718, 0.08038297271728516, 0.08039936065673828, 0.08046514892578124, 0.08037872314453125, 0.08056748962402344, 0.08055391693115234, 0.08057315063476563, 0.08064768218994141, 0.08050540924072265, 0.08045104217529297, 0.08051152038574219, 0.08067254638671875, 0.08063410949707031, 0.08062732696533204, 0.08076262664794921, 0.08055033874511719, 0.08053366088867188, 0.08055398559570312, 0.08045516967773438, 0.08048486328125, 0.08052867126464844, 0.0804751968383789, 0.08050841522216796, 0.08059439849853516, 0.08067465972900391, 0.09018617248535156, 0.08179727935791016, 0.08105574035644532, 0.08054573059082032, 0.0804588165283203, 0.08032918548583984, 0.08019993591308594, 0.08018265533447265, 0.08053852844238281, 0.08011804962158203, 0.08013212585449218, 0.08003334045410156, 0.080648193359375, 0.0804901123046875, 0.0805565414428711, 0.08037789154052734, 0.08038742065429688, 0.08080857849121094, 0.08047737884521484, 0.08039711761474609, 0.08044544219970703, 0.08030003356933593, 0.08020172882080077, 0.08028272247314452, 0.08021846771240235, 0.08028377532958984, 0.08038361358642578, 0.0805261459350586, 0.08032876586914063, 0.0809450912475586, 0.08087363433837891, 0.08214307403564453, 0.08042700958251953, 0.08046562957763671, 0.08060342407226563, 0.0817456283569336, 0.08101696014404297, 0.08074050903320312, 0.08064733123779297, 0.08058866882324218, 0.0805693130493164, 0.08068710327148437, 0.08089424133300781, 0.08064982604980468, 0.08051465606689454, 0.08040067291259766, 0.08074409484863282, 0.08082463836669922, 0.08120304107666015, 0.08057695770263672, 0.08046367645263672, 0.08351558685302735, 0.08134041595458984, 0.08083865356445312, 0.08048639678955079, 0.08036579132080078, 0.08037478637695313, 0.0804441909790039, 0.08049247741699218, 0.08045171356201172, 0.08056416320800781, 0.08066681671142578, 0.08099001312255859, 0.08953311920166016, 0.08120137786865235, 0.08082128143310546, 0.080321533203125, 0.08016041564941406, 0.0801119384765625, 0.08021987152099609, 0.07999862670898437, 0.08003062438964843, 0.0802240982055664, 0.08016780853271484, 0.08023737335205078, 0.08005990600585937, 0.0800693130493164, 0.08032041931152344, 0.08048758697509766, 0.08041388702392578, 0.08037350463867188, 0.08030111694335937, 0.08037677001953125, 0.0803051528930664, 0.08025395202636719, 0.08033484649658203, 0.08034406280517578, 0.08022029113769531, 0.08024269104003906, 0.08023129272460937, 0.08045308685302735, 0.08051136016845703, 0.08024285125732422, 0.08012518310546875, 0.08007737731933594, 0.0804335708618164, 0.08015980529785156, 0.08085574340820313, 0.08039222717285156, 0.08061542510986328, 0.08023638153076172, 0.08038006591796874, 0.0802344970703125, 0.08038947296142578, 0.0804985580444336, 0.08063046264648438, 0.0803960952758789, 0.08053113555908203, 0.08029039764404297, 0.08032428741455078, 0.08015904235839844, 0.08013823699951172, 0.08018745422363281, 0.0802581787109375, 0.08025727844238281, 0.08043958282470703, 0.08024502563476563, 0.08035123443603516, 0.08013622283935547, 0.08016687774658203, 0.08026435089111328, 0.08034185791015624, 0.08026924896240234, 0.08022022247314453, 0.08010377502441406, 0.08033859252929687, 0.0895567398071289, 0.08123212432861328, 0.08069097900390625, 0.08018466949462891, 0.0806143341064453, 0.08128505706787109, 0.08010956573486328, 0.07977574157714844, 0.07983219146728515, 0.08063065338134766, 0.07992115020751953, 0.07974864196777344, 0.07966515350341796, 0.07998694610595704, 0.08002365112304688, 0.07986502075195312, 0.07980707550048828, 0.0797003173828125, 0.07987535858154297, 0.07989462280273438, 0.07996018981933593, 0.0800588150024414, 0.08042201232910157, 0.08015353393554688, 0.08005760192871093, 0.08003151702880859, 0.08127919769287109, 0.080853759765625, 0.08044953918457032, 0.08042240142822266, 0.08043981170654296, 0.08052326202392578, 0.08047984313964844, 0.08029634857177734, 0.08052326202392578, 0.0803737564086914, 0.08034697723388672, 0.08033500671386719, 0.08039218902587891, 0.08094105529785156, 0.08022835540771485, 0.08011177825927734, 0.080174560546875, 0.08032498931884766, 0.08032624053955079, 0.08021177673339844, 0.080399169921875, 0.08054688262939454, 0.0804065933227539, 0.08044528198242187, 0.08055270385742187, 0.08048646545410157, 0.08083251190185547, 0.0805351333618164, 0.08052982330322266, 0.0804805450439453, 0.08047382354736328, 0.08049868774414062, 0.08050019073486328, 0.08057001495361328, 0.08057683563232422, 0.08054637145996094, 0.08039833831787109, 0.08946431732177734, 0.0810582046508789, 0.0806995849609375, 0.08015248107910156, 0.08009318542480469, 0.08024269104003906, 0.08025638580322265, 0.08008972930908204, 0.08006050872802735, 0.08045161437988281, 0.08009667205810547, 0.08019107055664063, 0.08011804962158203, 0.08008688354492187, 0.08029261016845703, 0.08027855682373047, 0.08020476531982422, 0.08001126098632813, 0.08032608032226562, 0.08008966064453125, 0.08016316986083985, 0.0799764175415039, 0.07994528198242187, 0.07992345428466797, 0.08021004486083984, 0.08018099212646485, 0.08035919952392578, 0.0804838104248047, 0.08006630706787109, 0.08011641693115235, 0.08006912231445312, 0.07999091339111328, 0.08035295867919921, 0.08009315490722656, 0.080200927734375, 0.08012995147705078, 0.08018326568603516, 0.08016316986083985, 0.08019004821777344, 0.0804491195678711, 0.0803309097290039, 0.08025142669677734, 0.08028166198730469, 0.08009693145751953, 0.08006886291503906, 0.08003533172607422, 0.08035763549804688, 0.08035456085205078, 0.08053622436523437, 0.08032617950439454, 0.08042345428466798, 0.08041474914550781, 0.08107008361816406, 0.08045478057861329, 0.08317632293701172, 0.08141414642333984, 0.08135884857177735, 0.08120259094238282, 0.08059964752197266, 0.08106598663330078, 0.0810533447265625, 0.08074082946777343, 0.08144384002685547, 0.090378173828125, 0.08283757019042969, 0.08203059387207032, 0.08073958587646485, 0.080408447265625, 0.08045359802246094, 0.08052035522460937, 0.0804615707397461, 0.08041203308105468, 0.08048214721679688, 0.08055683135986329, 0.08042082977294922, 0.08121887969970704, 0.08052809906005859, 0.0805292510986328, 0.0803034896850586, 0.08064694213867188, 0.08049254608154296, 0.08037785339355469, 0.08048255920410156, 0.08060288238525391, 0.08052941131591797, 0.08046793365478516, 0.08043318176269532, 0.08033507537841797, 0.08034457397460938, 0.08048668670654296, 0.08040038299560547, 0.08040067291259766, 0.08081356811523438, 0.0803435516357422, 0.08037923431396485, 0.08034137725830078, 0.08038604736328125, 0.08088371276855469, 0.08138137817382812, 0.08087312316894531, 0.08049494171142578, 0.080650146484375, 0.08063190460205077, 0.08073757171630859, 0.08102515411376954, 0.08142256164550782, 0.08096310424804687, 0.08061424255371094, 0.08068422698974609, 0.08041104125976563, 0.08041923522949218, 0.08067657470703125, 0.08053379058837891, 0.08066252899169922, 0.08090009307861327, 0.0807710723876953, 0.08082434844970703, 0.08076873779296875, 0.08072013092041015, 0.08081126403808593, 0.0811773452758789, 0.08148502349853516, 0.08098281860351562, 0.08086732482910156, 0.0816978530883789, 0.08106060791015625, 0.0896945571899414, 0.08142342376708984, 0.0807514877319336, 0.08054083251953124, 0.08027228546142579, 0.08027983856201172, 0.08027954864501953, 0.08024755096435547, 0.08020381164550781, 0.08001808166503906, 0.08018358612060547, 0.08026316833496094, 0.08022342681884766, 0.08033567810058594, 0.08031590270996093, 0.08035951995849609, 0.08041433715820312, 0.08018428802490235, 0.08023238372802734, 0.08023455810546876, 0.08016671752929687, 0.08036147308349609, 0.08050460815429687, 0.08042313385009765, 0.0803594207763672, 0.08042908477783203, 0.08053068542480468, 0.08039875030517578, 0.08101856231689453, 0.08116083526611328, 0.08102861022949219, 0.08049427032470703, 0.08049542236328125, 0.0815588150024414, 0.08049088287353516, 0.08027760314941407, 0.0802265625, 0.08049203491210938, 0.0808852767944336, 0.08058159637451172, 0.08076227569580079, 0.08063139343261719, 0.0804505615234375, 0.08026035308837891, 0.0802455062866211, 0.08033484649658203, 0.0811769256591797, 0.08074575805664062, 0.08055232238769532, 0.0806904296875, 0.08089622497558593, 0.08176399993896484, 0.08082316589355469, 0.08109670257568359, 0.08083251190185547, 0.08066252899169922, 0.0807701416015625, 0.08053043365478516, 0.08062761688232421, 0.0805580825805664, 0.08054598236083985, 0.08053126525878906, 0.08062684631347657, 0.09090048217773437, 0.08158182525634766, 0.08073651123046875, 0.08075878143310547, 0.08015872192382813, 0.08003926086425782, 0.08080633544921875, 0.08017292785644531, 0.0803147201538086, 0.0804823989868164, 0.0804494400024414, 0.08045069122314454, 0.08045452880859374, 0.08038329315185547, 0.08044409942626952, 0.080572509765625, 0.08088326263427735, 0.08036937713623046, 0.08049282836914062, 0.08039820861816406, 0.08052169799804687, 0.08041062164306641, 0.08034015655517578, 0.0806487045288086, 0.081004638671875, 0.08083849334716797, 0.08065184020996094, 0.08063369750976562, 0.08087615966796875, 0.08084310150146484, 0.08062566375732422, 0.08072803497314453, 0.081328125, 0.08127279663085937, 0.08091651153564453, 0.08076255798339843, 0.08070178985595704, 0.08091648101806641, 0.08089600372314452, 0.0806924819946289, 0.08068924713134766, 0.08057923126220704, 0.08056960296630859, 0.08063897705078125, 0.08069491577148437, 0.08060070037841797, 0.08072243499755859, 0.08055398559570312, 0.08089190673828126, 0.08144438171386718, 0.08105622100830077, 0.08100582122802734, 0.0812938232421875, 0.08092684936523438, 0.080595458984375, 0.08051673889160156, 0.08055193328857421, 0.08052531433105468, 0.08054956817626953, 0.08066079711914062, 0.08066047668457031, 0.08152019500732421, 0.08091910552978515, 0.09030242919921876, 0.08140595245361328, 0.08128921508789062, 0.08094310760498047, 0.081328125, 0.08047235107421875, 0.08038780975341797, 0.08071778869628907, 0.08046546936035157, 0.08031427001953124, 0.08027603149414063, 0.08050482940673828, 0.08063999938964844, 0.08068300628662109, 0.08058060455322266, 0.0806924819946289, 0.08088038635253907, 0.08060108947753906, 0.08068505859375, 0.08054771423339843, 0.08068828582763672, 0.08092301177978516, 0.08082697296142578, 0.08099174499511719, 0.08137779235839844, 0.08247296142578125, 0.08115977478027343, 0.08071119689941406, 0.08100543975830078, 0.08141747283935546, 0.08105856323242187, 0.08057453155517579, 0.08073516845703126, 0.08058573150634765, 0.08105577850341797, 0.08132809448242187, 0.08081574249267579, 0.0816316146850586, 0.08119910430908203, 0.08091033935546875, 0.0810758056640625, 0.08165590667724609, 0.08105551910400391, 0.0808156509399414, 0.08074121856689453, 0.08082038116455079, 0.08124777221679688, 0.08138502502441407, 0.08095123291015625, 0.08128406524658204, 0.08149795532226563, 0.08098729705810546, 0.0809013442993164, 0.08098365020751953, 0.0815617904663086, 0.08097138977050782, 0.08112985229492188, 0.08142550659179687, 0.08098032379150391, 0.08111353302001953, 0.08134636688232422, 0.08106835174560546, 0.08128102111816406, 0.08985260772705078, 0.0812743377685547, 0.08074089813232421, 0.08034019470214844, 0.08032937622070313, 0.08044898986816407, 0.08056678771972656, 0.08041069030761719, 0.08044454193115234, 0.08057545471191406, 0.08056832122802735, 0.08053533172607422, 0.08044156646728516, 0.0804853744506836, 0.080681884765625, 0.08044963073730468, 0.08044748687744141, 0.08051097869873047, 0.08046150207519531, 0.08042550659179687, 0.08037709045410156, 0.08048233795166015, 0.08057878112792968, 0.08049600219726563, 0.08058358764648438, 0.08118271636962891, 0.08057379150390626, 0.08061814117431641, 0.08063795471191407, 0.08059613037109375, 0.0804435806274414, 0.08058294677734375, 0.080512451171875, 0.08052774047851563, 0.08067046356201171, 0.08074527740478515, 0.08081990051269532, 0.08094969940185547, 0.08123104095458984, 0.0810340805053711, 0.0808194580078125, 0.08081283569335937, 0.08083993530273438, 0.08081590270996093, 0.08115894317626954, 0.08141619110107422, 0.08089164733886718, 0.08081871795654297, 0.08080355072021485, 0.08099635314941406, 0.08089190673828126, 0.0813017578125, 0.08140473937988281, 0.08192479705810547, 0.08108831787109375, 0.08078505706787109, 0.08127686309814453, 0.0809603500366211, 0.08097586822509766, 0.0810304946899414, 0.08157225799560547, 0.08118287658691406, 0.08090742492675781]",tokens/s,12.38418416204443,, @@ -3524,7 +3524,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 82181 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 71037 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,internlm,internlm/internlm-20b,internlm/internlm-20b,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -3602,7 +3602,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 34.12 MiB is free. Process 87085 has 14.71 GiB memory in use. Of the allocated memory 14.56 GiB is allocated by PyTorch, and 28.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 34.12 MiB is free. Process 76102 has 14.71 GiB memory in use. Of the allocated memory 14.56 GiB is allocated by PyTorch, and 28.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,codegen,Salesforce/codegen-16B-nl,Salesforce/codegen-16B-nl,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.223-212.873.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.44.2,,0.34.2,,,,1.21.4,,,,0.12.0,,,True,MB,1198.444544,10582.09792,0.0,10179.575808,10067.3536,s,1,22.452044921875,22.452044921875,0.0,22.452044921875,22.452044921875,22.452044921875,22.452044921875,[22.452044921875],,kWh,0.0004358892054416667,4.807473573073978e-05,0.00014503817158600246,0.000629002112758409,,MB,1408.946176,12521.96352,0.0,12106.858496,11264.997888,s,10,22.025080078125,2.2025080078125,0.012151026773434104,2.2043836669921877,2.2159740234375,2.216520703125,2.216958046875,"[2.176212158203125, 2.1932548828125, 2.192836669921875, 2.19731787109375, 2.204373779296875, 2.2043935546875, 2.210161865234375, 2.2170673828125, 2.2158525390625, 2.213609375]",tokens/s,116.2311324598795,kWh,6.432398427874924e-05,7.094660104198045e-06,4.286047873280005e-05,0.00011427912311574734,tokens/kWh,2240129.194382346,MB,1433.993216,12526.157824,0.0,12108.955648,11265.000448,s,10,108.8452109375,10.88452109375,0.02712600890432365,10.89675048828125,10.912278515625001,10.91612021484375,10.919193574218749,"[10.8512158203125, 10.8384873046875, 10.850037109375, 10.89648828125, 10.87683203125, 10.9199619140625, 10.90212109375, 10.9016298828125, 10.8970126953125, 10.9114248046875]",tokens/s,5.788036006120217,kWh,0.0003188599458741688,3.517261090855641e-05,0.00021167561378480083,0.0005657081705675261,tokens/kWh,111364.84017333097,,s,630,108.84104246520995,0.17276355946858724,0.0019008705304642802,0.1726689453125,0.17437670898437502,0.17481485061645508,0.18287327545166018,"[0.18123143005371095, 0.16884124755859375, 0.1698864288330078, 0.1694535675048828, 0.16980400085449218, 0.17391766357421876, 0.17219821166992189, 0.17003506469726562, 0.16972137451171876, 0.17039984130859376, 0.17010252380371094, 0.17304460144042969, 0.17199090576171874, 0.17009458923339843, 0.1698870086669922, 0.17112351989746094, 0.17137481689453124, 0.17271737670898438, 0.17186444091796876, 0.17133506774902343, 0.16930400085449218, 0.1692410888671875, 0.17280184936523438, 0.17293142700195313, 0.17225100708007812, 0.17232920837402343, 0.17161322021484374, 0.17191542053222655, 0.17265061950683594, 0.17347254943847656, 0.17347552490234375, 0.17267030334472655, 0.17283917236328125, 0.17308511352539063, 0.17524365234375, 0.1728419189453125, 0.17325059509277344, 0.17280242919921876, 0.171177734375, 0.17142640686035157, 0.17394688415527343, 0.172600830078125, 0.17188832092285156, 0.17245472717285157, 0.17216511535644533, 0.1722341766357422, 0.17210595703125, 0.17246421813964843, 0.17310508728027343, 0.17247673034667968, 0.17183949279785157, 0.17445401000976563, 0.17218832397460937, 0.17191535949707032, 0.1733632049560547, 0.17340937805175782, 0.17244866943359374, 0.1732052459716797, 0.173459716796875, 0.17327513122558594, 0.17274819946289063, 0.1745432586669922, 0.17215731811523438, 0.18242326354980468, 0.16967298889160157, 0.1696485137939453, 0.16967868041992187, 0.1691374053955078, 0.17443746948242186, 0.17374365234375, 0.16972840881347656, 0.17142098999023436, 0.170411865234375, 0.17159036254882812, 0.17380982971191405, 0.1727519073486328, 0.17074374389648436, 0.16964697265625, 0.17044085693359376, 0.1725028839111328, 0.17299148559570313, 0.17232371520996093, 0.17219766235351563, 0.1706724853515625, 0.17059840393066406, 0.17255833435058593, 0.17263597106933593, 0.1717220458984375, 0.17154905700683593, 0.17078057861328125, 0.17217391967773438, 0.17323426818847656, 0.17258486938476564, 0.1723408660888672, 0.17135446166992188, 0.17094044494628907, 0.17189718627929687, 0.17279971313476564, 0.17159776306152344, 0.17197900390625, 0.17137747192382813, 0.17130812072753906, 0.1721180419921875, 0.17249261474609376, 0.17197666931152344, 0.1719207000732422, 0.17092169189453124, 0.1718691864013672, 0.17255558776855467, 0.17226617431640626, 0.17223785400390626, 0.17203724670410156, 0.1714683837890625, 0.1731394500732422, 0.17237068176269532, 0.17178419494628908, 0.17258610534667967, 0.17182768249511718, 0.17225935363769532, 0.1725068817138672, 0.17211199951171874, 0.17232333374023437, 0.17258918762207032, 0.17306771850585936, 0.17303103637695313, 0.17321994018554687, 0.1830570831298828, 0.1692051239013672, 0.16866543579101562, 0.1706273956298828, 0.1694146270751953, 0.17474745178222656, 0.17273365783691405, 0.17148588562011718, 0.17113731384277345, 0.17039535522460938, 0.17113661193847657, 0.17392230224609376, 0.17260809326171875, 0.17085244750976564, 0.17160018920898437, 0.17135789489746095, 0.1725560302734375, 0.17246438598632813, 0.17185702514648438, 0.1710101776123047, 0.1706483154296875, 0.16993913269042968, 0.1722360382080078, 0.17167129516601562, 0.16999494934082032, 0.17109619140625, 0.17114070129394532, 0.17120297241210938, 0.17269132995605468, 0.1722574005126953, 0.17241702270507814, 0.17235968017578124, 0.17108522033691406, 0.1728287353515625, 0.17208192443847656, 0.17285916137695312, 0.17331199645996093, 0.17191526794433593, 0.1718970947265625, 0.17297113037109374, 0.1732010498046875, 0.17220851135253906, 0.17206947326660157, 0.17180467224121093, 0.17271192932128906, 0.1730908203125, 0.17206019592285157, 0.1729192657470703, 0.17183538818359376, 0.17184259033203125, 0.17311155700683595, 0.1729666290283203, 0.1727651824951172, 0.17352703857421875, 0.17219564819335936, 0.17253190612792968, 0.17262339782714844, 0.1728409881591797, 0.1727229461669922, 0.1740100555419922, 0.17296115112304689, 0.1731402587890625, 0.17305349731445313, 0.1853082580566406, 0.17081394958496093, 0.1701604461669922, 0.17026792907714844, 0.1698310089111328, 0.17575343322753906, 0.17244979858398438, 0.1716071319580078, 0.17046211242675782, 0.16919349670410155, 0.17160124206542968, 0.17533116149902345, 0.1725570831298828, 0.17214463806152344, 0.170446044921875, 0.17196112060546875, 0.17337522888183593, 0.17317298889160157, 0.17373933410644532, 0.17158624267578124, 0.1710672607421875, 0.17147421264648438, 0.17332940673828126, 0.17405039978027342, 0.17357632446289062, 0.17208604431152344, 0.17202128601074218, 0.17228367614746093, 0.17370780944824218, 0.17391001892089844, 0.17267916870117186, 0.17141123962402344, 0.17189222717285157, 0.17369906616210937, 0.17362757873535156, 0.17408883666992186, 0.1724144287109375, 0.17272047424316406, 0.17252362060546875, 0.1740597381591797, 0.173993408203125, 0.17379568481445312, 0.17256375122070314, 0.17261366271972656, 0.17380630493164062, 0.17351644897460938, 0.1744878692626953, 0.17404937744140625, 0.1721383056640625, 0.17273199462890626, 0.17382406616210938, 0.17305751037597655, 0.1742834930419922, 0.1718105010986328, 0.1720530548095703, 0.1734185028076172, 0.17295155334472656, 0.17447663879394532, 0.17287799072265625, 0.1732861785888672, 0.17324617004394532, 0.17430528259277345, 0.17438082885742187, 0.1816338806152344, 0.17068850708007813, 0.17040380859375, 0.168974365234375, 0.16962258911132813, 0.17552275085449218, 0.17239840698242187, 0.17102464294433595, 0.16939328002929688, 0.17024287414550782, 0.17165721130371095, 0.17437625122070313, 0.1731826934814453, 0.17163885498046874, 0.16991317749023438, 0.17148956298828125, 0.17299139404296876, 0.17458425903320313, 0.17228147888183593, 0.1720811767578125, 0.1713766326904297, 0.17118841552734376, 0.17448953247070312, 0.17361683654785157, 0.17207107543945313, 0.17141993713378906, 0.17140296936035157, 0.17253631591796875, 0.17350440979003906, 0.1725314636230469, 0.17216770935058595, 0.1728738555908203, 0.1724805145263672, 0.17308876037597656, 0.1738236083984375, 0.1726297607421875, 0.17188832092285156, 0.17124659729003905, 0.17338975524902345, 0.17315411376953124, 0.1729438018798828, 0.17337628173828126, 0.17332733154296875, 0.17302236938476562, 0.17478732299804686, 0.17393894958496095, 0.17353919982910157, 0.17146611022949218, 0.1722058868408203, 0.17481602478027344, 0.17557196044921874, 0.1734051818847656, 0.17318911743164062, 0.171863525390625, 0.1715328369140625, 0.17295677185058594, 0.17288841247558595, 0.17217088317871093, 0.17266578674316407, 0.1724559326171875, 0.17270364379882813, 0.1736846466064453, 0.17292921447753906, 0.18456454467773437, 0.16991439819335938, 0.17091941833496094, 0.17065621948242188, 0.17115213012695313, 0.17795989990234376, 0.17275830078125, 0.1715494384765625, 0.17105340576171876, 0.17086842346191405, 0.1720343322753906, 0.17609695434570313, 0.17358233642578125, 0.17158575439453125, 0.17124957275390626, 0.17145440673828125, 0.17333418273925782, 0.17373843383789062, 0.17346278381347657, 0.1719727020263672, 0.1712004089355469, 0.17220077514648438, 0.17323114013671875, 0.17415245056152343, 0.17446092224121093, 0.17134547424316407, 0.171078369140625, 0.1731715850830078, 0.1748017578125, 0.17316249084472657, 0.17172476196289063, 0.17259933471679687, 0.175177734375, 0.1736212463378906, 0.17459814453125, 0.17373114013671875, 0.17222930908203124, 0.17304495239257814, 0.17266758728027343, 0.17354351806640625, 0.1747538604736328, 0.1733031005859375, 0.17366464233398438, 0.173523193359375, 0.1747250213623047, 0.17485340881347655, 0.17385072326660156, 0.17280809020996094, 0.17280029296875, 0.17376045227050782, 0.172775390625, 0.17453097534179687, 0.17343446350097655, 0.17403692626953124, 0.17207366943359376, 0.17320755004882812, 0.17422726440429687, 0.17457171630859375, 0.1736681213378906, 0.17449168395996092, 0.17336058044433594, 0.17424050903320312, 0.1752241668701172, 0.18394454956054687, 0.17117893981933593, 0.1701965789794922, 0.17147740173339843, 0.17156256103515624, 0.17700703430175782, 0.17229580688476562, 0.172044677734375, 0.17031318664550782, 0.17063392639160158, 0.17210330200195312, 0.17562646484375, 0.17343283081054686, 0.17244979858398438, 0.17172402954101562, 0.17108822631835938, 0.1733470458984375, 0.17425836181640625, 0.17243341064453124, 0.17299871826171875, 0.17084614562988282, 0.17236941528320313, 0.17309336853027343, 0.17435647583007813, 0.17219110107421876, 0.17165171813964844, 0.17160911560058595, 0.17283193969726562, 0.17345747375488282, 0.17379913330078126, 0.17272146606445313, 0.17248941040039062, 0.17156626892089843, 0.17358447265625, 0.17360520935058593, 0.17359295654296875, 0.1722076416015625, 0.17221372985839845, 0.17318748474121093, 0.17318669128417968, 0.17482406616210938, 0.17306227111816405, 0.17286697387695313, 0.17195606994628906, 0.17295663452148438, 0.17385177612304686, 0.1730487060546875, 0.173295166015625, 0.1723410186767578, 0.17246412658691407, 0.17391081237792969, 0.1736143341064453, 0.1735350036621094, 0.17268234252929687, 0.17246797180175782, 0.173486083984375, 0.17351884460449218, 0.17297383117675783, 0.1731832275390625, 0.1744066619873047, 0.17315090942382813, 0.17470086669921875, 0.1747162628173828, 0.1832654724121094, 0.1705184326171875, 0.17141349792480468, 0.17093597412109374, 0.17148287963867187, 0.17794422912597657, 0.17450694274902342, 0.17143807983398437, 0.1714456024169922, 0.170990234375, 0.17288729858398438, 0.17481341552734375, 0.17339447021484375, 0.1720908203125, 0.17165132141113282, 0.17121542358398437, 0.1735615997314453, 0.17446269226074218, 0.17293061828613282, 0.17211056518554688, 0.17217741394042968, 0.17171229553222656, 0.17345472717285157, 0.17353610229492186, 0.17181007385253907, 0.1715063934326172, 0.1713880615234375, 0.17226634216308595, 0.17299971008300782, 0.17357513427734375, 0.17216677856445312, 0.17032435607910157, 0.17277714538574218, 0.17359698486328126, 0.17373423767089843, 0.17237571716308595, 0.17242489624023438, 0.17248605346679688, 0.17332083129882814, 0.17379971313476564, 0.1732515869140625, 0.17302630615234374, 0.1721750030517578, 0.1730194854736328, 0.17428889465332031, 0.17292288208007814, 0.17330390930175782, 0.17245919799804688, 0.1719385223388672, 0.17332345581054687, 0.1730806121826172, 0.17302793884277343, 0.17293125915527344, 0.1738936309814453, 0.1735925750732422, 0.17398988342285157, 0.17308502197265624, 0.17386691284179687, 0.17336090087890624, 0.17472671508789062, 0.17318547058105468, 0.17415577697753906, 0.17411891174316407, 0.18339814758300782, 0.171156005859375, 0.17198208618164063, 0.17031011962890624, 0.17186611938476562, 0.17667071533203124, 0.17309500122070312, 0.17112086486816405, 0.1704117126464844, 0.17079855346679687, 0.17282928466796876, 0.17576275634765626, 0.1725603485107422, 0.17109674072265624, 0.17124114990234374, 0.1716862030029297, 0.173764892578125, 0.17356361389160158, 0.1734261474609375, 0.17196479797363282, 0.17200682067871093, 0.17286607360839842, 0.17440403747558594, 0.17344284057617188, 0.17232691955566407, 0.17197261047363283, 0.17152204895019532, 0.17380787658691407, 0.173586181640625, 0.17397555541992188, 0.17260365295410157, 0.1713026580810547, 0.1730723876953125, 0.17270182800292969, 0.17265037536621095, 0.17329971313476564, 0.17364157104492187, 0.17218576049804687, 0.17244309997558593, 0.1741923522949219, 0.1739350128173828, 0.17252700805664062, 0.17270271301269532, 0.17262310791015625, 0.17229696655273438, 0.174339111328125, 0.17264906311035155, 0.17233552551269532, 0.17112451171875, 0.17222496032714843, 0.17357183837890625, 0.17446826171875, 0.17273942565917969, 0.1720863037109375, 0.1717884521484375, 0.17247894287109375, 0.17373219299316406, 0.17492985534667968, 0.17321171569824217, 0.1731995849609375, 0.17289193725585938, 0.17460415649414063, 0.17341786193847655, 0.1835221405029297, 0.1702010955810547, 0.17080099487304687, 0.17115335083007813, 0.17137277221679686, 0.17748127746582032, 0.17407955932617186, 0.1716413116455078, 0.17181715393066407, 0.17105123901367186, 0.17231820678710938, 0.17534413146972655, 0.17350198364257813, 0.17169378662109375, 0.17116236877441407, 0.17148931884765625, 0.17276719665527343, 0.17443020629882813, 0.1743953857421875, 0.17137586975097657, 0.17075177001953126, 0.17271609497070312, 0.17337641906738283, 0.17424998474121095, 0.17324832153320313, 0.17139651489257812, 0.17221209716796876, 0.1727804412841797, 0.17430908203125, 0.17395321655273438, 0.17265040588378905, 0.17186834716796875, 0.1722326965332031, 0.17415350341796876, 0.17486006164550782, 0.17386131286621093, 0.17311308288574218, 0.1726650848388672, 0.17331724548339844, 0.1742872314453125, 0.17370777893066405, 0.17301324462890624, 0.17346450805664063, 0.17224739074707032, 0.17319679260253906, 0.17313912963867187, 0.1729954833984375, 0.17319786071777343, 0.1719134063720703, 0.17226908874511718, 0.17374855041503906, 0.17503421020507812, 0.17401084899902344, 0.17349221801757814, 0.17274674987792968, 0.1738079376220703, 0.17379884338378906, 0.1748074188232422, 0.1738648376464844, 0.17301609802246093, 0.17296701049804689, 0.17279945373535155, 0.17417042541503908]",tokens/s,5.788257680473556,, @@ -3649,7 +3649,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 302.12 MiB is free. Process 154885 has 14.44 GiB memory in use. Of the allocated memory 14.20 GiB is allocated by PyTorch, and 132.89 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 302.12 MiB is free. Process 160809 has 14.44 GiB memory in use. Of the allocated memory 14.20 GiB is allocated by PyTorch, and 132.89 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,huggyllama/llama-30b,huggyllama/llama-30b,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -3696,7 +3696,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 54.12 MiB is free. Process 150142 has 14.69 GiB memory in use. Of the allocated memory 14.37 GiB is allocated by PyTorch, and 203.94 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 54.12 MiB is free. Process 155981 has 14.69 GiB memory in use. Of the allocated memory 14.37 GiB is allocated by PyTorch, and 203.94 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-13b-hf,meta-llama/Llama-2-13b-hf,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.224-212.876.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.0,,0.34.2,,,,1.22.0,,,,0.12.0,,,True,MB,825.176064,8535.277568,0.0,8132.755456,7824.681472,s,1,19.355048828125,19.355048828125,0.0,19.355048828125,19.355048828125,19.355048828125,19.355048828125,[19.355048828125],,kWh,0.000357421612099976,3.941356905418048e-05,0.00011472342511199829,0.0005115586062661547,,MB,1362.427904,9539.813376,0.0,9124.708352,8500.632064,s,10,17.740904541015624,1.7740904541015623,0.008706226093104712,1.7761414184570312,1.7815432250976564,1.7821738952636719,1.7826784313964845,"[1.7506591796875, 1.770615234375, 1.7712784423828125, 1.774149658203125, 1.7746822509765625, 1.77799755859375, 1.7776005859375, 1.7797139892578124, 1.7828045654296876, 1.781403076171875]",tokens/s,144.2992939892932,kWh,5.1742666187920645e-05,5.706854829461947e-06,3.4431388656198936e-05,9.188090967358153e-05,tokens/kWh,2786215.3401557747,MB,1399.656448,9544.00768,0.0,9126.805504,8500.634624,s,10,83.37062890625,8.337062890625,0.018678424742809827,8.340683105468749,8.35447490234375,8.354972802734375,8.355371123046876,"[8.294720703125, 8.3147724609375, 8.33323828125, 8.333140625, 8.334890625, 8.3464755859375, 8.3532841796875, 8.3543642578125, 8.350271484375, 8.355470703125]",tokens/s,7.556618059202035,kWh,0.00024398087841749578,2.6913000317783666e-05,0.00016191376841980242,0.00043280764715508186,tokens/kWh,145561.1988700054,,s,630,83.36781452941898,0.13232986433241103,0.0018008641998273776,0.132398681640625,0.1334996597290039,0.1338594841003418,0.14334174850463868,"[0.14481919860839843, 0.13183203125, 0.13035484313964843, 0.13028848266601561, 0.13021385192871093, 0.1302890167236328, 0.13022682189941406, 0.13251763916015624, 0.1320978240966797, 0.13074070739746094, 0.13034495544433594, 0.13034214782714842, 0.13036416625976563, 0.13039599609375, 0.13181149291992186, 0.13171452331542968, 0.13223939514160157, 0.13073373413085937, 0.13047042846679688, 0.13036976623535157, 0.13046588134765624, 0.13146316528320312, 0.1329747772216797, 0.13314186096191405, 0.1311072998046875, 0.1306929931640625, 0.13060957336425782, 0.13063343811035155, 0.1307667236328125, 0.13167015075683594, 0.13287628173828125, 0.13268992614746095, 0.1311270751953125, 0.13076502990722655, 0.13065420532226563, 0.13071286010742186, 0.13123606872558594, 0.132789794921875, 0.13278717041015625, 0.13275712585449218, 0.13237901306152344, 0.1311436767578125, 0.1308037109375, 0.13079756164550782, 0.13156338500976564, 0.1324701385498047, 0.13255552673339843, 0.1323489532470703, 0.1310866241455078, 0.13267628479003907, 0.13136224365234375, 0.13083705139160157, 0.1315392303466797, 0.13240031433105467, 0.1327222137451172, 0.13117543029785156, 0.13212672424316407, 0.13261415100097657, 0.1312027587890625, 0.13106434631347658, 0.13149366760253905, 0.1326878662109375, 0.13134025573730468, 0.14261453247070313, 0.13204074096679688, 0.13062696838378907, 0.13035891723632811, 0.13025173950195312, 0.13035282897949219, 0.13036166381835937, 0.13219020080566407, 0.13318553161621094, 0.13207347106933592, 0.13052281188964843, 0.13042105102539062, 0.1304239044189453, 0.13115676879882812, 0.130648193359375, 0.13253439331054687, 0.13253004455566406, 0.132495361328125, 0.13174732971191405, 0.130861572265625, 0.1305006103515625, 0.1311846466064453, 0.13174745178222655, 0.1330873260498047, 0.1328926696777344, 0.1320880889892578, 0.131557373046875, 0.13202348327636718, 0.1306337890625, 0.13061570739746095, 0.13254876708984376, 0.13292294311523437, 0.13267750549316407, 0.1319058837890625, 0.13119679260253905, 0.13080426025390626, 0.13111465454101562, 0.13256297302246095, 0.13212255859375, 0.13303176879882814, 0.13228419494628907, 0.13244288635253906, 0.1311928253173828, 0.13123379516601563, 0.13253330993652343, 0.13170375061035156, 0.13366886901855468, 0.13290701293945312, 0.13170072937011718, 0.1315752716064453, 0.1314343719482422, 0.13278880310058594, 0.13131085205078125, 0.13277679443359375, 0.13312205505371094, 0.132710205078125, 0.13129344177246094, 0.13107574462890625, 0.13160791015625, 0.1327277374267578, 0.1313278045654297, 0.1329189453125, 0.133555908203125, 0.1440358428955078, 0.13177778625488282, 0.130403076171875, 0.13034291076660157, 0.13029478454589843, 0.13028582763671875, 0.13075942993164064, 0.13374642944335938, 0.13345928955078126, 0.13242051696777343, 0.1309450225830078, 0.13046170043945313, 0.13038584899902345, 0.130418212890625, 0.133093017578125, 0.1321355895996094, 0.1328450927734375, 0.13297232055664063, 0.13147433471679687, 0.13058047485351562, 0.13047398376464844, 0.13190757751464843, 0.1331793975830078, 0.13338021850585938, 0.13299848937988282, 0.13253237915039062, 0.13205337524414062, 0.13074771118164064, 0.13115461730957031, 0.1325322265625, 0.13195797729492187, 0.13312643432617188, 0.13273344421386718, 0.13229884338378906, 0.13083177185058595, 0.13130607604980468, 0.13257081604003906, 0.13241981506347655, 0.1325240936279297, 0.13199559020996093, 0.13318553161621094, 0.1319710693359375, 0.13196697998046875, 0.1323970489501953, 0.13190757751464843, 0.13334323120117186, 0.13215461730957032, 0.13297311401367187, 0.13138691711425782, 0.13157398986816407, 0.1334291229248047, 0.1321922607421875, 0.13272543334960937, 0.13210202026367188, 0.1327389373779297, 0.13337408447265625, 0.13213603210449218, 0.13184912109375, 0.13263258361816407, 0.13255795288085936, 0.13331546020507812, 0.1322079620361328, 0.13329679870605468, 0.14353855895996093, 0.13185565185546874, 0.13040211486816405, 0.1302349090576172, 0.13026345825195312, 0.13043096923828126, 0.1306378173828125, 0.13441261291503906, 0.1324397430419922, 0.13136802673339842, 0.13200271606445313, 0.13061734008789064, 0.13037930297851563, 0.13051747131347657, 0.1333387145996094, 0.13267808532714845, 0.13269804382324218, 0.13187893676757811, 0.13093843078613282, 0.13194216918945312, 0.13078099060058593, 0.13164627075195312, 0.13247286987304688, 0.13285580444335937, 0.13196444702148438, 0.13237837219238283, 0.13172808837890626, 0.13232333374023436, 0.13102694702148437, 0.13251171875, 0.1326790771484375, 0.13268553161621094, 0.13198019409179687, 0.1323225860595703, 0.13160450744628907, 0.1322236785888672, 0.13190931701660155, 0.1331838073730469, 0.13266738891601562, 0.13291427612304688, 0.13123043823242186, 0.13266249084472656, 0.13146826171875, 0.13243807983398437, 0.132837158203125, 0.13228662109375, 0.13244825744628907, 0.13255885314941407, 0.13247283935546875, 0.13221478271484374, 0.13260418701171875, 0.13266102600097657, 0.13224748229980468, 0.1326796875, 0.13247897338867187, 0.13312205505371094, 0.1325875244140625, 0.1324400634765625, 0.13224453735351563, 0.1327687072753906, 0.1313420867919922, 0.13324029541015625, 0.13341981506347655, 0.14375071716308593, 0.131745849609375, 0.13029986572265626, 0.13022694396972656, 0.13069424438476562, 0.1302864990234375, 0.1308734130859375, 0.13475013732910157, 0.13320124816894532, 0.13155091857910156, 0.130396484375, 0.13034153747558594, 0.1304015350341797, 0.131046142578125, 0.13290447998046875, 0.1334639434814453, 0.13318328857421874, 0.13195138549804689, 0.13054754638671875, 0.13071990966796876, 0.13089955139160156, 0.13243801879882813, 0.1333784637451172, 0.13364134216308593, 0.13259190368652343, 0.13123440551757812, 0.1306419219970703, 0.13058026123046876, 0.1315919647216797, 0.13288479614257812, 0.1339412841796875, 0.13383485412597657, 0.13249317932128907, 0.131004638671875, 0.13064553833007814, 0.1308673858642578, 0.13256924438476564, 0.132589599609375, 0.134217529296875, 0.13354742431640626, 0.13216761779785155, 0.13082095336914062, 0.13067263793945313, 0.13148159790039063, 0.13284352111816405, 0.1340436553955078, 0.1339330596923828, 0.13243951416015626, 0.13126710510253906, 0.13100236511230468, 0.13315583801269532, 0.1316812744140625, 0.13301951599121092, 0.13399977111816405, 0.13328688049316406, 0.132421630859375, 0.13281893920898438, 0.13147544860839844, 0.13127433776855468, 0.13291529846191405, 0.133808349609375, 0.13252400207519532, 0.1335784912109375, 0.1441458282470703, 0.1318651885986328, 0.1305128936767578, 0.13071974182128906, 0.1303442840576172, 0.13047261047363282, 0.1316864013671875, 0.13605990600585938, 0.13276176452636718, 0.13115478515625, 0.13044876098632813, 0.13052359008789063, 0.1315569305419922, 0.1320597686767578, 0.1333719024658203, 0.13352326965332031, 0.13335494995117186, 0.1318943634033203, 0.130999267578125, 0.13095989990234375, 0.13105113220214842, 0.1337533721923828, 0.1331403503417969, 0.133378173828125, 0.13327769470214842, 0.13184819030761719, 0.13080166625976564, 0.13058216857910157, 0.13221270751953124, 0.13344309997558593, 0.13348077392578125, 0.13270252990722656, 0.13190969848632814, 0.13205279541015624, 0.13058493041992186, 0.13163880920410156, 0.13328022766113282, 0.13267575073242188, 0.13393898010253907, 0.13252204895019531, 0.13237657165527345, 0.13114169311523438, 0.13141191101074218, 0.13252198791503905, 0.13313221740722656, 0.13351651000976564, 0.1333830108642578, 0.13282322692871093, 0.13256617736816406, 0.1310847930908203, 0.13184159851074218, 0.13242556762695312, 0.133421630859375, 0.133714111328125, 0.13268988037109375, 0.13228402709960937, 0.13258384704589843, 0.13146258544921874, 0.13188359069824218, 0.13309327697753906, 0.1342710723876953, 0.1337012176513672, 0.1320800323486328, 0.14409318542480468, 0.13179017639160157, 0.130900634765625, 0.1304226531982422, 0.13033279418945312, 0.130844482421875, 0.13195896911621094, 0.1355404510498047, 0.1340429382324219, 0.1324349060058594, 0.1311129608154297, 0.1324073028564453, 0.13090815734863281, 0.13045555114746094, 0.13303366088867188, 0.13303021240234375, 0.13357437133789063, 0.13246287536621093, 0.13126559448242187, 0.13123004150390624, 0.13226194763183594, 0.131733154296875, 0.13336213684082032, 0.13290684509277345, 0.1335465545654297, 0.13193116760253906, 0.13067362976074218, 0.13252323913574218, 0.13232412719726563, 0.13258956909179687, 0.1325875244140625, 0.1326366424560547, 0.13308522033691406, 0.1322388153076172, 0.13105372619628905, 0.13253251647949219, 0.1317928924560547, 0.13326960754394532, 0.1327960662841797, 0.13261785888671876, 0.1319881896972656, 0.1327860107421875, 0.13305830383300782, 0.13140419006347656, 0.13260386657714843, 0.13275135803222657, 0.13258685302734374, 0.13331484985351563, 0.13295225524902343, 0.13344111633300781, 0.13205702209472656, 0.1326537628173828, 0.1323520050048828, 0.1325280303955078, 0.13367884826660156, 0.1328438720703125, 0.13287423706054688, 0.13226803588867186, 0.13336781311035156, 0.13225369262695313, 0.13277503967285156, 0.13271267700195313, 0.1334805145263672, 0.14332470703125, 0.13226646423339844, 0.13045298767089844, 0.13033625793457032, 0.13026611328125, 0.13026847839355468, 0.13154170227050782, 0.13490109252929688, 0.13394192504882813, 0.13215948486328125, 0.1310323181152344, 0.13043283081054688, 0.13052598571777344, 0.1311868438720703, 0.13342320251464843, 0.13351925659179686, 0.13354803466796875, 0.1322677459716797, 0.13192630004882813, 0.1306516876220703, 0.13124176025390624, 0.13301011657714842, 0.13379107666015624, 0.1337960662841797, 0.1325015106201172, 0.13270060729980468, 0.13125369262695313, 0.13086965942382814, 0.1322960662841797, 0.13368960571289062, 0.1327208709716797, 0.13265951538085938, 0.13352864074707033, 0.13258848571777343, 0.13104537963867188, 0.13188096618652342, 0.13348658752441406, 0.13327894592285155, 0.1328770294189453, 0.13345184326171874, 0.13323043823242187, 0.13160873413085938, 0.13168447875976563, 0.13323455810546875, 0.13259365844726562, 0.13265715026855468, 0.13347021484375, 0.1324419860839844, 0.13294534301757813, 0.13284556579589843, 0.13263296508789063, 0.13251820373535156, 0.1324541778564453, 0.13365382385253907, 0.13313526916503907, 0.1322904968261719, 0.1328661193847656, 0.13228851318359375, 0.13303094482421876, 0.13260694885253907, 0.13266943359375, 0.13404261779785157, 0.13257420349121093, 0.14285157775878907, 0.13190354919433595, 0.13045826721191406, 0.13031184387207032, 0.13021586608886718, 0.1303233642578125, 0.13216128540039063, 0.13490115356445312, 0.1339132843017578, 0.13231849670410156, 0.13092086791992188, 0.13066671752929687, 0.1303900146484375, 0.13136441040039062, 0.13384544372558593, 0.1337689208984375, 0.1333445739746094, 0.13176316833496093, 0.13090797424316405, 0.13046173095703126, 0.13111036682128907, 0.13374461364746093, 0.13283120727539063, 0.1333358154296875, 0.1328302459716797, 0.13141091918945313, 0.1305575714111328, 0.13117478942871094, 0.13311727905273438, 0.13298112487792968, 0.13306466674804687, 0.13256285095214843, 0.1326739501953125, 0.13302784729003905, 0.1312788543701172, 0.13151437377929687, 0.13221417236328126, 0.1331881561279297, 0.13291098022460937, 0.1324893798828125, 0.1328476104736328, 0.1330155487060547, 0.13228440856933593, 0.132071044921875, 0.1329256591796875, 0.1329701385498047, 0.13301737976074218, 0.13273980712890626, 0.13288406372070313, 0.1326508483886719, 0.13270201110839844, 0.13262448120117187, 0.13331248474121093, 0.13233126831054687, 0.13349778747558594, 0.13302784729003905, 0.1327941131591797, 0.13302195739746095, 0.13347430419921874, 0.1321448974609375, 0.133242431640625, 0.13285574340820314, 0.13273500061035157, 0.14334870910644532, 0.13180038452148438, 0.13031698608398437, 0.130322265625, 0.13017868041992187, 0.13030387878417968, 0.13165779113769532, 0.13593020629882813, 0.133791748046875, 0.13197747802734375, 0.1307269744873047, 0.1302947235107422, 0.13038796997070312, 0.13117666625976562, 0.13422569274902343, 0.1345797119140625, 0.13273458862304688, 0.13209481811523438, 0.1319239959716797, 0.13053543090820313, 0.1308847961425781, 0.13306553649902345, 0.13397811889648437, 0.13354803466796875, 0.13165977478027344, 0.13258956909179687, 0.13165951538085938, 0.13089190673828124, 0.1321985321044922, 0.13368634033203125, 0.13308615112304686, 0.13277162170410156, 0.13204092407226561, 0.13232102966308593, 0.13114393615722655, 0.13254197692871095, 0.13303855895996095, 0.1338709716796875, 0.13308787536621094, 0.1326796875, 0.13291651916503905, 0.13284835815429688, 0.1322434539794922, 0.13258656311035155, 0.1330410919189453, 0.1337507781982422, 0.13278207397460937, 0.13288447570800782, 0.13254579162597657, 0.1328053436279297, 0.1331682891845703, 0.13324159240722655, 0.13300953674316407, 0.13324662780761717, 0.13241993713378905, 0.1325875244140625, 0.13271449279785155, 0.13342105102539062, 0.13261993408203124, 0.1326266632080078, 0.13245404052734375, 0.13247731018066405, 0.13370991516113281]",tokens/s,7.556873159697436,, @@ -3748,7 +3748,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 38.12 MiB is free. Process 183152 has 14.70 GiB memory in use. Of the allocated memory 14.42 GiB is allocated by PyTorch, and 176.08 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 38.12 MiB is free. Process 189130 has 14.70 GiB memory in use. Of the allocated memory 14.42 GiB is allocated by PyTorch, and 176.08 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,01-ai/Yi-6B,01-ai/Yi-6B,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,810.307584,4683.923456,0.0,4288.67584,4213.842432,s,1,13.3825947265625,13.3825947265625,0.0,13.3825947265625,13.3825947265625,13.3825947265625,13.3825947265625,[13.3825947265625],,kWh,0.00018314163358749863,2.0194795041710082e-05,5.9438658661999924e-05,0.00026277508729120863,,MB,1203.580928,5107.54816,0.0,4699.717632,4535.245312,s,10,8.61562255859375,0.861562255859375,0.008665497944930763,0.8641178894042969,0.8676427673339844,0.8691550628662109,0.8703648992919921,"[0.8376939697265625, 0.8578470458984375, 0.8613602294921875, 0.86381201171875, 0.86074853515625, 0.8644237670898437, 0.8673067016601562, 0.866231689453125, 0.86553125, 0.8706673583984375]",tokens/s,297.13465075678124,kWh,2.512609172499987e-05,2.77097493549076e-06,1.671614763216668e-05,4.461321429265731e-05,tokens/kWh,5738210.170660891,MB,1254.387712,5115.936768,0.0,4708.10624,4535.247872,s,10,40.49503979492187,4.049503979492187,0.008818434365495443,4.05219580078125,4.058032958984375,4.060212646484375,4.0619563964843755,"[4.034101318359375, 4.038423095703125, 4.041629638671875, 4.04458203125, 4.0519599609375, 4.055001220703125, 4.052431640625, 4.062392333984375, 4.057548583984375, 4.056969970703125]",tokens/s,15.557460943130193,kWh,0.00011892334101375033,1.3116958640399224e-05,7.873059539183332e-05,0.00021077089504598282,tokens/kWh,298902.7492920956,,s,630,40.49235707473757,0.06427358265831357,0.0015613808815573744,0.06408670425415039,0.06476814651489259,0.06502831153869629,0.07511216720581056,"[0.07883795166015625, 0.06517282867431641, 0.06399033737182618, 0.06358208084106445, 0.06323929595947266, 0.06272499084472656, 0.06272784042358398, 0.06273904037475586, 0.06267811203002929, 0.0627658576965332, 0.06272927856445312, 0.06436297607421874, 0.0639450569152832, 0.06363328170776367, 0.06317478561401367, 0.06414745330810546, 0.06385391998291015, 0.06370131301879883, 0.06434236907958985, 0.06395510482788086, 0.06379913711547852, 0.06296960067749023, 0.06361520004272461, 0.06377679824829102, 0.06308883285522461, 0.06342943954467774, 0.06381792068481446, 0.0633469123840332, 0.06313430404663085, 0.06461849975585937, 0.06422099304199219, 0.0638199691772461, 0.0634093132019043, 0.06483977508544922, 0.06380006408691406, 0.06366617584228515, 0.06442111968994141, 0.06395775985717773, 0.06388051223754883, 0.0632364158630371, 0.06434371185302734, 0.0640823974609375, 0.0639526710510254, 0.06328163146972657, 0.06322134399414063, 0.06354927825927735, 0.06443417358398437, 0.06413369750976562, 0.0636313591003418, 0.06385459136962891, 0.06440464019775391, 0.0640021743774414, 0.06467219543457031, 0.06434342193603515, 0.06393334579467773, 0.06358425521850586, 0.0645481948852539, 0.06434678649902344, 0.06397747039794922, 0.06351433563232421, 0.06330352020263671, 0.06500188446044922, 0.06455219268798829, 0.07693769836425782, 0.06492147064208985, 0.06408739471435547, 0.06373033523559571, 0.06331763076782226, 0.06295158386230469, 0.06320095825195313, 0.06293056106567382, 0.06303382492065429, 0.06298790359497071, 0.06296275329589844, 0.06297983932495117, 0.06299609756469726, 0.0629961280822754, 0.0649551010131836, 0.06710681915283204, 0.06476179504394532, 0.06423149108886719, 0.06352719879150391, 0.06341603088378907, 0.06429417419433593, 0.06403663635253906, 0.06366227340698243, 0.06314006423950196, 0.06325827026367188, 0.06452877044677735, 0.06380323028564452, 0.06352758407592773, 0.06308438491821289, 0.06336959838867187, 0.06432745361328125, 0.0637583351135254, 0.06394204711914063, 0.064340576171875, 0.06398566436767578, 0.06374582290649414, 0.06383023834228516, 0.06456524658203125, 0.06415564727783203, 0.06390364837646484, 0.06357772827148438, 0.06443465423583984, 0.06453775787353516, 0.06418716430664062, 0.06393376159667968, 0.06361983871459961, 0.06444236755371094, 0.0639365119934082, 0.06369270324707031, 0.0644486083984375, 0.06403260803222656, 0.06384630584716797, 0.06356198501586914, 0.06447872161865234, 0.06412544250488281, 0.0639766731262207, 0.06348880004882812, 0.06452355194091797, 0.0646595230102539, 0.06434883117675781, 0.06393241500854492, 0.0634511375427246, 0.06363478469848632, 0.07382377624511718, 0.06468851470947265, 0.06397673416137695, 0.06359932708740235, 0.06322995376586914, 0.06435558319091797, 0.06392460632324219, 0.06360486221313477, 0.06313008117675781, 0.0629634895324707, 0.06299238586425782, 0.06474748992919922, 0.06424578857421875, 0.0637279052734375, 0.06355116653442383, 0.0643318099975586, 0.06368255996704102, 0.06357820892333985, 0.06436831665039063, 0.06392380905151367, 0.06372825622558594, 0.06377676773071289, 0.06486016082763672, 0.0642779541015625, 0.06379987335205078, 0.06348185729980468, 0.06330275344848633, 0.06440335845947266, 0.06384873580932617, 0.06357270431518555, 0.06407901000976562, 0.06407657623291016, 0.06385446548461914, 0.06494636535644531, 0.06439247894287109, 0.06387171173095703, 0.0638130226135254, 0.06351878356933593, 0.06440809631347656, 0.06471065521240234, 0.06398099136352539, 0.0633554573059082, 0.06436857604980468, 0.06392975997924805, 0.06367299270629882, 0.06373750305175781, 0.0644529571533203, 0.06406940460205078, 0.06442620849609375, 0.06395814514160156, 0.06356579208374023, 0.06376335906982422, 0.06454271697998047, 0.06424937438964844, 0.06495894622802735, 0.06439437103271485, 0.06396540832519532, 0.06380147171020507, 0.06445625305175781, 0.06410105895996093, 0.06381084823608399, 0.06449443054199219, 0.06414556884765625, 0.07492995452880859, 0.06452464294433594, 0.06403494262695313, 0.06359116744995118, 0.06315097427368165, 0.06425202941894531, 0.06392575836181641, 0.06389503860473633, 0.0632490234375, 0.06422566223144531, 0.0636819839477539, 0.06346912002563476, 0.06346566390991211, 0.06417286682128906, 0.06402845001220703, 0.06360496139526367, 0.0644106216430664, 0.0642325439453125, 0.06332611083984375, 0.06400972747802734, 0.06429478454589843, 0.06395065689086914, 0.06383084869384766, 0.06354358291625976, 0.06425296020507812, 0.06387782287597656, 0.06363750457763671, 0.06427017974853516, 0.06387868881225586, 0.0638490867614746, 0.06392623901367188, 0.06436156463623047, 0.06398223876953125, 0.06379983901977539, 0.06418816375732422, 0.064505859375, 0.06404096221923829, 0.0638914566040039, 0.0645090560913086, 0.06422617340087891, 0.06386003112792969, 0.06355424118041993, 0.06439910125732422, 0.06388761520385743, 0.06350438308715821, 0.0647906265258789, 0.06441522979736328, 0.06397699356079102, 0.06372236633300782, 0.0644582061767578, 0.06425039672851562, 0.06454608154296874, 0.06416588592529297, 0.06391471862792969, 0.06410614776611329, 0.06426787567138671, 0.06443084716796875, 0.06365343856811523, 0.0645263671875, 0.06400144195556641, 0.06385532760620118, 0.06457782745361328, 0.06448099517822266, 0.07537216186523438, 0.06461004638671874, 0.06401206207275391, 0.06360969543457032, 0.06322550582885743, 0.06404905700683594, 0.06376902389526368, 0.06357606506347656, 0.06314422225952149, 0.0641632308959961, 0.0636473617553711, 0.06421984100341797, 0.06379110336303712, 0.06360892868041992, 0.06380944061279296, 0.06426419067382813, 0.06401974487304687, 0.0645429458618164, 0.06417430114746094, 0.06360086441040039, 0.06428438568115234, 0.06390000152587891, 0.0641719970703125, 0.06391350555419922, 0.0637199363708496, 0.06364889526367187, 0.06416883087158202, 0.06385657501220703, 0.06374201583862305, 0.06441165161132813, 0.06381977462768555, 0.06365305709838867, 0.06437366485595702, 0.06487849426269532, 0.06428070068359375, 0.06387247848510742, 0.06446707153320312, 0.06407968139648437, 0.06493414306640626, 0.06615676879882812, 0.06381977462768555, 0.06358220672607422, 0.0634769287109375, 0.06500355529785157, 0.06440239715576172, 0.06374399948120117, 0.06419149017333985, 0.06447801971435548, 0.06456034851074219, 0.06441244506835937, 0.06415974426269531, 0.06456934356689453, 0.06422528076171875, 0.06504649353027343, 0.06456508636474609, 0.06413533020019531, 0.06357590484619141, 0.06401657867431641, 0.06507520294189453, 0.06457875061035157, 0.06402738952636719, 0.0638853759765625, 0.06462464141845703, 0.07551795196533204, 0.06456934356689453, 0.06394879913330079, 0.06416582489013672, 0.06370105743408203, 0.0635228157043457, 0.06346748733520508, 0.06398064041137695, 0.06359750366210938, 0.0632699203491211, 0.06420601654052735, 0.06385232162475586, 0.06363056182861328, 0.06320134353637695, 0.06418915557861328, 0.06505010986328125, 0.06477008056640625, 0.06423091125488281, 0.06377532958984375, 0.06418470764160156, 0.06426195526123046, 0.06387756729125976, 0.06366124725341797, 0.06336700820922851, 0.06432841491699219, 0.0644505615234375, 0.06410034942626953, 0.06381260681152344, 0.06378742218017579, 0.06438358306884766, 0.06542745971679688, 0.06431737518310547, 0.06476806640625, 0.06437593841552734, 0.06416063690185547, 0.06491526031494141, 0.06424508666992187, 0.06384316635131836, 0.06370297622680664, 0.06437689971923828, 0.0649583969116211, 0.0644109115600586, 0.06398236846923828, 0.06366761779785156, 0.06386368179321289, 0.06446665954589843, 0.06380953598022461, 0.06574845123291016, 0.06378675079345703, 0.06438790130615234, 0.06409827423095703, 0.06466873931884766, 0.0642011489868164, 0.06403727722167969, 0.06445065307617187, 0.06514073944091797, 0.06456934356689453, 0.06424761962890625, 0.06402272033691406, 0.0640610580444336, 0.0646123504638672, 0.06455567932128906, 0.06399151992797851, 0.07492991638183594, 0.06456793975830079, 0.06400204467773438, 0.06357196807861328, 0.06315827178955079, 0.06459552001953126, 0.06398611068725586, 0.06355699157714843, 0.0632856330871582, 0.06357958221435547, 0.06406790161132812, 0.06367283248901368, 0.0633449592590332, 0.06420652770996094, 0.06389126586914062, 0.06417635345458984, 0.06461436462402344, 0.0643399658203125, 0.06400819396972657, 0.06369459152221679, 0.0637768325805664, 0.0643094711303711, 0.06341363143920899, 0.06392483139038085, 0.0643276824951172, 0.0641630401611328, 0.06345721435546875, 0.06539759826660156, 0.06430105590820312, 0.06407955169677734, 0.06437715148925781, 0.0644382095336914, 0.06483542633056641, 0.06432790374755859, 0.06378438568115234, 0.06439993286132813, 0.06397235107421875, 0.0644775390625, 0.06417030334472656, 0.06380886459350586, 0.06344931030273437, 0.0644016342163086, 0.0639309425354004, 0.06461440277099609, 0.06449152374267578, 0.06452601623535156, 0.06423725128173828, 0.06501644897460937, 0.06437586975097656, 0.06398252868652343, 0.06384608078002929, 0.0640904312133789, 0.06449056243896484, 0.0640992660522461, 0.06395391845703124, 0.06415795135498047, 0.06456374359130859, 0.06418447875976563, 0.0645033950805664, 0.06420291137695312, 0.0642174072265625, 0.06462834930419922, 0.0651960678100586, 0.07654399871826172, 0.06484716796875, 0.06434886169433594, 0.06389350509643554, 0.06374409484863282, 0.06655133056640625, 0.06360271835327148, 0.06421129608154297, 0.06356172943115235, 0.0636231689453125, 0.0635814094543457, 0.06414828491210937, 0.06371865463256836, 0.06331606292724609, 0.064837890625, 0.06497318267822266, 0.06433177947998046, 0.06402252960205078, 0.06434611511230469, 0.06422732543945313, 0.06344467163085937, 0.0643703384399414, 0.06394464111328126, 0.06436265563964844, 0.0640456314086914, 0.06382953643798828, 0.06429878234863282, 0.06384662246704101, 0.06363820648193359, 0.06449517059326172, 0.0648153305053711, 0.06443417358398437, 0.06454681396484375, 0.06426214599609376, 0.06384435272216797, 0.06695116424560547, 0.06386687850952148, 0.06350787353515625, 0.0654131851196289, 0.06441629028320313, 0.06393856048583985, 0.06378905487060547, 0.06389680099487305, 0.06436943817138673, 0.06390934371948243, 0.06397139358520508, 0.06449199676513671, 0.06437423706054687, 0.0645633316040039, 0.0645206069946289, 0.0644814682006836, 0.064325439453125, 0.06422105407714844, 0.06557884979248046, 0.06444866943359374, 0.06401017761230468, 0.06391212844848633, 0.06395699310302734, 0.0646123504638672, 0.0644874267578125, 0.06414950561523437, 0.06510963439941406, 0.06422470092773437, 0.07518659210205078, 0.06450534057617187, 0.06406195068359374, 0.06460415649414063, 0.06399987030029297, 0.06354137420654297, 0.06339567947387695, 0.06402655792236328, 0.06361110305786133, 0.06428262329101563, 0.06396723175048828, 0.06374195098876953, 0.06334678268432617, 0.06427638244628907, 0.0638397102355957, 0.06467763519287109, 0.06433052825927735, 0.064036865234375, 0.06418637084960938, 0.0637317771911621, 0.06471612548828125, 0.06420950317382812, 0.06381913757324219, 0.06381427383422851, 0.06424371337890625, 0.06479180908203125, 0.06419942474365234, 0.06384844970703125, 0.06359878540039063, 0.06426809692382812, 0.06408601379394531, 0.06441983795166016, 0.06430643463134765, 0.06385283279418945, 0.06432975769042969, 0.0639799690246582, 0.06444009399414062, 0.06412310028076172, 0.06381296157836915, 0.06418294525146484, 0.06475971221923828, 0.06495552062988282, 0.06438114929199219, 0.06378726577758789, 0.06422742462158203, 0.0644758071899414, 0.06495945739746094, 0.0643755874633789, 0.06413648223876953, 0.06489775848388672, 0.06409625244140625, 0.06494380950927735, 0.06434425354003906, 0.06388748931884766, 0.06459091186523437, 0.06418118286132812, 0.06499737548828124, 0.06451824188232422, 0.06576118469238282, 0.06392217636108398, 0.06448025512695313, 0.06399897766113281, 0.0642007064819336, 0.07526537322998048, 0.06485635375976563, 0.06396556854248046, 0.06350848007202148, 0.0634224624633789, 0.06410240173339844, 0.06363919830322265, 0.06352262496948242, 0.0630338897705078, 0.064104736328125, 0.06366547012329102, 0.06352278518676757, 0.06342873764038086, 0.06422278594970703, 0.06473554992675781, 0.06542176055908203, 0.06533734130859375, 0.06435750579833985, 0.06411148834228515, 0.06333235168457031, 0.06362863922119141, 0.06534210968017579, 0.06421702575683594, 0.06390313720703125, 0.06368483352661133, 0.06379286575317383, 0.0641370849609375, 0.06380563354492187, 0.06349177551269532, 0.06483197021484376, 0.06454118347167968, 0.0647759017944336, 0.06444624328613281, 0.06436637115478516, 0.06458211517333984, 0.06432998657226563, 0.06473113250732422, 0.06507315063476563, 0.06415315246582032, 0.06381232070922851, 0.06364950561523437, 0.06436204528808594, 0.06400249481201171, 0.06389775848388672, 0.06392345428466797, 0.06458428955078124, 0.06442915344238281, 0.06457027435302734, 0.06465945434570312, 0.0645010528564453, 0.06423126220703125, 0.06476886749267578, 0.06454476928710938, 0.06442598724365234, 0.06395465469360352, 0.06388281631469726, 0.0648259506225586, 0.06449779510498047, 0.06388121414184571, 0.06503801727294922, 0.06453414154052735, 0.06417052459716797, 0.06414498901367187]",tokens/s,15.558491663925528,, @@ -3801,7 +3801,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 101184 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 107245 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-0.5B,Qwen/Qwen1.5-0.5B,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,1300.811776,1093.599232,0.0,698.351616,690.178048,s,1,9.0914990234375,9.0914990234375,0.0,9.0914990234375,9.0914990234375,9.0914990234375,9.0914990234375,[9.0914990234375],,kWh,4.987506294163874e-05,5.4943037951315234e-06,1.7437791727969998e-05,7.280715846474025e-05,,MB,1435.410432,1408.172032,0.0,1000.341504,957.775872,s,10,0.6234354209899903,0.062343542098999026,0.0003477882052090371,0.062364255905151364,0.06268343505859375,0.06280832557678223,0.06290823799133301,"[0.06265568161010743, 0.06167334365844727, 0.06215894317626953, 0.06193344116210937, 0.06265401458740234, 0.0624372787475586, 0.0629332160949707, 0.06226099014282226, 0.062372673034667966, 0.06235583877563477]",tokens/s,4106.279357587388,kWh,1.909804073257004e-06,2.104694802594633e-07,1.2647159718300294e-06,3.3849895253464964e-06,tokens/kWh,75628003.59738046,MB,1462.214656,1420.754944,0.0,1012.924416,957.778432,s,10,27.223863525390627,2.7223863525390626,0.009427315305820406,2.7192373046875,2.733354736328125,2.7382030029296875,2.7420816162109376,"[2.719857666015625, 2.74305126953125, 2.718616943359375, 2.712888671875, 2.7300703125, 2.73227734375, 2.7155078125, 2.71289599609375, 2.714979248046875, 2.72371826171875]",tokens/s,23.141461880031237,kWh,7.949106879882261e-05,8.767922032024084e-06,3.395020108897235e-05,0.00012220919191981903,tokens/kWh,515509.50472967734,,s,630,27.221254787445105,0.04320834093245249,0.000528114719663762,0.04307913780212402,0.043606554794311525,0.04398572731018066,0.045682334403991706,"[0.04331292724609375, 0.04353705596923828, 0.04330876922607422, 0.04297286224365234, 0.04332556915283203, 0.04328905487060547, 0.043525375366210935, 0.04314966583251953, 0.04310467147827148, 0.04306262588500977, 0.04283824157714844, 0.04278726577758789, 0.04292748641967774, 0.043000446319580075, 0.042723232269287106, 0.04281967926025391, 0.04296192169189453, 0.04319302368164062, 0.04285984039306641, 0.04284108734130859, 0.04309401702880859, 0.043036670684814454, 0.04317113494873047, 0.04297798538208008, 0.043749183654785154, 0.04318207931518555, 0.04293593597412109, 0.0432624626159668, 0.04321900939941406, 0.044125953674316404, 0.043983104705810544, 0.043259361267089846, 0.04342428970336914, 0.043177982330322266, 0.04291142272949219, 0.04317740631103516, 0.04305964660644531, 0.04289785766601562, 0.04282598495483399, 0.042845279693603515, 0.0428243522644043, 0.043062335968017576, 0.04314822387695313, 0.043169792175292966, 0.0433889274597168, 0.04375542449951172, 0.04366723251342773, 0.04313119888305664, 0.04322035217285156, 0.04311715316772461, 0.04311151885986328, 0.04311088180541992, 0.0431129264831543, 0.043222881317138674, 0.04331660842895508, 0.04331804656982422, 0.043046913146972655, 0.04304399871826172, 0.04364988708496094, 0.04316556930541992, 0.04310371017456055, 0.04298614501953125, 0.04307129669189453, 0.04358121490478516, 0.04322601699829102, 0.043093311309814454, 0.042998462677001956, 0.04298342514038086, 0.04320665740966797, 0.04326009750366211, 0.04308176040649414, 0.04334979248046875, 0.04304230499267578, 0.042797569274902345, 0.04351385498046875, 0.04329372787475586, 0.04331414413452148, 0.04313497543334961, 0.04342281723022461, 0.04294902420043945, 0.042895870208740236, 0.04317184066772461, 0.04345446395874023, 0.04315913772583008, 0.04325827026367188, 0.04321484756469727, 0.04322089767456055, 0.042948703765869144, 0.043038719177246096, 0.043597824096679685, 0.043243518829345705, 0.04331235122680664, 0.043408096313476564, 0.04340073776245117, 0.04339766311645508, 0.04352000045776367, 0.04429414367675781, 0.04410508728027344, 0.04393638229370117, 0.04432486343383789, 0.04337667083740234, 0.04334982299804688, 0.04304003143310547, 0.04301001739501953, 0.04333353424072266, 0.04302707290649414, 0.0455601921081543, 0.043480064392089846, 0.043240447998046876, 0.043235328674316405, 0.04336844635009766, 0.044423168182373046, 0.04343952178955078, 0.04313763046264649, 0.043218944549560545, 0.043142814636230466, 0.04346505737304687, 0.04343369674682617, 0.04901286315917969, 0.04343619155883789, 0.045797279357910156, 0.044500225067138674, 0.043916126251220704, 0.04352000045776367, 0.04396819305419922, 0.0432061767578125, 0.04354339218139648, 0.04353023910522461, 0.04313497543334961, 0.043140705108642576, 0.04313731384277344, 0.04325593566894531, 0.042985183715820316, 0.04349900817871094, 0.04306537628173828, 0.043340286254882815, 0.04307942581176758, 0.043143871307373044, 0.042968894958496096, 0.04340326309204102, 0.043053054809570314, 0.04294819259643555, 0.043006305694580076, 0.04301942443847656, 0.04306403350830078, 0.04291392135620117, 0.04290943908691406, 0.04313958358764648, 0.04306515121459961, 0.04350886535644531, 0.04284288024902344, 0.04305897521972656, 0.042928417205810546, 0.042942527770996095, 0.04389068984985352, 0.04332313537597656, 0.04322739028930664, 0.0429486083984375, 0.04315135955810547, 0.04291788864135742, 0.042995712280273435, 0.04340550231933594, 0.043111743927001955, 0.043364864349365234, 0.04288716888427734, 0.0429027214050293, 0.04285862350463867, 0.043005790710449215, 0.043299678802490235, 0.04647222518920899, 0.04318038558959961, 0.042870944976806644, 0.04300009536743164, 0.0429917106628418, 0.04297475051879883, 0.043149505615234375, 0.042772254943847655, 0.04297500610351562, 0.04275008010864258, 0.042947265625, 0.04288694381713867, 0.04284147262573242, 0.04283014297485352, 0.04289590454101563, 0.04319641494750977, 0.04319641494750977, 0.04328857421875, 0.043020320892333985, 0.04319638442993164, 0.04338278579711914, 0.043046913146972655, 0.042949790954589846, 0.04304326248168945, 0.042970783233642576, 0.04273638534545898, 0.042643329620361325, 0.04326768112182617, 0.04267385482788086, 0.044544864654541015, 0.043257217407226566, 0.04276287841796875, 0.044203231811523434, 0.04311324691772461, 0.042979328155517575, 0.04294377517700195, 0.0428100814819336, 0.04300799942016602, 0.0430489616394043, 0.044076671600341795, 0.04280767822265625, 0.04281958389282227, 0.042864543914794925, 0.043140480041503906, 0.042909950256347654, 0.04300233459472656, 0.04299980926513672, 0.04300764846801758, 0.04282198333740234, 0.042799102783203126, 0.04255081558227539, 0.04291823959350586, 0.04311257553100586, 0.0432988166809082, 0.04283343887329102, 0.04346928024291992, 0.0426776008605957, 0.04306358337402344, 0.042946945190429686, 0.043009342193603514, 0.04275651168823242, 0.04286707305908203, 0.04303766250610352, 0.04277958297729492, 0.04321279907226563, 0.04302188873291016, 0.04343427276611328, 0.04293257522583008, 0.042854209899902344, 0.04310537719726563, 0.04291676712036133, 0.042698879241943356, 0.04298124694824219, 0.04295270538330078, 0.042995712280273435, 0.0430489616394043, 0.042935871124267576, 0.04336051177978516, 0.043853759765625, 0.043398944854736325, 0.04289788818359375, 0.043063297271728515, 0.04301615905761719, 0.04342950439453125, 0.043686176300048826, 0.04339279937744141, 0.043112319946289064, 0.04310844802856445, 0.043450721740722655, 0.04312188720703125, 0.04293711853027344, 0.04311856079101563, 0.0431016960144043, 0.04301375961303711, 0.0430294075012207, 0.04290764617919922, 0.0430571517944336, 0.04295065689086914, 0.043364158630371095, 0.04347859191894531, 0.04414323043823242, 0.04336819076538086, 0.04299599838256836, 0.042958465576171875, 0.04307612609863281, 0.04299142456054687, 0.043087390899658205, 0.0431416015625, 0.04309196853637695, 0.0440709114074707, 0.0432182731628418, 0.04351990509033203, 0.043610527038574216, 0.04306774520874023, 0.043140670776367185, 0.04336240005493164, 0.04652431869506836, 0.04355920028686523, 0.043233440399169924, 0.04339494323730469, 0.043469024658203126, 0.043267871856689455, 0.043094142913818356, 0.04362444686889649, 0.04385177612304687, 0.04328243255615234, 0.04350566482543945, 0.043261951446533206, 0.043200511932373044, 0.04305920028686523, 0.04333315277099609, 0.042923904418945315, 0.04351587295532226, 0.04314380645751953, 0.04310630416870117, 0.043104255676269534, 0.04324723052978516, 0.04347942352294922, 0.04353843307495117, 0.04329203033447265, 0.04327619171142578, 0.04382547378540039, 0.043413921356201174, 0.04303664016723633, 0.043078849792480466, 0.043053184509277344, 0.0435316162109375, 0.04356121444702148, 0.04343235015869141, 0.043409374237060545, 0.044272960662841795, 0.043686622619628905, 0.04333363342285156, 0.04520959854125976, 0.04346060943603516, 0.043225086212158204, 0.04300566482543945, 0.043098209381103515, 0.04302608108520508, 0.0434672966003418, 0.04306243133544922, 0.04299043273925781, 0.04484719848632813, 0.04319551849365234, 0.043635009765625, 0.043467231750488285, 0.043278175354003905, 0.04308803176879883, 0.04317753601074219, 0.043199935913085935, 0.04315785598754883, 0.043251583099365234, 0.04360611343383789, 0.04343840026855469, 0.043176319122314455, 0.04327219009399414, 0.04302345657348633, 0.04430665588378906, 0.04368864059448242, 0.04332940673828125, 0.042903678894042965, 0.04308377456665039, 0.04295430374145508, 0.04397715377807617, 0.04289263916015625, 0.042789215087890624, 0.042936641693115236, 0.04312268829345703, 0.04296438217163086, 0.043072097778320315, 0.043071487426757815, 0.043757312774658205, 0.042938625335693356, 0.04282483291625976, 0.042875232696533205, 0.0430695686340332, 0.042961055755615235, 0.043456768035888674, 0.04512768173217773, 0.04321279907226563, 0.04338390350341797, 0.04312771224975586, 0.04374118423461914, 0.043159358978271486, 0.04323142242431641, 0.043374591827392575, 0.04364492797851562, 0.043415550231933595, 0.0430301742553711, 0.043464702606201173, 0.043587039947509766, 0.04299216079711914, 0.04273779296875, 0.04291929626464844, 0.04280985641479492, 0.04284524917602539, 0.04270585632324219, 0.04292393493652344, 0.04301193618774414, 0.04262937545776367, 0.04298652648925781, 0.042646495819091794, 0.04301811218261719, 0.042821758270263674, 0.04291788864135742, 0.04281686401367187, 0.04616463851928711, 0.04294460678100586, 0.04307958221435547, 0.04362444686889649, 0.042967041015625, 0.043103649139404294, 0.04319049453735352, 0.04335577774047852, 0.04310009765625, 0.0429444808959961, 0.043404193878173826, 0.04290460968017578, 0.04308198547363281, 0.04296732711791992, 0.04308211135864258, 0.04290150451660156, 0.042893310546875, 0.043151168823242186, 0.043270336151123044, 0.04348928070068359, 0.0428928337097168, 0.04288876724243164, 0.0429202880859375, 0.04272800064086914, 0.042819263458251954, 0.04301446533203125, 0.04298950576782227, 0.042700672149658205, 0.0426740493774414, 0.04279299163818359, 0.044973728179931644, 0.043171489715576175, 0.04297830581665039, 0.04313699340820312, 0.04308915328979492, 0.04289779281616211, 0.04301968002319336, 0.042937313079833984, 0.04354048156738281, 0.04287612915039062, 0.04281238555908203, 0.043736385345458983, 0.04322140884399414, 0.04288726425170898, 0.043063297271728515, 0.043020286560058595, 0.04394128036499023, 0.04323705673217773, 0.04304777526855469, 0.042958911895751954, 0.04312460708618164, 0.04356927871704101, 0.04311228942871094, 0.042931934356689454, 0.04291628646850586, 0.04331315231323242, 0.04278793716430664, 0.04287376022338867, 0.042913791656494144, 0.04338687896728516, 0.04290764617919922, 0.0428642578125, 0.04534214401245117, 0.04285740661621094, 0.04300595092773438, 0.04288476943969727, 0.04316732788085938, 0.04296371078491211, 0.04296192169189453, 0.04295372772216797, 0.04302403259277344, 0.04320086288452148, 0.04312255859375, 0.04303007888793945, 0.043033153533935546, 0.04300163269042969, 0.04294697570800781, 0.04270371246337891, 0.04403094482421875, 0.04295212936401367, 0.04383318328857422, 0.04317871856689453, 0.04277248001098633, 0.04288003158569336, 0.042855392456054686, 0.042962944030761716, 0.043053054809570314, 0.04324726486206055, 0.042924385070800784, 0.042821632385253904, 0.04273971176147461, 0.042643455505371096, 0.042828895568847655, 0.04291267013549805, 0.04302438354492188, 0.04287897491455078, 0.0431629753112793, 0.04278054428100586, 0.04274665451049805, 0.04290457534790039, 0.042787841796875, 0.04315964889526367, 0.042831775665283206, 0.04279203033447266, 0.04288358306884766, 0.04282614517211914, 0.04301728057861328, 0.0430621452331543, 0.04305417633056641, 0.043278335571289066, 0.042999168395996094, 0.04303116989135742, 0.04303244781494141, 0.043014015197753906, 0.043016384124755856, 0.04287289428710937, 0.04285955047607422, 0.042912734985351565, 0.042796222686767575, 0.04293305587768555, 0.042978401184082034, 0.042793632507324215, 0.04269068908691406, 0.042762081146240236, 0.043010337829589844, 0.0429417610168457, 0.04293292617797852, 0.04284620666503906, 0.0429219856262207, 0.04295065689086914, 0.04294220733642578, 0.04281779098510742, 0.04283801651000976, 0.04319609451293945, 0.04308614349365234, 0.043499519348144534, 0.04290150451660156, 0.0431800651550293, 0.0429951057434082, 0.04386377716064453, 0.04373385620117187, 0.04396236801147461, 0.0432657585144043, 0.043296993255615236, 0.043320926666259765, 0.043065536499023435, 0.043355873107910156, 0.04292870330810547, 0.042689697265625, 0.04293427276611328, 0.04292256164550781, 0.042942752838134764, 0.043090110778808595, 0.04313065719604492, 0.044378143310546875, 0.04294831848144531, 0.04303244781494141, 0.04281590270996094, 0.043345569610595706, 0.04263955307006836, 0.043159713745117186, 0.04272742462158203, 0.042660896301269534, 0.04355753707885742, 0.04448281478881836, 0.04276025772094726, 0.04283564758300781, 0.04304313659667969, 0.04314217758178711, 0.043987873077392575, 0.04300316619873047, 0.0426644172668457, 0.0430362548828125, 0.04306179046630859, 0.04347804641723633, 0.04297987365722656, 0.04308329772949219, 0.04576950454711914, 0.04435968017578125, 0.04356915283203125, 0.042897407531738284, 0.043068992614746095, 0.04286918258666992, 0.04297727966308594, 0.042942272186279294, 0.04310371017456055, 0.042963550567626956, 0.04291551971435547, 0.042856094360351565, 0.042928993225097654, 0.04279289627075195, 0.04353551864624024, 0.0431808967590332, 0.04286646270751953, 0.042899070739746095, 0.04286323165893555, 0.04283184051513672, 0.0429136962890625, 0.04300352096557617, 0.04308816146850586, 0.04351961517333985, 0.04313350296020508, 0.04307558441162109, 0.0432988166809082, 0.04390092849731445, 0.04332748794555664, 0.043328510284423825, 0.04436067199707031, 0.04342284774780274, 0.043289505004882815, 0.04304076766967774, 0.042990718841552734, 0.04297203063964844, 0.04298137664794922, 0.04309196853637695, 0.04287286376953125, 0.04294652938842773, 0.042897151947021483, 0.04314751815795898, 0.0433070068359375, 0.04314278411865234, 0.0429428482055664, 0.04294192123413086, 0.04319855880737305, 0.04299753570556641, 0.042877601623535155, 0.04314089584350586, 0.043001697540283206, 0.04573222351074219, 0.043284481048583984, 0.043804672241210936, 0.04327654266357422, 0.04304886245727539, 0.043326366424560545, 0.04299462509155273]",tokens/s,23.14367963267319,, @@ -3852,7 +3852,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 128.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 20.12 MiB is free. Process 200378 has 14.72 GiB memory in use. Of the allocated memory 14.54 GiB is allocated by PyTorch, and 67.18 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 128.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 20.12 MiB is free. Process 206190 has 14.72 GiB memory in use. Of the allocated memory 14.54 GiB is allocated by PyTorch, and 67.18 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gemma,google/gemma-7b,google/gemma-7b,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.223-212.873.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.44.2,,0.34.2,,,,1.21.4,,,,0.12.0,,,True,MB,902.107136,6589.120512,0.0,6186.5984,6098.951168,s,1,14.927158203125,14.927158203125,0.0,14.927158203125,14.927158203125,14.927158203125,14.927158203125,[14.927158203125],,kWh,0.00023453259544164667,2.585945365850905e-05,7.786450673599732e-05,0.00033825655583615303,,MB,1400.89344,7306.346496,0.0,6891.241472,6575.183872,s,10,11.447591552734373,1.1447591552734377,0.0066483095537777275,1.1478670043945312,1.1498494506835937,1.1500173767089843,1.1501517175292968,"[1.12945166015625, 1.135996337890625, 1.1472989501953126, 1.1492005615234375, 1.1413291015625, 1.150185302734375, 1.1464898681640625, 1.14843505859375, 1.1498121337890626, 1.149392578125]",tokens/s,223.62782496275534,kWh,3.3365034042085424e-05,3.6795308681339335e-06,2.2279823379398044e-05,5.932438828961741e-05,tokens/kWh,4315257.306155883,MB,1456.68096,7310.5408,0.0,6893.338624,6575.186432,s,10,57.65196728515625,5.765196728515625,0.02080366906057327,5.76491748046875,5.7923897460937495,5.793774462890625,5.794882236328125,"[5.72457958984375, 5.74243505859375, 5.7578486328125, 5.75404638671875, 5.7656162109375, 5.76421875, 5.7739482421875, 5.782033203125, 5.79208203125, 5.7951591796875]",tokens/s,10.927640975093095,kWh,0.00016985143455291036,1.8736066947479215e-05,0.00011250575667120233,0.00030109325817159196,tokens/kWh,209237.49798508117,,s,630,57.64940335845947,0.09150698945787218,0.0016950634204783265,0.09123726654052734,0.09286253204345704,0.09322157020568847,0.101445630569458,"[0.10121647644042969, 0.08966780853271485, 0.08922537231445313, 0.08915487670898438, 0.09003692626953125, 0.08983961486816407, 0.08947917175292969, 0.09059532928466797, 0.08961023712158203, 0.09085660552978515, 0.09208303833007812, 0.09220438385009766, 0.09039740753173828, 0.08951769256591798, 0.0895450210571289, 0.09141001892089844, 0.08947958374023438, 0.0914698257446289, 0.08953788757324219, 0.0909208984375, 0.08975775909423828, 0.09286835479736329, 0.09045699310302735, 0.0915241928100586, 0.08966851043701173, 0.09152102661132812, 0.08971017456054688, 0.09150498962402344, 0.08960556793212891, 0.09152780914306641, 0.08961228942871094, 0.09190809631347656, 0.09010336303710938, 0.09192214202880859, 0.09018646240234375, 0.09156198120117187, 0.08967782592773438, 0.09199929809570312, 0.08970950317382813, 0.09160054779052734, 0.08963107299804687, 0.09187657928466797, 0.09007315063476562, 0.09160492706298828, 0.09037257385253906, 0.09103330993652343, 0.09028668975830079, 0.09203481292724609, 0.09042537689208985, 0.09145571136474609, 0.09033113861083984, 0.09146572875976562, 0.09111750030517578, 0.09093497467041016, 0.09116915130615234, 0.09134454345703125, 0.09070217895507812, 0.09110291290283203, 0.09078982543945313, 0.09091487884521485, 0.09079952239990234, 0.09054009246826172, 0.09168284606933594, 0.10139017486572266, 0.08969174194335937, 0.09037001800537109, 0.08935485076904297, 0.090640380859375, 0.08945664215087891, 0.09170086669921874, 0.08948365020751953, 0.09100262451171875, 0.09016140747070313, 0.09541232299804688, 0.09139600372314453, 0.09091891479492188, 0.09050630187988282, 0.08950675201416015, 0.09040076446533203, 0.09044985961914062, 0.09014265441894531, 0.0903578872680664, 0.08976179504394531, 0.09309184265136719, 0.09180364990234376, 0.09114796447753906, 0.0908639373779297, 0.09072230529785157, 0.09059123229980469, 0.09023078155517578, 0.08986809539794922, 0.09058528137207031, 0.09040220642089844, 0.09055702209472656, 0.0919019546508789, 0.09110733032226563, 0.09238118743896484, 0.09095513916015625, 0.09222412872314453, 0.08983248138427734, 0.09218883514404297, 0.08979948425292969, 0.09198579406738282, 0.0897100830078125, 0.09271974182128906, 0.09098854064941406, 0.09213712310791015, 0.09093564605712891, 0.09181292724609375, 0.09072278594970704, 0.09037667083740235, 0.09083673858642578, 0.09036998748779297, 0.09073827362060546, 0.09103433227539062, 0.09114419555664062, 0.09107004547119141, 0.09293567657470703, 0.0913130874633789, 0.09229043579101562, 0.09051197052001952, 0.09167769622802735, 0.09055744171142578, 0.09177702331542968, 0.09058665466308594, 0.09160243225097656, 0.10264575958251954, 0.09019187164306641, 0.09170329284667969, 0.08954879760742188, 0.09166438293457031, 0.09041919708251953, 0.0918363494873047, 0.08955430603027344, 0.09165484619140625, 0.09019596862792968, 0.09459916687011719, 0.09141452789306641, 0.09145279693603516, 0.09027238464355469, 0.09097984313964844, 0.09035008239746094, 0.09144319915771484, 0.09037004852294922, 0.0912212142944336, 0.09038313293457031, 0.09215580749511719, 0.09109718322753907, 0.09172480010986328, 0.09140038299560548, 0.09038726043701172, 0.09060294342041016, 0.09033990478515624, 0.09088124847412109, 0.09023158264160157, 0.09143004608154297, 0.09100319671630859, 0.09175421142578125, 0.09121692657470704, 0.09231747436523438, 0.09041468811035157, 0.09281986999511718, 0.08968192291259766, 0.09274956512451171, 0.08981529235839844, 0.0923156509399414, 0.09091417694091797, 0.0923592987060547, 0.09069686126708984, 0.09299849700927734, 0.09096141052246094, 0.09121616363525391, 0.09057091522216797, 0.09105241394042969, 0.09052310180664062, 0.09085155487060546, 0.09131593322753906, 0.09124073791503906, 0.09198387145996094, 0.09140016174316407, 0.0930570526123047, 0.09055010986328126, 0.09247350311279297, 0.0906055679321289, 0.09169058990478515, 0.09055785369873047, 0.09122303771972656, 0.09117286682128906, 0.09196297454833985, 0.10156230163574219, 0.09029017639160156, 0.09007984161376953, 0.09041731262207031, 0.09024716949462891, 0.09015090942382813, 0.09080825805664063, 0.08971180725097656, 0.0909648666381836, 0.09030802917480468, 0.09574774169921875, 0.09189263916015625, 0.09077760314941406, 0.09129984283447265, 0.09010176086425781, 0.09058509063720703, 0.08998502349853515, 0.08987612915039063, 0.09058544158935547, 0.09011814117431641, 0.09172809600830079, 0.09310950469970704, 0.09163565063476563, 0.09323785400390625, 0.09032825469970703, 0.09164268493652344, 0.09031849670410157, 0.09128569793701172, 0.09035298919677734, 0.090947998046875, 0.09105206298828125, 0.09196173095703125, 0.09110844421386718, 0.0919766082763672, 0.09117820739746094, 0.09121257781982423, 0.09103769683837891, 0.09020620727539062, 0.09142998504638672, 0.09011510467529296, 0.09125446319580079, 0.09093897247314453, 0.09284690856933593, 0.0911951675415039, 0.09304473876953125, 0.09109503936767578, 0.09125682830810547, 0.09090866851806641, 0.09049088287353516, 0.09079190063476562, 0.09110470581054687, 0.09105862426757813, 0.09128771209716798, 0.09121711730957031, 0.09197872161865234, 0.09255302429199219, 0.0912691192626953, 0.09279283142089843, 0.0905502700805664, 0.09164383697509766, 0.09047456359863282, 0.09191171264648437, 0.09076719665527344, 0.10215833282470703, 0.09035065460205079, 0.09016995239257812, 0.08977238464355469, 0.08957513427734375, 0.0908864288330078, 0.09039052581787109, 0.0906688003540039, 0.09020960235595703, 0.09092995452880859, 0.09434915161132812, 0.09378438568115234, 0.09082669067382812, 0.091840576171875, 0.09077750396728515, 0.09144086456298828, 0.09026358032226563, 0.09150089263916016, 0.08957286071777344, 0.09143551635742188, 0.09098239898681641, 0.0939391326904297, 0.09181603240966797, 0.0911200942993164, 0.09139417266845704, 0.09033920288085938, 0.09097740936279297, 0.09051840209960937, 0.0910355224609375, 0.08971046447753907, 0.09105795288085937, 0.09225059509277343, 0.09352191925048828, 0.09122406768798828, 0.09279692840576172, 0.09046985626220704, 0.0915297622680664, 0.0909656982421875, 0.0915766372680664, 0.09101312255859376, 0.0904599380493164, 0.09120956420898438, 0.09112204742431641, 0.09225830078125, 0.09166175842285157, 0.09293385314941406, 0.09054908752441407, 0.09258092498779297, 0.09091168212890625, 0.09167871856689454, 0.09093318176269531, 0.0916808319091797, 0.09107977294921875, 0.09135945892333984, 0.09204601287841797, 0.09132441711425782, 0.0921743392944336, 0.09144115447998047, 0.09288089752197265, 0.09043312072753906, 0.0924224624633789, 0.09062963104248047, 0.0924534683227539, 0.10141020965576172, 0.09019248199462891, 0.09008451080322266, 0.09047030639648437, 0.09102022552490234, 0.09014476776123047, 0.08947731018066406, 0.09133856201171875, 0.09017139434814453, 0.09084928131103516, 0.09460518646240235, 0.09316365051269532, 0.09069977569580077, 0.09201010894775391, 0.0901349105834961, 0.09245491027832031, 0.0903024673461914, 0.09112576293945313, 0.0901242904663086, 0.09138585662841797, 0.09148947143554688, 0.09286943817138672, 0.09161433410644532, 0.09100291442871093, 0.09156301116943359, 0.09045565032958984, 0.09191817474365234, 0.09026719665527344, 0.09181475067138672, 0.09029631805419921, 0.09318195343017578, 0.09116835021972657, 0.09304220581054687, 0.09118790435791016, 0.09089984130859376, 0.09105680084228515, 0.09109315490722657, 0.0912056655883789, 0.09095574188232422, 0.09136911773681641, 0.09099858856201172, 0.09231209564208985, 0.09117036437988281, 0.09326432037353516, 0.09088956451416015, 0.0918736343383789, 0.09122191619873046, 0.09069551849365234, 0.09149225616455078, 0.09056534576416016, 0.09120146942138672, 0.09111254119873047, 0.09277327728271484, 0.09197567749023437, 0.09308319854736329, 0.09136582183837891, 0.09119033813476562, 0.09163785552978515, 0.09048355102539063, 0.09140428924560547, 0.09122815704345703, 0.0912523193359375, 0.09115471649169922, 0.10269324493408204, 0.09141862487792969, 0.09014697265625, 0.09181577301025391, 0.09017139434814453, 0.09228278350830078, 0.0902053451538086, 0.09170630645751954, 0.09024307250976563, 0.09203305816650391, 0.09287881469726562, 0.09213951873779297, 0.09153065490722656, 0.09098262023925781, 0.091432861328125, 0.09021692657470703, 0.09136089324951172, 0.09021887969970703, 0.09215122985839844, 0.09051155090332032, 0.09339737701416016, 0.09118659210205078, 0.09306758117675781, 0.09121974182128906, 0.09139401245117187, 0.09074127960205078, 0.0909271011352539, 0.09044172668457032, 0.09096924591064454, 0.09097711944580078, 0.09110118103027344, 0.09378585815429688, 0.09133286285400391, 0.09300502777099609, 0.09118595123291015, 0.09097974395751954, 0.09155763244628906, 0.09106031799316407, 0.09041177368164062, 0.09095362854003906, 0.09096403503417969, 0.09157756805419921, 0.09304966735839844, 0.09125254058837891, 0.09286188507080079, 0.09126579284667968, 0.09176678466796875, 0.09115647888183594, 0.09054726409912109, 0.09125910186767579, 0.09106854248046875, 0.09166108703613281, 0.09130496215820312, 0.09306768035888671, 0.09153372955322266, 0.09287184143066406, 0.09142972564697266, 0.0911911392211914, 0.0912088623046875, 0.09096089935302734, 0.09147142028808594, 0.09125523376464843, 0.09310569763183593, 0.10366973114013672, 0.09028937530517578, 0.0913620834350586, 0.09063795471191406, 0.09198963165283203, 0.09009011077880859, 0.09199014282226563, 0.09014886474609375, 0.09206169891357421, 0.09041059112548828, 0.09555804443359375, 0.09187052917480469, 0.090908447265625, 0.09234114837646484, 0.09024259185791016, 0.09269705963134765, 0.09040275573730469, 0.0922542724609375, 0.09056230163574219, 0.0913466567993164, 0.09153145599365234, 0.09285257720947265, 0.09170697784423829, 0.09103404998779296, 0.09189542388916015, 0.09099292755126953, 0.09288854217529297, 0.09037910461425781, 0.09227852630615234, 0.09052159881591797, 0.0916883544921875, 0.09114450836181641, 0.09198210906982422, 0.09136029052734375, 0.09187161254882813, 0.09194489288330078, 0.09117558288574219, 0.09296470642089844, 0.09048489379882813, 0.09230038452148437, 0.09058544158935547, 0.09129427337646484, 0.0921190414428711, 0.09133897399902344, 0.09106604766845704, 0.09249187469482421, 0.09364198303222657, 0.09128947448730469, 0.09166118621826172, 0.0912691192626953, 0.09107046508789063, 0.09123379516601562, 0.09105059051513673, 0.09198560333251952, 0.09134246063232422, 0.09273404693603515, 0.09158793640136718, 0.09339561462402343, 0.0913563232421875, 0.09112012481689453, 0.09172415924072265, 0.09143292999267578, 0.0911475830078125, 0.10435788726806641, 0.09028755187988281, 0.0912266845703125, 0.09029017639160156, 0.0921144027709961, 0.09020470428466797, 0.09199411010742188, 0.09095967864990234, 0.09219296264648437, 0.09089024353027343, 0.09495120239257812, 0.092253662109375, 0.09111833953857422, 0.09260147094726562, 0.09063308715820312, 0.09234226989746094, 0.09036956787109375, 0.09147440338134766, 0.0912547836303711, 0.09158656311035156, 0.09188524627685547, 0.0926104965209961, 0.09202291107177735, 0.09128937530517578, 0.09292642974853516, 0.09048831939697266, 0.09298297882080078, 0.09065760040283204, 0.09161231994628906, 0.09045247650146485, 0.09225865936279297, 0.092065185546875, 0.09184111785888673, 0.091580322265625, 0.09169673919677734, 0.09371273803710937, 0.09116687774658203, 0.09214771270751954, 0.0909097900390625, 0.09163459014892578, 0.09096742248535156, 0.09152780914306641, 0.09228902435302734, 0.09166643524169922, 0.0932715835571289, 0.0917418212890625, 0.09217865753173828, 0.09193126678466797, 0.09094758605957032, 0.09138380432128906, 0.09111734771728516, 0.09273065948486328, 0.09158707427978516, 0.09291926574707031, 0.09136022186279297, 0.0912702407836914, 0.09252249908447266, 0.09195315551757813, 0.0931583023071289, 0.09145958709716796, 0.09208627319335938, 0.09133257293701172, 0.09134697723388673, 0.10146009826660156, 0.09062973022460938, 0.09108723449707032, 0.090833984375, 0.090251708984375, 0.09109490966796875, 0.0908909149169922, 0.09223065948486328, 0.0908338851928711, 0.09241603088378907, 0.09315225219726563, 0.09313775634765625, 0.09122988891601562, 0.09091053009033204, 0.09173011016845703, 0.09083740997314453, 0.09170130920410156, 0.09057689666748046, 0.09371033477783203, 0.09081231689453124, 0.09320166778564454, 0.09199702453613282, 0.09194493103027344, 0.09177295684814453, 0.09119676971435547, 0.09367619323730468, 0.0908431396484375, 0.09206121826171874, 0.09158499145507812, 0.09200230407714843, 0.09126834869384766, 0.09211119842529297, 0.092170654296875, 0.09201805114746094, 0.09293875122070312, 0.09115046691894531, 0.09165542602539062, 0.09118386840820313, 0.09130802917480468, 0.09106022644042969, 0.0920821762084961, 0.09167839813232422, 0.09141305541992187, 0.09445756530761719, 0.09153708648681641, 0.09223203277587891, 0.09159423828125, 0.09156835174560547, 0.09279427337646484, 0.0910831069946289, 0.09275590515136718, 0.09085398101806641, 0.09253017425537109, 0.0918075180053711, 0.09201123046875, 0.09254297637939453, 0.09201590728759766, 0.09356502532958984, 0.09128409576416016, 0.0928047332763672, 0.09166639709472656, 0.09145999908447265, 0.09243430328369141]",tokens/s,10.928126976140748,, @@ -3895,7 +3895,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.52 GiB is free. Process 111459 has 13.22 GiB memory in use. Of the allocated memory 13.10 GiB is allocated by PyTorch, and 6.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.52 GiB is free. Process 117507 has 13.22 GiB memory in use. Of the allocated memory 13.10 GiB is allocated by PyTorch, and 6.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,opt,facebook/opt-350m,facebook/opt-350m,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.223-212.873.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.44.2,,0.34.2,,,,1.21.4,,,,0.12.0,,,True,MB,819.63008,652.0832,0.0,249.561088,236.183552,s,1,8.46774609375,8.46774609375,0.0,8.46774609375,8.46774609375,8.46774609375,8.46774609375,[8.46774609375],,kWh,2.948941470418352e-05,3.2457430116192323e-06,9.203062917989557e-06,4.1938220633792314e-05,,MB,1289.990144,756.9408,0.0,341.835776,295.187456,s,10,0.5692166366577148,0.056921663665771484,0.00019196185068508464,0.05693083190917969,0.057161652755737306,0.05721489009857177,0.05725747997283935,"[0.05726812744140625, 0.056790206909179686, 0.05685891342163086, 0.05668412780761719, 0.05702627182006836, 0.05691455841064453, 0.056947105407714846, 0.05714982223510742, 0.056599296569824216, 0.05697820663452149]",tokens/s,4497.408956687604,kWh,1.7516466204091658e-06,1.9305352877369698e-07,1.1612284738922726e-06,3.105928623075135e-06,tokens/kWh,82423014.52070656,MB,1322.528768,794.689536,0.0,379.584512,300.017664,s,10,22.01041723632812,2.201041723632813,0.006431360996584057,2.2009888916015625,2.2076276123046874,2.211374719238281,2.214372404785156,"[2.201778076171875, 2.19366552734375, 2.198657470703125, 2.206794921875, 2.192263671875, 2.2041005859375, 2.215121826171875, 2.20248828125, 2.19534716796875, 2.20019970703125]",tokens/s,28.622810428154306,kWh,6.389133519250488e-05,7.047098314760946e-06,2.3644680259707176e-05,9.4583113766973e-05,tokens/kWh,666080.8413985483,,s,630,22.003552829742453,0.034926274332924494,0.0005763521804872355,0.03481252861022949,0.03523347625732422,0.035550552749633785,0.03739608402252199,"[0.03458649444580078, 0.03475263977050781, 0.03467468643188477, 0.034772991180419925, 0.03588323211669922, 0.03486697769165039, 0.0348600959777832, 0.03484153747558594, 0.03538556671142578, 0.03483606338500977, 0.03547155380249024, 0.03490560150146484, 0.03508889770507812, 0.03470547103881836, 0.03510470581054687, 0.03498416137695313, 0.034993953704833984, 0.034960960388183596, 0.034939327239990235, 0.03522969436645508, 0.03476492691040039, 0.03482611083984375, 0.03505155181884766, 0.034799583435058595, 0.03487526321411133, 0.03465836715698242, 0.03500038528442383, 0.035227840423583984, 0.034778526306152344, 0.03479776000976562, 0.03471987152099609, 0.03475868988037109, 0.03479344177246094, 0.03482444763183594, 0.03561888122558594, 0.03530115127563477, 0.03655680084228516, 0.03482419204711914, 0.03509417724609375, 0.034689342498779294, 0.034782463073730466, 0.0348454704284668, 0.0347589111328125, 0.034774784088134766, 0.034727745056152344, 0.03496777725219727, 0.0349159049987793, 0.03478073501586914, 0.035255390167236327, 0.034862846374511716, 0.03474249649047852, 0.03453740692138672, 0.034518878936767576, 0.034846942901611326, 0.03459612655639648, 0.03492095947265625, 0.03570687866210937, 0.03468479919433594, 0.03482575988769531, 0.03521356964111328, 0.03472828674316406, 0.03470252990722656, 0.03460742568969727, 0.03450281524658203, 0.03496780776977539, 0.03501702499389649, 0.03484902572631836, 0.03482406234741211, 0.03473436737060547, 0.03458428955078125, 0.034840576171875, 0.03465011215209961, 0.03466854476928711, 0.03460121536254883, 0.03483126449584961, 0.03518960189819336, 0.034748416900634765, 0.034649120330810544, 0.034695457458496094, 0.03455408096313477, 0.03463756942749024, 0.0345972785949707, 0.0349136962890625, 0.03467516708374024, 0.03471404647827148, 0.034646305084228515, 0.03459033584594726, 0.034865249633789064, 0.034619102478027346, 0.03678031921386719, 0.03500249481201172, 0.03476601409912109, 0.035678241729736326, 0.03491088104248047, 0.03470336151123047, 0.03478348922729492, 0.03480140686035156, 0.03456723022460938, 0.034864063262939456, 0.03466009521484375, 0.03471747207641602, 0.03483884811401367, 0.03457219314575195, 0.034684513092041014, 0.035506847381591794, 0.03483427047729492, 0.03475487899780273, 0.03472777557373047, 0.03471974563598633, 0.03474563217163086, 0.034777984619140626, 0.03478307342529297, 0.03470892715454101, 0.03475750350952148, 0.03478051376342774, 0.03466864013671875, 0.03472969436645508, 0.034681472778320316, 0.03456195068359375, 0.035108863830566404, 0.03496755218505859, 0.034705406188964845, 0.03476684951782227, 0.03472339248657227, 0.0346190414428711, 0.034863903045654294, 0.03452489471435547, 0.03501289749145508, 0.0349648323059082, 0.03496435165405273, 0.03543427276611328, 0.035194847106933595, 0.035340320587158205, 0.034993377685546875, 0.03470655822753906, 0.034868896484375, 0.0349349136352539, 0.03489769744873047, 0.03465020751953125, 0.035027328491210936, 0.034873184204101564, 0.034998046875, 0.0349200325012207, 0.034595230102539065, 0.03467164611816406, 0.034789985656738284, 0.03491187286376953, 0.0347371826171875, 0.035902175903320316, 0.034729248046875, 0.03478255844116211, 0.03482799911499023, 0.03475088119506836, 0.034730239868164064, 0.03817385482788086, 0.03470003128051758, 0.03495043182373047, 0.03475724792480469, 0.03467695999145508, 0.035019840240478516, 0.034958240509033206, 0.03537424087524414, 0.03491721725463867, 0.03482172775268555, 0.034557567596435544, 0.034694049835205076, 0.03487321472167969, 0.034692832946777344, 0.03466009521484375, 0.034628128051757814, 0.03495110321044922, 0.03456726455688477, 0.03462470245361328, 0.034678558349609374, 0.03481625747680664, 0.03491984176635742, 0.03454524612426758, 0.0347061767578125, 0.034680831909179685, 0.03480899047851563, 0.034820606231689456, 0.034550113677978514, 0.03459664154052734, 0.034916416168212894, 0.03490012741088867, 0.034823646545410154, 0.03464246368408203, 0.03458867263793945, 0.03464191818237305, 0.03482611083984375, 0.035175838470458985, 0.03515055847167969, 0.03470950317382813, 0.03471571350097656, 0.03476678466796875, 0.03492454528808594, 0.03506758499145508, 0.03493900680541992, 0.03488358306884766, 0.034606624603271484, 0.034616222381591795, 0.034608062744140626, 0.034929473876953124, 0.03502905654907226, 0.0351591682434082, 0.03469305419921875, 0.034723968505859376, 0.03472256088256836, 0.0348359375, 0.03466870498657226, 0.0354574089050293, 0.040374305725097655, 0.03490403366088867, 0.03481190490722656, 0.03482419204711914, 0.034885215759277347, 0.03493497467041016, 0.03541219329833984, 0.034985183715820316, 0.035174560546875, 0.034961536407470704, 0.0349285774230957, 0.035037151336669924, 0.034677631378173826, 0.03558355331420898, 0.03474198532104492, 0.03474419021606445, 0.034772991180419925, 0.034759231567382816, 0.03477231979370117, 0.034530303955078126, 0.034649761199951175, 0.0344637451171875, 0.034782367706298827, 0.034648929595947266, 0.03475247955322266, 0.0346399040222168, 0.034672767639160156, 0.034912128448486325, 0.034815841674804685, 0.03509027099609375, 0.034972129821777345, 0.03485606384277344, 0.034809921264648436, 0.034703296661376955, 0.034872318267822264, 0.03558371353149414, 0.03491823959350586, 0.034885345458984376, 0.034810497283935544, 0.034727073669433596, 0.03848777770996094, 0.03473788833618164, 0.03496956634521484, 0.034896705627441404, 0.034890846252441404, 0.03506867218017578, 0.03551420974731445, 0.03495670318603516, 0.03534716796875, 0.03459423828125, 0.03487705612182617, 0.03477561569213867, 0.034672382354736325, 0.03478799819946289, 0.03472959899902344, 0.034613632202148435, 0.034781150817871094, 0.034608577728271486, 0.034654815673828124, 0.03455590438842773, 0.03472505569458008, 0.03461775970458984, 0.03500070571899414, 0.035906913757324216, 0.034957759857177736, 0.034762622833251954, 0.03490832138061523, 0.03468729782104492, 0.03459040069580078, 0.034637409210205077, 0.03466700744628906, 0.034871425628662106, 0.03479667282104492, 0.03570163345336914, 0.03459107208251953, 0.03504707336425781, 0.03459052658081055, 0.03523334503173828, 0.034769664764404296, 0.034627166748046875, 0.034465473175048826, 0.034726593017578126, 0.03474422454833984, 0.03451289749145508, 0.034598911285400394, 0.03457843017578125, 0.034715648651123046, 0.034582527160644534, 0.034539039611816404, 0.034724319458007816, 0.03461702346801758, 0.03454937744140625, 0.034936641693115236, 0.03466976165771484, 0.0347022705078125, 0.03479014587402344, 0.03472588729858399, 0.03450217437744141, 0.03453180694580078, 0.034662654876708984, 0.03469465637207031, 0.03467001724243164, 0.034668479919433594, 0.03493308639526367, 0.03582969665527344, 0.03489388656616211, 0.0347685432434082, 0.034853214263916014, 0.03470764923095703, 0.03474003219604492, 0.034764801025390625, 0.034816001892089846, 0.03469881439208984, 0.034855358123779295, 0.034648063659667966, 0.03471516799926758, 0.03479167938232422, 0.03487772750854492, 0.03467900848388672, 0.034816001892089846, 0.03480137634277344, 0.034932319641113284, 0.03492291259765625, 0.03505500793457031, 0.03512790298461914, 0.034971519470214846, 0.03474444961547852, 0.03480575942993164, 0.035059711456298825, 0.03494086456298828, 0.034949184417724606, 0.035160064697265625, 0.03502284622192383, 0.03496992111206055, 0.03528262329101563, 0.03484201431274414, 0.035011390686035156, 0.035274528503417966, 0.03477004623413086, 0.03496217727661133, 0.034968894958496095, 0.03495151901245117, 0.034964160919189455, 0.0348034553527832, 0.034875423431396484, 0.03491839981079101, 0.034805023193359375, 0.03506454467773437, 0.03499808120727539, 0.03500870513916016, 0.034879615783691406, 0.03505548858642578, 0.034821182250976565, 0.03483334350585938, 0.03486294555664062, 0.03485295867919922, 0.03500185775756836, 0.034781150817871094, 0.03766947174072265, 0.034933761596679686, 0.03533750534057617, 0.0355970573425293, 0.03500572967529297, 0.03474003219604492, 0.03486240005493164, 0.03479779052734375, 0.034957889556884766, 0.0346940803527832, 0.03495766448974609, 0.03493798446655273, 0.03493904113769531, 0.034831775665283206, 0.034894817352294924, 0.03501433563232422, 0.034894302368164064, 0.03515606307983398, 0.035095584869384765, 0.03503606414794922, 0.03501251220703125, 0.03492172622680664, 0.03511715316772461, 0.03506604766845703, 0.03494847869873047, 0.0349071044921875, 0.034949054718017576, 0.03491644668579102, 0.034740222930908206, 0.03504947280883789, 0.03469107055664063, 0.03475046539306641, 0.03478726577758789, 0.0346646728515625, 0.034940769195556644, 0.036071136474609376, 0.04329081726074219, 0.03529264068603516, 0.0352856330871582, 0.03504035186767578, 0.034939327239990235, 0.03558028793334961, 0.035215614318847656, 0.03495004653930664, 0.03496371078491211, 0.0352446403503418, 0.035348575592041014, 0.03513753509521484, 0.03520102310180664, 0.03495964813232422, 0.03469190216064453, 0.03496790313720703, 0.035060287475585934, 0.035175487518310546, 0.03484175872802735, 0.03498688125610352, 0.03479235076904297, 0.0352369270324707, 0.03542931365966797, 0.03488166427612305, 0.03473571014404297, 0.03476044845581055, 0.0349087028503418, 0.03473408126831055, 0.03474198532104492, 0.03753398513793945, 0.03495731353759766, 0.03477913665771484, 0.03479532623291016, 0.034711326599121094, 0.034605472564697266, 0.03467782211303711, 0.03475711822509766, 0.03523583984375, 0.035043327331542966, 0.0353177604675293, 0.03515801620483398, 0.03504537582397461, 0.035001953125, 0.034912094116210934, 0.03495292663574219, 0.03523465728759766, 0.03496918487548828, 0.03473683166503906, 0.03497264099121094, 0.03510095977783203, 0.03479177474975586, 0.03465536117553711, 0.03488240051269531, 0.03492265701293945, 0.03501260757446289, 0.03473408126831055, 0.03474227142333984, 0.0356824951171875, 0.034857887268066406, 0.03497478485107422, 0.034753982543945315, 0.0347259521484375, 0.034912609100341795, 0.03508639907836914, 0.03478214263916016, 0.03517139053344726, 0.034947006225585935, 0.03481315231323242, 0.03477705764770508, 0.0346550407409668, 0.034944671630859375, 0.03484502410888672, 0.0346255989074707, 0.03469686508178711, 0.03479609680175781, 0.034964256286621094, 0.03496166229248047, 0.0351137924194336, 0.03539980697631836, 0.0349666862487793, 0.03522361755371094, 0.03468947219848633, 0.03466473770141602, 0.03469043350219726, 0.03465571212768555, 0.03643695831298828, 0.03580928039550781, 0.03477423858642578, 0.03478607940673828, 0.03462758255004883, 0.034661823272705075, 0.03460563278198242, 0.034805313110351566, 0.03471001434326172, 0.03549177551269531, 0.034977790832519534, 0.03500236892700195, 0.03476483154296875, 0.03478524780273438, 0.03425068664550781, 0.03468489456176758, 0.03462815856933594, 0.03494902420043945, 0.0348304328918457, 0.03494854354858398, 0.034828544616699215, 0.034609470367431644, 0.0346847038269043, 0.03464995193481445, 0.03463520050048828, 0.034589630126953125, 0.034632991790771485, 0.03448704147338867, 0.03473430252075195, 0.03535638427734375, 0.034406688690185545, 0.034528289794921875, 0.03484316635131836, 0.03470764923095703, 0.03459686279296875, 0.03462963104248047, 0.03456409454345703, 0.03531980895996094, 0.03495743942260742, 0.036923263549804686, 0.034731166839599606, 0.03526332855224609, 0.034625537872314455, 0.03458832168579101, 0.03452755355834961, 0.03462790298461914, 0.034777854919433596, 0.034791553497314456, 0.034861438751220705, 0.035178462982177736, 0.034701023101806644, 0.034826751708984374, 0.0348502082824707, 0.03516915130615234, 0.034669921875, 0.03461356735229492, 0.035186622619628904, 0.03507235336303711, 0.03480710220336914, 0.03454140853881836, 0.03453839874267578, 0.03445113754272461, 0.034619712829589845, 0.034797569274902344, 0.03474156951904297, 0.03459347152709961, 0.03536076736450195, 0.03502899169921875, 0.03500032043457031, 0.03455926513671875, 0.03472048187255859, 0.034625537872314455, 0.03541721725463867, 0.03600032043457031, 0.03488729476928711, 0.03476924896240234, 0.03515849685668945, 0.034496097564697265, 0.034777503967285156, 0.03479347229003906, 0.03473625564575195, 0.035675838470458986, 0.03492019271850586, 0.03470175933837891, 0.03465359878540039, 0.03476115036010742, 0.034752670288085936, 0.034680992126464846, 0.03477897644042969, 0.03461759948730469, 0.03485260772705078, 0.03469916915893555, 0.03469321441650391, 0.035198974609375, 0.0349672966003418, 0.034583839416503906, 0.03455209732055664, 0.03457542419433594, 0.0346077766418457, 0.03452617645263672, 0.0344453125, 0.034753982543945315, 0.03516902542114258, 0.035039390563964844, 0.03560208129882812, 0.03489996719360351, 0.0347710075378418, 0.03495481491088867, 0.03492707061767578, 0.034934688568115234, 0.0349306869506836, 0.034598209381103515, 0.034619937896728514, 0.03508563232421875, 0.03453833770751953, 0.034998302459716794, 0.034543582916259766, 0.03462758255004883, 0.034633438110351564, 0.034646305084228515, 0.03468697738647461, 0.034770942687988284, 0.03570211029052734, 0.03810985565185547, 0.034844833374023436, 0.03465760040283203, 0.03626364898681641, 0.03465078353881836, 0.03478953552246094, 0.034599040985107424, 0.03472723388671875, 0.03473823928833008, 0.034824382781982424, 0.03466921615600586, 0.03705846405029297, 0.0348873291015625, 0.034847999572753904, 0.03487353515625, 0.034726558685302736, 0.03477916717529297]",tokens/s,28.631739831961248,, @@ -4304,7 +4304,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 68.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 68.12 MiB is free. Process 78893 has 14.67 GiB memory in use. Of the allocated memory 14.27 GiB is allocated by PyTorch, and 293.36 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 68.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 68.12 MiB is free. Process 67768 has 14.67 GiB memory in use. Of the allocated memory 14.27 GiB is allocated by PyTorch, and 293.36 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neo,EleutherAI/gpt-neo-125m,EleutherAI/gpt-neo-125m,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -4457,7 +4457,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 422.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 328.12 MiB is free. Process 205002 has 14.42 GiB memory in use. Of the allocated memory 14.22 GiB is allocated by PyTorch, and 91.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 422.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 328.12 MiB is free. Process 210719 has 14.42 GiB memory in use. Of the allocated memory 14.22 GiB is allocated by PyTorch, and 91.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neo,EleutherAI/gpt-neo-1.3B,EleutherAI/gpt-neo-1.3B,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -4540,7 +4540,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 28.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 4.12 MiB is free. Process 179850 has 14.73 GiB memory in use. Of the allocated memory 14.49 GiB is allocated by PyTorch, and 137.12 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 28.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 4.12 MiB is free. Process 185682 has 14.73 GiB memory in use. Of the allocated memory 14.49 GiB is allocated by PyTorch, and 137.12 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen,Qwen/Qwen-72B,Qwen/Qwen-72B,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -4658,7 +4658,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 280.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 232.12 MiB is free. Process 21820 has 14.51 GiB memory in use. Of the allocated memory 14.07 GiB is allocated by PyTorch, and 332.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 280.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 232.12 MiB is free. Process 22915 has 14.51 GiB memory in use. Of the allocated memory 14.07 GiB is allocated by PyTorch, and 332.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,EleutherAI/pythia-12b,EleutherAI/pythia-12b,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,1085.898752,8106.47552,0.0,7711.227904,7603.953664,s,1,18.641384765625,18.641384765625,0.0,18.641384765625,18.641384765625,18.641384765625,18.641384765625,[18.641384765625],,kWh,0.00033724906177082703,3.719273954371149e-05,0.0001315656608080057,0.0005060074621225442,,MB,1238.069248,9870.180352,0.0,9462.349824,8756.504576,s,10,16.571733886718746,1.657173388671875,0.005695760134307078,1.6594805297851563,1.6617247680664062,1.6619040100097655,1.662047403564453,"[1.642461181640625, 1.65444482421875, 1.6555794677734375, 1.6585115966796875, 1.6541756591796875, 1.6610345458984375, 1.660449462890625, 1.662083251953125, 1.6616849365234374, 1.6613089599609374]",tokens/s,154.4799124521114,kWh,4.8202024260833316e-05,5.316259229355053e-06,3.2103109015799156e-05,8.562139250598752e-05,tokens/kWh,2989906.990616835,MB,1256.022016,9870.180352,0.0,9462.349824,8756.507136,s,10,82.34577734374999,8.234577734375,0.016255357341610887,8.241509277343749,8.25075615234375,8.251183740234374,8.251525810546875,"[8.2009775390625, 8.2143671875, 8.223310546875, 8.228544921875, 8.2426123046875, 8.24040625, 8.2453994140625, 8.24788671875, 8.251611328125, 8.2506611328125]",tokens/s,7.650665526783283,kWh,0.0002409144947095827,2.657342281018311e-05,0.0001601282392136,0.0004276161567333656,tokens/kWh,147328.39021160468,,s,630,82.34075679016118,0.13069961395263674,0.0019530654863721322,0.13057984161376954,0.13235903625488282,0.1331479217529297,0.1400793441772461,"[0.13893551635742188, 0.12739881896972657, 0.1281716766357422, 0.1281764831542969, 0.12808396911621095, 0.12900732421875, 0.12900901794433595, 0.13369232177734375, 0.13002336120605468, 0.12795699310302736, 0.1282478790283203, 0.12825497436523436, 0.129184326171875, 0.12880531311035157, 0.13193011474609376, 0.13184819030761719, 0.13043276977539062, 0.12846925354003907, 0.12811241149902344, 0.1283865966796875, 0.1292335968017578, 0.13049856567382812, 0.13147955322265625, 0.13099417114257814, 0.12988134765625, 0.1286023406982422, 0.12846339416503907, 0.12916841125488282, 0.1301554870605469, 0.13091839599609376, 0.13138729858398437, 0.1306288604736328, 0.12922966003417968, 0.12963623046875, 0.12919171142578126, 0.1294476776123047, 0.13091845703125, 0.1312957763671875, 0.13156556701660158, 0.13001231384277342, 0.12954217529296874, 0.12889170837402344, 0.13035519409179688, 0.12991693115234376, 0.13176422119140624, 0.13158108520507814, 0.1312407989501953, 0.1301724853515625, 0.12928025817871094, 0.13052044677734376, 0.13092332458496095, 0.13003366088867188, 0.13144985961914063, 0.13050064086914062, 0.1318668212890625, 0.12899346923828126, 0.13061897277832032, 0.13022227478027343, 0.13028025817871094, 0.13159014892578125, 0.132210693359375, 0.13058805847167967, 0.13121315002441405, 0.13991322326660155, 0.1275998077392578, 0.1280213165283203, 0.12786447906494142, 0.12774230194091796, 0.1277675552368164, 0.13016575622558593, 0.13545875549316405, 0.1306542663574219, 0.1288970184326172, 0.12824986267089844, 0.12756537628173828, 0.12829945373535157, 0.12905062866210937, 0.13305145263671875, 0.13252703857421874, 0.1309327392578125, 0.12860826110839843, 0.12896617126464843, 0.1283973388671875, 0.1281212158203125, 0.13095698547363283, 0.132129150390625, 0.13204249572753907, 0.12965846252441407, 0.12837135314941406, 0.1293758087158203, 0.12868858337402345, 0.1301436767578125, 0.13119340515136718, 0.13201773071289064, 0.13140013122558594, 0.12929776000976562, 0.12912503051757812, 0.12901376342773438, 0.13061695861816405, 0.1302408905029297, 0.13229055786132812, 0.13236972045898437, 0.13076141357421875, 0.12902787780761718, 0.1294844512939453, 0.12975765991210939, 0.13076693725585936, 0.1304470977783203, 0.13214288330078125, 0.13153660583496093, 0.12992530822753906, 0.13021650695800782, 0.13024575805664063, 0.13055442810058593, 0.131830078125, 0.13075421142578125, 0.13242594909667968, 0.13120729064941405, 0.1299488983154297, 0.13001808166503906, 0.1299571228027344, 0.1305771484375, 0.13168412780761718, 0.1315616912841797, 0.13172735595703125, 0.13051235961914062, 0.14113037109375, 0.12770079803466797, 0.1281025848388672, 0.12817768859863282, 0.12807420349121093, 0.12819046020507813, 0.13012991333007812, 0.13605679321289063, 0.1302139129638672, 0.12912435913085937, 0.12855705261230468, 0.1283108215332031, 0.12955081176757813, 0.12886416625976563, 0.1329439697265625, 0.13226393127441408, 0.13094912719726562, 0.12882534790039063, 0.12912640380859375, 0.12847718811035155, 0.12873274230957032, 0.13159059143066407, 0.13233561706542968, 0.13138893127441406, 0.12994309997558592, 0.1294775695800781, 0.12870761108398437, 0.1293076171875, 0.13037132263183593, 0.13155743408203124, 0.13206716918945313, 0.13080812072753906, 0.12963145446777344, 0.12936402893066407, 0.12912669372558594, 0.1300955810546875, 0.13120716857910156, 0.13225091552734375, 0.13093244934082032, 0.13088870239257813, 0.12988211059570312, 0.12900146484375, 0.12990821838378908, 0.1315527648925781, 0.13109075927734376, 0.1322421112060547, 0.13127430725097655, 0.1305072021484375, 0.12969334411621095, 0.12982716369628905, 0.13123583984375, 0.1306480712890625, 0.1312522277832031, 0.1315635223388672, 0.13078732299804688, 0.13080986022949218, 0.12949913024902343, 0.13116578674316406, 0.13121168518066406, 0.13136607360839844, 0.13177305603027345, 0.13169273376464843, 0.13022207641601563, 0.14015536499023437, 0.12814131164550782, 0.12783926391601563, 0.12821160888671876, 0.12814118957519532, 0.12914837646484376, 0.13038485717773438, 0.13575932312011718, 0.13007516479492187, 0.12905673217773436, 0.12838899230957032, 0.12839543151855468, 0.12823904418945312, 0.1293679656982422, 0.13350775146484375, 0.13243382263183595, 0.12989244079589843, 0.1287118377685547, 0.1287560272216797, 0.12811318969726562, 0.12890521240234376, 0.13169218444824218, 0.13325045776367186, 0.13136370849609375, 0.1295311737060547, 0.12929922485351564, 0.128901123046875, 0.1284931182861328, 0.13039231872558593, 0.13224159240722655, 0.1323702392578125, 0.1308605499267578, 0.13014291381835938, 0.1295064392089844, 0.1300693817138672, 0.12980224609375, 0.1313804473876953, 0.13241629028320312, 0.13214720153808593, 0.13099203491210937, 0.1301973114013672, 0.1295870361328125, 0.13003411865234374, 0.13066444396972657, 0.13156556701660158, 0.13164544677734374, 0.132347900390625, 0.13004595947265624, 0.13087129211425783, 0.12948069763183595, 0.13137510681152345, 0.13122122192382812, 0.13173788452148438, 0.13170384216308595, 0.1305425567626953, 0.13078486633300782, 0.1300413818359375, 0.1302959747314453, 0.13059756469726563, 0.1317069091796875, 0.13183180236816405, 0.13174374389648438, 0.13154917907714844, 0.14049481201171876, 0.1283507843017578, 0.12919430541992188, 0.1283706817626953, 0.12837673950195314, 0.12830915832519532, 0.13027468872070314, 0.13690963745117188, 0.13102284240722656, 0.1295626220703125, 0.12902400207519532, 0.12853363037109375, 0.12835311889648438, 0.13054556274414061, 0.13333721923828126, 0.13229055786132812, 0.13028457641601562, 0.13049728393554688, 0.1287006378173828, 0.12937152099609375, 0.12869007873535157, 0.131152099609375, 0.13243843078613282, 0.131874755859375, 0.1308625030517578, 0.13033139038085936, 0.12958924865722657, 0.13054566955566407, 0.12919923400878905, 0.13162342834472657, 0.13164991760253905, 0.13192752075195313, 0.13001373291015625, 0.13120863342285155, 0.13006402587890625, 0.1314231414794922, 0.12982272338867187, 0.13268569946289063, 0.13119850158691407, 0.13076130676269532, 0.13126783752441407, 0.13021852111816407, 0.13056431579589844, 0.13036338806152345, 0.13152870178222656, 0.13144009399414064, 0.13100086975097655, 0.13126246643066405, 0.130616455078125, 0.13119541931152343, 0.13120547485351564, 0.13072515869140625, 0.13122396850585938, 0.1311808624267578, 0.13150822448730468, 0.1304289245605469, 0.13058253479003906, 0.1306409912109375, 0.13108901977539061, 0.13035871887207032, 0.131402587890625, 0.13177037048339843, 0.13162413024902345, 0.13996461486816406, 0.12824986267089844, 0.12899690246582032, 0.12820938110351562, 0.1282027587890625, 0.12900338745117187, 0.13072543334960937, 0.135719482421875, 0.13090815734863281, 0.12924844360351562, 0.12854150390625, 0.12843008422851562, 0.12855705261230468, 0.13024870300292968, 0.13279641723632812, 0.1326755828857422, 0.13027122497558594, 0.12992889404296876, 0.1286961212158203, 0.1288605194091797, 0.12978378295898438, 0.13254060363769532, 0.1318338623046875, 0.13180928039550782, 0.1308112030029297, 0.1294015350341797, 0.12876153564453124, 0.12974432373046876, 0.13192076110839843, 0.1318666229248047, 0.13188710021972655, 0.13182917785644532, 0.13013241577148438, 0.1293429718017578, 0.1296328887939453, 0.13080166625976564, 0.13147048950195311, 0.13216444396972657, 0.1321697235107422, 0.13107814025878906, 0.12951513671875, 0.12986105346679688, 0.1301329345703125, 0.13050221252441407, 0.13167864990234374, 0.13249932861328126, 0.131217529296875, 0.1313846435546875, 0.12980653381347657, 0.1308134765625, 0.12995606994628905, 0.13157452392578126, 0.13173146057128907, 0.13190943908691405, 0.13118278503417968, 0.13089727783203126, 0.1306956787109375, 0.13019354248046874, 0.13020159912109375, 0.13052291870117189, 0.13164361572265626, 0.1316532440185547, 0.13308735656738283, 0.14012620544433593, 0.12829696655273437, 0.12820457458496093, 0.12826751708984374, 0.12816073608398437, 0.12899737548828125, 0.13085081481933594, 0.13553567504882813, 0.1309697265625, 0.12925340270996094, 0.12850051879882812, 0.12850994873046875, 0.12881846618652343, 0.13050726318359376, 0.13317747497558594, 0.13320101928710937, 0.13059170532226563, 0.130110595703125, 0.12919692993164061, 0.1286264953613281, 0.12942970275878907, 0.13166160583496095, 0.1332943115234375, 0.131778564453125, 0.13052674865722655, 0.1299471435546875, 0.12895443725585937, 0.12917648315429686, 0.13096754455566406, 0.1318968963623047, 0.13240777587890626, 0.13143650817871094, 0.13108837890625, 0.129544189453125, 0.1289581756591797, 0.13014653015136718, 0.13153286743164064, 0.13238636779785157, 0.13170938110351563, 0.13111488342285157, 0.13149606323242188, 0.12948419189453125, 0.13031024169921876, 0.13073049926757813, 0.132927490234375, 0.13182879638671874, 0.13113235473632812, 0.13153689575195313, 0.13080342102050782, 0.12925570678710938, 0.1304289245605469, 0.13183193969726562, 0.13213848876953124, 0.13127308654785155, 0.13165568542480469, 0.13163427734375, 0.13061827087402345, 0.13018464660644533, 0.13065887451171876, 0.131842041015625, 0.13173554992675782, 0.13281607055664063, 0.13068780517578124, 0.14088627624511718, 0.12847657775878907, 0.12828726196289061, 0.12817864990234376, 0.1282721252441406, 0.13027101135253907, 0.13066493225097656, 0.13762098693847657, 0.13073458862304688, 0.12884378051757814, 0.12850143432617187, 0.1285246124267578, 0.12947251892089845, 0.13043302917480468, 0.1336376953125, 0.13255520629882814, 0.13028483581542968, 0.1293585205078125, 0.1285509490966797, 0.13010943603515626, 0.1297097930908203, 0.13242396545410157, 0.13242274475097657, 0.13157635498046874, 0.12990882873535156, 0.12923695373535157, 0.12992953491210937, 0.1295474548339844, 0.13100099182128908, 0.13325328063964845, 0.13212608337402343, 0.13156005859375, 0.12980429077148437, 0.12916940307617186, 0.12946636962890626, 0.13043096923828126, 0.13245216369628907, 0.13192965698242187, 0.13177468872070314, 0.13058908081054688, 0.13014630126953125, 0.1291673583984375, 0.1298303680419922, 0.13090255737304687, 0.13227801513671875, 0.13193855285644532, 0.13237452697753907, 0.13075833129882813, 0.13074464416503906, 0.12919375610351563, 0.1314959411621094, 0.13160418701171875, 0.13327615356445313, 0.13225779724121095, 0.13110067749023438, 0.13025074768066405, 0.13086309814453126, 0.13005413818359374, 0.13144677734375, 0.131693603515625, 0.13102998352050782, 0.1325137939453125, 0.1304718780517578, 0.1415303955078125, 0.1285926055908203, 0.12825190734863282, 0.1283656005859375, 0.12828361511230468, 0.1293097229003906, 0.13108460998535157, 0.13580560302734376, 0.13170396423339845, 0.12993417358398437, 0.12859187316894533, 0.12940083312988282, 0.12825907897949218, 0.12942437744140625, 0.13311180114746093, 0.1330708465576172, 0.13054937744140624, 0.12996237182617187, 0.12910797119140624, 0.12851405334472657, 0.12912828063964843, 0.1317111358642578, 0.1323970489501953, 0.13160652160644531, 0.13123487854003907, 0.12992813110351562, 0.12900965881347656, 0.12895436096191407, 0.13127679443359375, 0.13221612548828124, 0.131932861328125, 0.13253631591796874, 0.13065414428710936, 0.13003372192382812, 0.12951962280273438, 0.130702392578125, 0.13098080444335938, 0.13345587158203126, 0.13215335083007812, 0.13170687866210937, 0.1302650909423828, 0.13012991333007812, 0.12970188903808594, 0.13027122497558594, 0.1325972442626953, 0.13206723022460937, 0.1317255096435547, 0.1314349060058594, 0.13033401489257812, 0.13015315246582032, 0.13065362548828124, 0.13179058837890625, 0.13291807556152344, 0.1315450897216797, 0.1315691223144531, 0.13074858093261718, 0.13038812255859375, 0.12999293518066407, 0.13138534545898437, 0.13204071044921875, 0.1320028533935547, 0.1316414031982422, 0.13171580505371094, 0.14180744934082032, 0.12851321411132813, 0.12808647155761718, 0.12903887939453124, 0.12826214599609376, 0.12901951599121095, 0.13128533935546874, 0.13619609069824218, 0.13124610900878905, 0.1289318389892578, 0.129225830078125, 0.12940176391601563, 0.12878224182128906, 0.1296315155029297, 0.1337425537109375, 0.1322852783203125, 0.13041664123535157, 0.12938380432128907, 0.12946905517578125, 0.12852345275878907, 0.1294569549560547, 0.13196493530273437, 0.13230047607421874, 0.1317952575683594, 0.13015449523925782, 0.12972393798828125, 0.12955859375, 0.12914073181152344, 0.13074269104003905, 0.1320939483642578, 0.1321553955078125, 0.13203660583496094, 0.13028099060058593, 0.12952943420410157, 0.13014309692382814, 0.13009642028808593, 0.13128572082519532, 0.132115966796875, 0.13234768676757813, 0.1316829833984375, 0.12997843933105468, 0.13084262084960938, 0.13017219543457031, 0.13082083129882813, 0.13194444274902345, 0.1323534698486328, 0.13208018493652343, 0.13037158203125, 0.13044940185546874, 0.13119488525390624, 0.1305432891845703, 0.13164166259765625, 0.13235784912109375, 0.13192218017578125, 0.13110176086425782, 0.13091119384765626, 0.13022547912597657, 0.13100898742675782, 0.13138665771484376, 0.1316790466308594, 0.13319180297851563, 0.1308078155517578, 0.1314009246826172]",tokens/s,7.651132009941381,, @@ -4777,7 +4777,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 110.12 MiB is free. Process 164769 has 14.63 GiB memory in use. Of the allocated memory 14.47 GiB is allocated by PyTorch, and 51.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 110.12 MiB is free. Process 170741 has 14.63 GiB memory in use. Of the allocated memory 14.47 GiB is allocated by PyTorch, and 51.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,Deci/DeciCoder-1b,Deci/DeciCoder-1b,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -4865,7 +4865,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 448.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 436.12 MiB is free. Process 171195 has 14.31 GiB memory in use. Of the allocated memory 14.14 GiB is allocated by PyTorch, and 61.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 448.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 436.12 MiB is free. Process 177130 has 14.31 GiB memory in use. Of the allocated memory 14.14 GiB is allocated by PyTorch, and 61.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,togethercomputer/RedPajama-INCITE-Base-7B-v0.1,togethercomputer/RedPajama-INCITE-Base-7B-v0.1,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,,True,MB,1171.74272,5037.228032,0.0,4634.70592,4621.451776,s,1,14.017248046875,14.017248046875,0.0,14.017248046875,14.017248046875,14.017248046875,14.017248046875,[14.017248046875],,kWh,0.00019905916797079802,2.1950498733250368e-05,6.75647762739956e-05,0.000288574442978044,,MB,1344.139264,5911.740416,0.0,5496.635392,5336.871424,s,10,9.381483093261718,0.9381483093261718,0.0052608552739976085,0.9388148498535156,0.9424588684082031,0.9425970489501952,0.9427075933837891,"[0.9242491455078125, 0.9354083251953125, 0.9369375610351562, 0.94210546875, 0.937936279296875, 0.9377211303710937, 0.9427352294921875, 0.9422683715820312, 0.9396934204101562, 0.9424281616210938]",tokens/s,272.87796338286097,kWh,2.7226402124618733e-05,3.0025769358617312e-06,1.81140296427273e-05,4.834300870320775e-05,tokens/kWh,5295491.67226353,MB,1387.823104,5911.740416,0.0,5496.635392,5336.873984,s,10,44.90041455078125,4.490041455078126,0.02078786908101345,4.492143310546876,4.5145380859375,4.51455126953125,4.51456181640625,"[4.45762060546875, 4.46370556640625, 4.47185888671875, 4.4781611328125, 4.48135107421875, 4.502935546875, 4.50408837890625, 4.514564453125, 4.51159375, 4.51453515625]",tokens/s,14.031050855610376,kWh,0.0001319513217583023,1.4555179242530698e-05,8.755458266987098e-05,0.00023406108367070397,tokens/kWh,269160.507214576,,s,630,44.897264289855926,0.07126549887278723,0.0016062593425575284,0.07099377822875977,0.07233937225341798,0.07277121505737305,0.08117747116088868,"[0.08214636993408203, 0.07056832122802735, 0.06992720031738281, 0.06955622100830078, 0.06930223846435547, 0.06932691192626952, 0.07043714904785156, 0.06996371459960937, 0.06969548797607422, 0.07021071624755859, 0.06988428497314453, 0.06931283569335937, 0.0699345932006836, 0.07270467376708985, 0.0723329620361328, 0.07126582336425781, 0.07043360137939453, 0.07017078399658203, 0.06964822387695313, 0.06995097351074218, 0.0706012191772461, 0.06988800048828125, 0.06964444732666016, 0.06980592346191407, 0.07027871704101563, 0.06983468627929687, 0.07130915069580078, 0.07179507446289063, 0.07138333129882812, 0.07097865295410156, 0.07072454071044922, 0.07038579559326172, 0.07043875122070313, 0.07036275482177734, 0.07013426971435546, 0.06976831817626954, 0.06990300750732421, 0.07062537384033203, 0.07014399719238282, 0.0708647689819336, 0.07128844451904297, 0.07132825469970704, 0.07147519683837891, 0.07083622741699219, 0.07126127624511719, 0.07061090850830078, 0.07077983856201171, 0.07093641662597656, 0.07054470062255859, 0.07043363189697266, 0.07023001861572266, 0.07124281311035156, 0.07052992248535156, 0.07170496368408204, 0.07093548583984376, 0.07118515014648437, 0.07137078094482421, 0.07116966247558594, 0.07073827362060547, 0.07096889495849609, 0.0709986572265625, 0.07060054779052734, 0.07054700469970702, 0.08119439697265625, 0.07060540771484375, 0.06990217590332032, 0.06957209777832031, 0.069493408203125, 0.0695173110961914, 0.06975897979736329, 0.07006208038330078, 0.06951936340332031, 0.06935475158691407, 0.07020211029052735, 0.06951321411132813, 0.07071324920654297, 0.07357449340820313, 0.07270333099365234, 0.07138575744628907, 0.07054252624511718, 0.07004780578613282, 0.06967295837402344, 0.06985616302490234, 0.07028508758544921, 0.06976313781738282, 0.0700047378540039, 0.07044271850585937, 0.07007395172119141, 0.06980473327636719, 0.07083401489257812, 0.07245155334472657, 0.07219884490966796, 0.07211203002929688, 0.0717674560546875, 0.07054592132568359, 0.07018515014648438, 0.07008048248291016, 0.07062735748291016, 0.07002025604248047, 0.07044950103759766, 0.07038963317871094, 0.07046617889404297, 0.07050870513916016, 0.07171379089355469, 0.07149040222167968, 0.07239190673828125, 0.0715865249633789, 0.07091584014892578, 0.07045318603515625, 0.07047615814208984, 0.07118409729003906, 0.07057027435302735, 0.06994944000244141, 0.07048799896240235, 0.07009506988525391, 0.0704776611328125, 0.070698974609375, 0.07184925079345703, 0.07185081481933593, 0.07206668853759765, 0.07136492919921875, 0.07069856262207032, 0.07045164489746093, 0.07117132568359374, 0.07077891540527344, 0.07053724670410157, 0.08113603210449219, 0.07055206298828125, 0.07010508728027344, 0.06952451324462891, 0.07062828826904297, 0.07012764739990235, 0.06956646728515625, 0.06957868957519531, 0.07009286499023437, 0.06981581115722656, 0.07028947448730469, 0.0695607681274414, 0.07073792266845703, 0.07359510040283203, 0.07241641235351562, 0.07120140838623047, 0.0709421157836914, 0.07030435180664063, 0.06977126312255859, 0.06964390563964844, 0.07114790344238281, 0.07020655822753906, 0.07044802856445312, 0.06997615814208985, 0.07006134033203125, 0.07058303833007812, 0.07171881866455078, 0.07226979064941406, 0.07175065612792969, 0.07161913299560548, 0.07144662475585938, 0.0706949462890625, 0.07021417236328124, 0.07042845153808594, 0.0707192611694336, 0.07048623657226563, 0.07082109069824219, 0.07050873565673828, 0.07035327911376953, 0.07074428558349609, 0.07141785430908203, 0.07194409942626953, 0.07202323150634765, 0.07201270294189453, 0.07061113739013672, 0.07101945495605469, 0.07058303833007812, 0.07026028442382813, 0.07051526641845703, 0.0711436767578125, 0.07064351654052735, 0.07040755462646485, 0.07099974060058593, 0.07106870269775391, 0.07184349060058594, 0.0719832992553711, 0.07241276550292969, 0.07131523132324219, 0.07104166412353516, 0.07055350494384766, 0.07055165100097656, 0.07058822631835937, 0.0708650894165039, 0.08226515197753906, 0.071000732421875, 0.07042281341552735, 0.0697853775024414, 0.07019747161865235, 0.07010018920898438, 0.0697147216796875, 0.06992829132080078, 0.06964044952392578, 0.06964995574951172, 0.07002729797363282, 0.07084480285644532, 0.07100454711914063, 0.07351920318603515, 0.0727162857055664, 0.07164514923095704, 0.07092192077636719, 0.07034095764160156, 0.07044505310058594, 0.07016448211669922, 0.06971392059326172, 0.06978479766845704, 0.0708815689086914, 0.07023859405517578, 0.06974681854248047, 0.07064575958251954, 0.07191331481933594, 0.07231504058837891, 0.07221862030029297, 0.07201900482177734, 0.0711994857788086, 0.07069305419921874, 0.070508544921875, 0.07023616027832032, 0.07009254455566406, 0.06985958099365235, 0.07106073760986328, 0.07050342559814453, 0.07040150451660156, 0.0706728286743164, 0.07173001861572266, 0.07200601959228516, 0.07197145843505859, 0.07152611541748047, 0.07143443298339844, 0.07104694366455078, 0.07161885070800782, 0.07073590087890624, 0.07050351715087891, 0.07047670745849609, 0.0710843505859375, 0.07048105621337891, 0.07054137420654297, 0.07113168334960937, 0.07181622314453125, 0.0720208969116211, 0.07158930969238281, 0.07130361938476562, 0.0711943359375, 0.07193987274169922, 0.07143097686767579, 0.07068367767333984, 0.07060765075683594, 0.08222509002685546, 0.07104889678955079, 0.06995897674560547, 0.06984134674072266, 0.06998483276367187, 0.06953984069824219, 0.06980812835693359, 0.0701112289428711, 0.06982806396484376, 0.07022998046875, 0.06993977355957032, 0.07057965087890625, 0.071159423828125, 0.07402137756347656, 0.07232502746582031, 0.07119312286376953, 0.07070476531982423, 0.07024668884277344, 0.07129535675048829, 0.07042633819580078, 0.07019725036621094, 0.06995782470703125, 0.07122720336914062, 0.0702852783203125, 0.06999657440185547, 0.07072895812988281, 0.07226652526855469, 0.07207933044433594, 0.07203424072265625, 0.07121414184570313, 0.07080989074707031, 0.07113187408447266, 0.07138285064697265, 0.07034899139404296, 0.07009279632568359, 0.07053635406494141, 0.07045616149902344, 0.07014809417724609, 0.0706170883178711, 0.07180691528320313, 0.07170873260498047, 0.07239807891845704, 0.07176070404052734, 0.07104067230224609, 0.07157904052734375, 0.0712867202758789, 0.07093138885498047, 0.07046963500976562, 0.07049996948242188, 0.07084038543701172, 0.07121965026855469, 0.07048998260498048, 0.07201382446289062, 0.07142121887207031, 0.07181177520751954, 0.07205590057373047, 0.07187500762939453, 0.07111116790771485, 0.07116595458984375, 0.07128678131103515, 0.07072077178955079, 0.07039871978759765, 0.07118224334716797, 0.0809559326171875, 0.07100729370117187, 0.07051324462890625, 0.06985113525390625, 0.06994290924072266, 0.07004608154296875, 0.07039100646972657, 0.07075714874267579, 0.07047372436523437, 0.07054541015625, 0.07009279632568359, 0.07064575958251954, 0.07183094024658203, 0.07430000305175781, 0.07283916473388671, 0.0714097900390625, 0.07092425537109374, 0.07046662139892577, 0.0703595199584961, 0.07079974365234375, 0.07055750274658203, 0.07057017517089843, 0.07148908996582032, 0.07051718139648437, 0.0709591064453125, 0.07168195343017578, 0.07273686218261718, 0.07254386901855468, 0.07210393524169922, 0.07172748565673828, 0.0712841567993164, 0.07068470764160156, 0.07196521759033203, 0.07076454162597656, 0.07050383758544922, 0.07054918670654296, 0.07058236694335937, 0.07128761291503906, 0.07136685180664062, 0.07219181060791016, 0.07286784362792968, 0.07173529815673828, 0.07162258911132813, 0.07227961730957032, 0.07123538970947266, 0.07149967956542969, 0.07108892822265625, 0.07093862152099609, 0.07114093017578126, 0.07074861145019531, 0.07076863861083985, 0.07240473937988282, 0.07188636779785157, 0.07304656219482422, 0.07235398101806641, 0.07165955352783203, 0.0723351058959961, 0.07139350128173828, 0.07176834869384766, 0.07216432189941406, 0.07124230194091796, 0.07092777252197266, 0.0712609634399414, 0.08091648101806641, 0.07091410827636718, 0.07054351806640625, 0.07002294158935547, 0.07017068481445313, 0.07004668426513672, 0.07011395263671875, 0.07010851287841798, 0.07084537506103515, 0.0704738540649414, 0.07062723541259766, 0.07074201965332032, 0.07184585571289062, 0.07399836730957031, 0.07312134552001953, 0.07206883239746094, 0.07091887664794921, 0.07055372619628907, 0.07026233673095703, 0.0709202880859375, 0.07057635498046876, 0.07041382598876954, 0.07080207824707031, 0.07050822448730469, 0.07055580902099609, 0.07150128173828126, 0.07345206451416016, 0.07291085052490234, 0.07235363006591797, 0.07154704284667969, 0.0711305923461914, 0.07081629180908203, 0.07076454162597656, 0.07158988952636719, 0.07087299346923828, 0.07141999816894531, 0.07101849365234375, 0.07069193267822266, 0.07100035095214843, 0.07217625427246094, 0.07267961883544923, 0.07262802886962891, 0.07275917053222657, 0.07165145874023438, 0.07134822082519532, 0.0710054702758789, 0.0708512954711914, 0.0712640609741211, 0.07116786956787109, 0.07081401824951172, 0.07154710388183594, 0.07094866943359375, 0.07143603515625, 0.07277133178710937, 0.07198767852783203, 0.07231283569335938, 0.07175904083251954, 0.07231520080566406, 0.07164979553222656, 0.07119657897949219, 0.07204386901855468, 0.07117286682128907, 0.07111804962158202, 0.0825099868774414, 0.07117753601074218, 0.07071971130371094, 0.07031874847412109, 0.07078460693359374, 0.07045903778076172, 0.07011382293701172, 0.07094480133056641, 0.07051673889160157, 0.07068262481689454, 0.0709265594482422, 0.07054694366455078, 0.07186255645751953, 0.07399382019042969, 0.07293788909912109, 0.07209149169921875, 0.0715777587890625, 0.07135350036621094, 0.0708878402709961, 0.07134838104248047, 0.07096348571777343, 0.07063334655761719, 0.07080973052978516, 0.07100621032714843, 0.07068409729003906, 0.07107231903076172, 0.07233526611328125, 0.07277107238769531, 0.0721945571899414, 0.07254761505126953, 0.07161529541015625, 0.07164934539794922, 0.0713604507446289, 0.07085584259033204, 0.07077958679199219, 0.07105782318115235, 0.07117990112304687, 0.07095708465576171, 0.07126640319824219, 0.07178854370117188, 0.07230226898193359, 0.07237036895751953, 0.07222198486328125, 0.0723683853149414, 0.07163164520263672, 0.07128656005859375, 0.07211551666259766, 0.07114115142822265, 0.07091705322265625, 0.07139328002929687, 0.07156326293945313, 0.07104467010498047, 0.07167635345458985, 0.071804931640625, 0.07229609680175782, 0.073210205078125, 0.07190528106689453, 0.0723394546508789, 0.07141580963134765, 0.0717127685546875, 0.07112406158447265, 0.07168230438232422, 0.07135465240478515, 0.08324569702148438, 0.07128601837158204, 0.070662109375, 0.0702311019897461, 0.07076236724853516, 0.07041455841064453, 0.07012242889404297, 0.07036524963378907, 0.07075305938720704, 0.07056301116943359, 0.07076777648925782, 0.07050614166259765, 0.07185209655761719, 0.0748625259399414, 0.07318553924560547, 0.07165497589111328, 0.07150614166259765, 0.07072764587402344, 0.07113442993164062, 0.07059126281738282, 0.07065325164794922, 0.07039810943603515, 0.07073436737060547, 0.070204833984375, 0.07149622344970703, 0.07130461120605469, 0.07280502319335938, 0.07266508483886719, 0.07271424102783203, 0.07203868865966796, 0.0710994873046875, 0.07144322967529297, 0.07093222045898437, 0.07086294555664062, 0.0708629150390625, 0.07078701019287109, 0.0709365463256836, 0.07056492614746093, 0.07229519653320313, 0.0714651870727539, 0.07286780548095703, 0.07225363159179687, 0.07188614654541016, 0.0716746597290039, 0.07219094085693359, 0.07177471923828126, 0.07102259063720703, 0.07151427459716797, 0.07098287963867188, 0.07106854248046875, 0.07100214385986328, 0.0713338851928711, 0.07157689666748047, 0.07233936309814454, 0.07268428802490234, 0.07170480346679688, 0.07191324615478516, 0.07198515319824218, 0.07181520080566406, 0.07126934051513673, 0.07171155548095703, 0.0715265884399414, 0.0716759033203125, 0.08157004547119141, 0.07102457427978516, 0.070723388671875, 0.07035603332519531, 0.07088137817382813, 0.07047779083251954, 0.07085148620605469, 0.07056563568115234, 0.07066957092285156, 0.07085356903076172, 0.07045532989501953, 0.07049212646484375, 0.07219830322265625, 0.07426217651367187, 0.07274425506591797, 0.071480224609375, 0.07119033813476562, 0.07127468872070312, 0.07082189178466797, 0.07082592010498047, 0.07079347229003906, 0.0706819839477539, 0.07081619262695313, 0.0708136978149414, 0.0713543701171875, 0.07148544311523437, 0.07342694091796875, 0.0721981430053711, 0.07180902099609375, 0.07150118255615234, 0.07145116424560546, 0.072569091796875, 0.0712143325805664, 0.07145942687988281, 0.07093609619140626, 0.07081571197509766, 0.07096896362304687, 0.0711668472290039, 0.07110041809082031, 0.07235727691650391, 0.07255510711669921, 0.07271206665039062, 0.07198323059082032, 0.07178841400146484, 0.071823486328125, 0.07136870574951172, 0.07190528106689453, 0.07151628875732421, 0.0709468765258789, 0.07192095947265625, 0.07110707092285157, 0.0712499237060547, 0.07174505615234375, 0.07238639831542969, 0.07205919647216796, 0.07215545654296875, 0.07167747497558594, 0.07230716705322265, 0.071731201171875, 0.07166067504882813, 0.07251033782958985, 0.07140966033935547, 0.07098889923095703]",tokens/s,14.032035358161936,, @@ -4911,7 +4911,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 83834 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 72807 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,internlm,internlm/internlm-20b,internlm/internlm-20b,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -4989,7 +4989,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 34.12 MiB is free. Process 88661 has 14.71 GiB memory in use. Of the allocated memory 14.56 GiB is allocated by PyTorch, and 28.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 34.12 MiB is free. Process 77676 has 14.71 GiB memory in use. Of the allocated memory 14.56 GiB is allocated by PyTorch, and 28.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,codegen,Salesforce/codegen-16B-nl,Salesforce/codegen-16B-nl,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -5070,7 +5070,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 302.12 MiB is free. Process 156530 has 14.44 GiB memory in use. Of the allocated memory 14.20 GiB is allocated by PyTorch, and 132.89 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 302.12 MiB is free. Process 162486 has 14.44 GiB memory in use. Of the allocated memory 14.20 GiB is allocated by PyTorch, and 132.89 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,huggyllama/llama-30b,huggyllama/llama-30b,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -5117,7 +5117,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 54.12 MiB is free. Process 151658 has 14.69 GiB memory in use. Of the allocated memory 14.37 GiB is allocated by PyTorch, and 203.94 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 54.12 MiB is free. Process 157531 has 14.69 GiB memory in use. Of the allocated memory 14.37 GiB is allocated by PyTorch, and 203.94 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-13b-hf,meta-llama/Llama-2-13b-hf,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.224-212.876.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.0,,0.34.2,,,,1.22.0,,,,0.12.0,,,True,MB,822.530048,8535.277568,0.0,8132.755456,7824.681472,s,1,19.363349609375,19.363349609375,0.0,19.363349609375,19.363349609375,19.363349609375,19.363349609375,[19.363349609375],,kWh,0.00035991572860831034,3.9689742300941714e-05,0.0001184187058459879,0.00051802417675524,,MB,1317.277696,9539.813376,0.0,9124.708352,8500.500992,s,10,17.646905761718752,1.764690576171875,0.007826338434277208,1.766254577636719,1.7710271606445311,1.771575653076172,1.7720144470214845,"[1.744111328125, 1.758710693359375, 1.764703857421875, 1.7649429931640626, 1.764416015625, 1.767566162109375, 1.76920263671875, 1.77022265625, 1.7721241455078125, 1.7709052734375]",tokens/s,145.0679249136912,kWh,5.141618405000069e-05,5.67070418568476e-06,3.420905514500139e-05,9.129594338068683e-05,tokens/kWh,2804067.6345555508,MB,1341.882368,9541.910528,0.0,9126.805504,8500.503552,s,10,83.08388867187502,8.308388867187501,0.013627102147594676,8.31012353515625,8.32286650390625,8.324238916015625,8.325336845703125,"[8.2831162109375, 8.291119140625, 8.296341796875, 8.3096630859375, 8.305431640625, 8.318150390625, 8.310583984375, 8.3213095703125, 8.325611328125, 8.3225615234375]",tokens/s,7.582697561112871,kWh,0.00024312225124250062,2.6818406767784186e-05,0.00016131454571819833,0.00043125520372848305,tokens/kWh,146085.19376768984,,s,630,83.08128448486335,0.13187505473787822,0.0017931074401041948,0.13183060455322265,0.13309696197509766,0.133570947265625,0.1425301985168457,"[0.14336204528808594, 0.13153074645996093, 0.1298570556640625, 0.12978839111328125, 0.13047398376464844, 0.1301360626220703, 0.12976332092285156, 0.13173286437988282, 0.13229116821289064, 0.13057388305664064, 0.1300455322265625, 0.13068582153320313, 0.13145497131347655, 0.13009446716308593, 0.1302487335205078, 0.13198602294921874, 0.13237615966796876, 0.13121133422851564, 0.1311130828857422, 0.13177436828613281, 0.13055007934570312, 0.13019664001464842, 0.13174844360351562, 0.13151068115234374, 0.13099810791015626, 0.13145021057128906, 0.13109111022949219, 0.13215948486328125, 0.13084669494628906, 0.12996528625488282, 0.13167709350585938, 0.13198326110839845, 0.13163725280761718, 0.13153280639648438, 0.13204031372070313, 0.13062384033203125, 0.1300951385498047, 0.13071907043457032, 0.1315681915283203, 0.1315998077392578, 0.13230464172363282, 0.13229557800292968, 0.13136300659179687, 0.13183570861816407, 0.1303166046142578, 0.13028457641601562, 0.13196354675292968, 0.13137100219726563, 0.13180117797851562, 0.13251779174804687, 0.13241661071777344, 0.1307058563232422, 0.1302034912109375, 0.13293411254882812, 0.13179241943359374, 0.1319983367919922, 0.131716796875, 0.1328172149658203, 0.1318154296875, 0.13043302917480468, 0.13119631958007813, 0.13226637268066407, 0.13200611877441407, 0.143204345703125, 0.1316884460449219, 0.12993276977539062, 0.12972067260742187, 0.1299027862548828, 0.13164054870605468, 0.13056224060058594, 0.13329866027832032, 0.13185142517089843, 0.1317809295654297, 0.13043510437011718, 0.12992063903808593, 0.1299640350341797, 0.1303822021484375, 0.13200796508789062, 0.1327744903564453, 0.13241754150390625, 0.13149501037597655, 0.13011651611328126, 0.1298695373535156, 0.12989468383789063, 0.13122969055175782, 0.1325194549560547, 0.13213945007324218, 0.13252511596679686, 0.13087849426269532, 0.13031826782226563, 0.13000090026855468, 0.1309020233154297, 0.13178598022460938, 0.13195062255859374, 0.13282701110839842, 0.13154106140136718, 0.13018540954589844, 0.1319508819580078, 0.13054124450683594, 0.1317091522216797, 0.13170524597167968, 0.13203456115722656, 0.1323469696044922, 0.13075753784179686, 0.13151846313476562, 0.13171705627441407, 0.13039418029785158, 0.13111820983886718, 0.13191871643066405, 0.13304550170898438, 0.1316800994873047, 0.13178562927246093, 0.13167820739746094, 0.13195468139648436, 0.1304637451171875, 0.1308625030517578, 0.13187747192382812, 0.1323520050048828, 0.13158195495605468, 0.1319403533935547, 0.13213081359863282, 0.1316884460449219, 0.1309165802001953, 0.13259344482421875, 0.1328037109375, 0.1321460418701172, 0.14190634155273438, 0.13134243774414062, 0.1298836212158203, 0.12979869079589842, 0.12982441711425782, 0.1298189697265625, 0.13039942932128906, 0.1342779541015625, 0.131917724609375, 0.13192195129394532, 0.1307213134765625, 0.12987242126464843, 0.12984906005859376, 0.13066677856445313, 0.13265087890625, 0.13286566162109376, 0.13177702331542968, 0.131845947265625, 0.13087875366210938, 0.12998133850097657, 0.13000630187988282, 0.13139222717285157, 0.13267088317871092, 0.132008544921875, 0.13263626098632814, 0.13088096618652345, 0.1305855712890625, 0.13189462280273437, 0.13117251586914064, 0.13186099243164062, 0.1329332733154297, 0.13181578063964844, 0.13154304504394532, 0.1301930847167969, 0.13188946533203125, 0.1305518035888672, 0.1319219207763672, 0.13300502014160157, 0.1322183380126953, 0.13236306762695313, 0.13065011596679688, 0.13077667236328125, 0.1317728271484375, 0.13066009521484376, 0.13306002807617187, 0.13196780395507812, 0.13321830749511718, 0.13182550048828126, 0.13065373229980468, 0.13204934692382814, 0.13060319519042968, 0.13144064331054686, 0.13185775756835938, 0.13299778747558594, 0.1323984375, 0.13216015625, 0.1313955841064453, 0.1317061767578125, 0.1303948211669922, 0.13129318237304688, 0.13311180114746093, 0.132117919921875, 0.13223158264160156, 0.143544677734375, 0.1313217010498047, 0.12981468200683594, 0.13000895690917968, 0.12984947204589845, 0.12993280029296875, 0.13051910400390626, 0.1355200958251953, 0.132998046875, 0.1320813446044922, 0.13028108215332032, 0.12986344909667968, 0.12978031921386718, 0.13010105895996094, 0.13377247619628907, 0.13362640380859375, 0.13279075622558595, 0.13153805541992186, 0.1301246795654297, 0.12995123291015626, 0.1299311981201172, 0.13198176574707032, 0.13221490478515624, 0.1337159729003906, 0.13207962036132812, 0.1309388732910156, 0.13003366088867188, 0.13001139831542968, 0.13168336486816407, 0.1325003204345703, 0.13312806701660157, 0.13277743530273436, 0.1318507843017578, 0.13126422119140624, 0.13016444396972657, 0.13003628540039064, 0.13245234680175783, 0.13251484680175782, 0.13305340576171876, 0.13281893920898438, 0.13197500610351562, 0.13076495361328125, 0.13032383728027344, 0.13186111450195312, 0.13207347106933592, 0.13256105041503907, 0.13247811889648436, 0.13285446166992188, 0.13183590698242187, 0.13044122314453124, 0.1310834197998047, 0.13194940185546875, 0.13220672607421874, 0.13203648376464844, 0.13280487060546875, 0.13239033508300782, 0.1323726043701172, 0.1308323516845703, 0.13218380737304689, 0.13299533081054687, 0.13261846923828124, 0.13322409057617188, 0.13294435119628906, 0.14304421997070313, 0.13134477233886718, 0.13023794555664062, 0.1299113006591797, 0.12978346252441406, 0.1299019775390625, 0.13078009033203125, 0.1349376983642578, 0.13267216491699219, 0.13080982971191407, 0.12987411499023438, 0.13001449584960936, 0.12988088989257812, 0.13069517517089843, 0.13310771179199218, 0.1341441650390625, 0.1322125701904297, 0.13060096740722657, 0.12994345092773438, 0.13179498291015626, 0.13028153991699218, 0.13089170837402345, 0.13307090759277343, 0.13284573364257812, 0.13158297729492188, 0.13047439575195313, 0.13139398193359375, 0.13139762878417968, 0.13032652282714843, 0.132421630859375, 0.13237449645996094, 0.13267884826660156, 0.13083120727539063, 0.13158399963378906, 0.13083648681640625, 0.1320201873779297, 0.13136221313476562, 0.13190774536132813, 0.13297235107421876, 0.13188552856445312, 0.13194464111328125, 0.13141996765136718, 0.13190162658691407, 0.13233935546875, 0.1313723907470703, 0.13195376586914062, 0.13183753967285156, 0.1314981689453125, 0.13165904235839843, 0.13252899169921875, 0.13231103515625, 0.13177407836914062, 0.13224179077148437, 0.13288426208496093, 0.13220681762695313, 0.13076480102539062, 0.13314598083496093, 0.13139353942871093, 0.13193075561523437, 0.13202841186523437, 0.13261935424804688, 0.13261302185058593, 0.13199880981445314, 0.14318243408203124, 0.13144454956054688, 0.13009919738769532, 0.129778076171875, 0.13156947326660157, 0.13044956970214844, 0.13171282958984376, 0.13397943115234376, 0.13325946044921874, 0.13134083557128906, 0.1299148864746094, 0.13018931579589843, 0.1317184295654297, 0.1304071350097656, 0.131548828125, 0.13340403747558593, 0.13200236511230468, 0.13047235107421876, 0.13022207641601563, 0.13185420227050781, 0.13176812744140626, 0.13057049560546874, 0.13237254333496093, 0.13268377685546875, 0.13148159790039063, 0.13032585144042969, 0.1329670715332031, 0.13165180969238283, 0.13168966674804689, 0.13182217407226562, 0.13199974060058595, 0.13252931213378907, 0.13171388244628907, 0.1312761993408203, 0.13290147399902344, 0.1317264709472656, 0.13138624572753907, 0.13256454467773438, 0.1320821075439453, 0.13171440124511719, 0.13170140075683595, 0.13277781677246095, 0.13264834594726563, 0.13107685852050782, 0.13311386108398438, 0.13189045715332032, 0.1330501708984375, 0.13226821899414062, 0.1315499267578125, 0.1336269073486328, 0.1316116180419922, 0.13197071838378907, 0.13181167602539062, 0.132834716796875, 0.13242839050292968, 0.13191302490234375, 0.13297254943847656, 0.1325198974609375, 0.13114151000976562, 0.13256358337402344, 0.13150405883789062, 0.13311318969726563, 0.13203756713867187, 0.14307554626464844, 0.13148374938964844, 0.130052001953125, 0.12977561950683594, 0.12978994750976564, 0.1297810516357422, 0.13145993041992188, 0.1358720703125, 0.13280863952636718, 0.13074978637695311, 0.12978012084960938, 0.12980899047851563, 0.12977516174316406, 0.13207391357421874, 0.13401695251464843, 0.13310581970214844, 0.13168374633789062, 0.13159043884277344, 0.13019711303710937, 0.13001280212402344, 0.12990264892578124, 0.1322382354736328, 0.13339033508300782, 0.1324522247314453, 0.13234188842773437, 0.1317560272216797, 0.13034895324707033, 0.1299737548828125, 0.13119754028320313, 0.13287014770507813, 0.13217304992675782, 0.13245516967773438, 0.13263381958007814, 0.13162985229492188, 0.13023638916015626, 0.1309163818359375, 0.13186770629882813, 0.13270710754394532, 0.1325897216796875, 0.1321158447265625, 0.13225798034667968, 0.13136326599121093, 0.13107609558105468, 0.1317417297363281, 0.13295849609375, 0.13201478576660156, 0.13309645080566407, 0.13194650268554686, 0.13230624389648438, 0.13124195861816407, 0.1321267852783203, 0.13171363830566407, 0.13184005737304688, 0.13245053100585938, 0.13272860717773438, 0.132140380859375, 0.13147203063964844, 0.13195161437988281, 0.13210520935058595, 0.13171717834472657, 0.1330298309326172, 0.13210806274414064, 0.13224368286132812, 0.14255532836914062, 0.13150198364257812, 0.1300581512451172, 0.1297407989501953, 0.12979811096191407, 0.129762939453125, 0.1313140869140625, 0.1354895324707031, 0.13285372924804686, 0.13116549682617187, 0.13005398559570314, 0.13001210021972656, 0.12991276550292968, 0.13153468322753906, 0.13397116088867186, 0.13379049682617186, 0.13211782836914063, 0.13096022033691407, 0.12997625732421875, 0.12997978210449218, 0.13207212829589843, 0.1324564514160156, 0.13347634887695312, 0.1323251495361328, 0.13227638244628906, 0.13064198303222657, 0.13011311340332032, 0.13117890930175782, 0.13267881774902343, 0.1340015106201172, 0.13238067626953126, 0.13272816467285156, 0.13203485107421875, 0.13102896118164062, 0.13016233825683593, 0.13136358642578125, 0.13353575134277343, 0.13325222778320311, 0.13269296264648436, 0.13205494689941405, 0.13176368713378905, 0.13218460083007813, 0.1310576629638672, 0.133093017578125, 0.13262217712402344, 0.13324954223632812, 0.13225100708007811, 0.1314370574951172, 0.13147520446777344, 0.13238674926757812, 0.13162950134277343, 0.13316242980957033, 0.1323643798828125, 0.13323721313476564, 0.13201408386230468, 0.13140534973144533, 0.13207600402832032, 0.13188710021972655, 0.1322001647949219, 0.13241343688964843, 0.13359951782226562, 0.13199942016601562, 0.13253453063964843, 0.14246867370605468, 0.13145494079589845, 0.13006463623046874, 0.12988858032226563, 0.12987942504882813, 0.1297681579589844, 0.13288249206542968, 0.13623699951171875, 0.1331955261230469, 0.13142796325683595, 0.13084022521972657, 0.12984214782714842, 0.12985139465332032, 0.13211366271972655, 0.1340648956298828, 0.13347196960449217, 0.1320020751953125, 0.13168435668945314, 0.13094297790527343, 0.13013754272460937, 0.1300198974609375, 0.13323263549804687, 0.13358489990234376, 0.13228195190429687, 0.13265116882324218, 0.1308275909423828, 0.13006658935546875, 0.13043586730957032, 0.13207267761230468, 0.1328873291015625, 0.13237452697753907, 0.13350898742675782, 0.1323245086669922, 0.13124911499023437, 0.13036749267578124, 0.13162495422363282, 0.13256410217285156, 0.13307321166992186, 0.1330386505126953, 0.1322208709716797, 0.13198057556152343, 0.13064408874511718, 0.13152928161621094, 0.1317540740966797, 0.1331015625, 0.13238272094726564, 0.13343046569824218, 0.131887939453125, 0.13213900756835936, 0.13121945190429687, 0.13215904235839843, 0.132291015625, 0.13246975708007813, 0.1329792022705078, 0.1327229766845703, 0.13189126586914063, 0.1320592956542969, 0.1321942138671875, 0.13175613403320313, 0.13251513671875, 0.13355389404296875, 0.132567138671875, 0.133446533203125, 0.1422878723144531, 0.13119500732421874, 0.12993263244628905, 0.12983351135253907, 0.12989645385742188, 0.12986778259277343, 0.1321285400390625, 0.13679624938964843, 0.13272837829589842, 0.13104570007324218, 0.13024691772460936, 0.1298588409423828, 0.13105203247070313, 0.1312617950439453, 0.13388890075683593, 0.13333056640625, 0.13278399658203124, 0.13157017517089845, 0.13045350646972656, 0.13162495422363282, 0.1302118377685547, 0.13238412475585937, 0.13322093200683593, 0.1323868865966797, 0.13302989196777343, 0.1311846466064453, 0.13088307189941406, 0.13144729614257813, 0.13087344360351563, 0.13242767333984376, 0.13273843383789063, 0.13234034729003907, 0.13309295654296874, 0.1311664581298828, 0.13195036315917968, 0.13057472229003905, 0.13156556701660158, 0.13360537719726562, 0.1324416046142578, 0.1329402618408203, 0.13209324645996093, 0.1320885467529297, 0.13104537963867188, 0.13059686279296875, 0.13255474853515625, 0.13312527465820312, 0.13212144470214843, 0.1330931854248047, 0.13197273254394531, 0.13248159790039063, 0.13082009887695312, 0.13170278930664062, 0.13233766174316405, 0.132384765625, 0.13343257141113282, 0.1324297332763672, 0.13207638549804687, 0.13173504638671876, 0.13123635864257813, 0.13246054077148436, 0.13283731079101563, 0.13267359924316408, 0.1327626953125]",tokens/s,7.582935241146648,, @@ -5203,7 +5203,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 38.12 MiB is free. Process 184939 has 14.70 GiB memory in use. Of the allocated memory 14.42 GiB is allocated by PyTorch, and 176.08 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 38.12 MiB is free. Process 190870 has 14.70 GiB memory in use. Of the allocated memory 14.42 GiB is allocated by PyTorch, and 176.08 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,01-ai/Yi-6B,01-ai/Yi-6B,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,807.604224,4683.923456,0.0,4288.67584,4213.842432,s,1,13.5787783203125,13.5787783203125,0.0,13.5787783203125,13.5787783203125,13.5787783203125,13.5787783203125,[13.5787783203125],,kWh,0.00018497083590416515,2.03963994005206e-05,6.22283831159991e-05,0.00026759561842068486,,MB,1192.230912,5107.54816,0.0,4699.717632,4535.11424,s,10,8.553767395019532,0.855376739501953,0.007683236508205055,0.8569793090820312,0.8610069885253906,0.8619868316650391,0.8627707061767578,"[0.8337443237304687, 0.852953857421875, 0.85672607421875, 0.8551661376953125, 0.8570130615234375, 0.8589288330078125, 0.8607892456054688, 0.8629666748046875, 0.856945556640625, 0.8585336303710938]",tokens/s,299.28333116593416,kWh,2.48531106444444e-05,2.7390895892942575e-06,1.6417744615666776e-05,4.400994484940543e-05,tokens/kWh,5816867.093925897,MB,1232.44544,5115.936768,0.0,4708.10624,4535.1168,s,10,40.23405615234375,4.023405615234375,0.0061507504575199655,4.025970825195312,4.029557641601563,4.030256481933594,4.030815554199219,"[4.015683837890625, 4.010197021484375, 4.019560546875, 4.02631640625, 4.025761962890625, 4.022799560546875, 4.0261796875, 4.027199462890625, 4.030955322265625, 4.02940234375]",tokens/s,15.658376516017778,kWh,0.0001180998546972224,1.3028841168202256e-05,7.851052577133302e-05,0.00020963922163675768,tokens/kWh,300516.2846347533,,s,630,40.23095680236817,0.06385866159106059,0.0015793124110772148,0.06369526290893554,0.06435266647338868,0.06462068290710449,0.07443461128234863,"[0.0795832290649414, 0.06496665954589843, 0.06382553482055664, 0.0633240966796875, 0.06302150344848632, 0.06267267227172851, 0.06260348892211914, 0.06262319946289062, 0.06255990219116211, 0.06251110458374023, 0.0625406723022461, 0.0626558723449707, 0.0636995849609375, 0.06333849716186524, 0.06316032028198242, 0.0640646743774414, 0.06386947250366211, 0.06376176071166992, 0.06457443237304687, 0.06376364898681641, 0.06333523178100586, 0.06309888076782226, 0.0625459213256836, 0.0625172462463379, 0.06259507369995117, 0.06265753555297851, 0.062499679565429685, 0.06303321456909179, 0.06385036849975587, 0.06337561416625977, 0.0632762565612793, 0.06387731170654297, 0.06397161483764649, 0.06438889312744141, 0.06388556671142578, 0.06335715103149414, 0.06384867095947265, 0.06398745727539062, 0.06370655822753907, 0.0633639030456543, 0.06284284973144531, 0.06338313674926757, 0.06284649658203124, 0.06317763137817382, 0.06418812561035156, 0.06393593597412109, 0.06363631820678711, 0.0637050895690918, 0.06414070129394531, 0.06332067108154296, 0.0645630111694336, 0.06414915466308593, 0.06368515014648438, 0.063246337890625, 0.0632828483581543, 0.06490048217773438, 0.06420988464355469, 0.06366207885742188, 0.06340329742431641, 0.06301123046875, 0.06356784057617187, 0.06435033416748047, 0.06386483383178711, 0.07312252807617188, 0.06415724945068359, 0.0634764175415039, 0.06313958358764649, 0.06327705764770508, 0.06235340881347656, 0.06369859313964844, 0.06333190536499024, 0.06303577423095703, 0.06275542449951171, 0.062441280364990234, 0.06250700759887695, 0.062476287841796874, 0.0642231674194336, 0.0637768325805664, 0.06471459197998047, 0.06404045104980469, 0.06379996871948242, 0.06357401657104492, 0.06384569549560547, 0.06351119995117188, 0.06330972671508789, 0.0628040657043457, 0.06274051284790039, 0.06330368041992188, 0.06374560165405273, 0.06343718338012695, 0.06299619293212891, 0.06276287841796875, 0.06390422439575195, 0.06333440017700195, 0.0637248649597168, 0.0641173095703125, 0.06391158294677735, 0.06370352172851562, 0.06410854339599609, 0.0637393913269043, 0.06341856002807617, 0.0632111358642578, 0.06312825775146484, 0.06380915069580079, 0.0633081283569336, 0.06303696060180664, 0.0627143669128418, 0.06400780487060546, 0.06390412902832031, 0.0633733139038086, 0.06342009735107422, 0.06392863845825195, 0.0635431022644043, 0.06429920196533204, 0.06409830474853516, 0.06335279846191406, 0.06341020965576172, 0.06396627044677734, 0.06368966293334961, 0.06398102569580078, 0.06364543914794922, 0.06342086410522461, 0.06347398376464844, 0.06394473648071289, 0.06361280059814453, 0.06330364990234374, 0.07437910461425781, 0.06458464050292968, 0.06365695953369141, 0.06322102355957031, 0.06286921691894531, 0.0636409912109375, 0.06352137756347656, 0.06316646575927734, 0.06252665710449219, 0.06272697448730469, 0.06378086471557617, 0.0632845115661621, 0.06389014434814454, 0.06371680068969726, 0.06330182266235351, 0.06348393630981446, 0.06390819168090821, 0.06345523071289062, 0.06521616363525391, 0.06425122833251953, 0.06332115173339843, 0.06318073654174805, 0.06293289566040039, 0.06377251052856445, 0.0633182716369629, 0.06302505493164062, 0.06321734237670898, 0.06329385757446289, 0.06403616333007812, 0.0636110725402832, 0.06324684906005859, 0.06309478378295899, 0.06383206558227539, 0.0639360008239746, 0.06431385803222656, 0.06393446350097656, 0.0636467514038086, 0.06354838562011719, 0.06402047729492187, 0.06362931060791016, 0.06343270492553711, 0.0631596794128418, 0.06395967864990235, 0.063825439453125, 0.06346185684204102, 0.06345913696289063, 0.06400364685058593, 0.06367295837402344, 0.06465948486328126, 0.06430003356933593, 0.06346441650390625, 0.06420480346679687, 0.0636701774597168, 0.06337340927124023, 0.0633507843017578, 0.06420025634765625, 0.06356832122802734, 0.0633481903076172, 0.06321763229370117, 0.06413497924804687, 0.06366396713256836, 0.0634090576171875, 0.06420233917236329, 0.07695600128173828, 0.06464921569824218, 0.06372963333129883, 0.06340790557861328, 0.06316249465942383, 0.06255628967285157, 0.06267497634887695, 0.06326800155639649, 0.06375507354736328, 0.06320876693725586, 0.06297875213623047, 0.06274383926391601, 0.06377510452270507, 0.06346550369262695, 0.06382009506225586, 0.06616886138916016, 0.06432262420654297, 0.06374697494506835, 0.06332412719726563, 0.06311939239501953, 0.0638914566040039, 0.06340403366088868, 0.06329958343505859, 0.0628936653137207, 0.06392464065551758, 0.06333030319213867, 0.0631621437072754, 0.06397974395751953, 0.0635590705871582, 0.06332851028442382, 0.06389139175415039, 0.06386633682250976, 0.0639119987487793, 0.06394355010986329, 0.06377459335327149, 0.06328537750244141, 0.06301283264160157, 0.06409219360351562, 0.06359775924682617, 0.06438790130615234, 0.06381568145751954, 0.0634511375427246, 0.06315766525268554, 0.06394326400756836, 0.06369894409179687, 0.06428444671630859, 0.06386825561523438, 0.06373260879516601, 0.06383580780029297, 0.06431308746337891, 0.06373430252075195, 0.06334265518188477, 0.06407782745361328, 0.0639815673828125, 0.06369865417480469, 0.06437328338623047, 0.0643663330078125, 0.06412249755859376, 0.06380992126464843, 0.06351033782958984, 0.06418450927734375, 0.06368438339233398, 0.06358780670166016, 0.07429385375976562, 0.06462054443359375, 0.06384022521972656, 0.06342838287353515, 0.06279193496704101, 0.06415468597412109, 0.06344595336914062, 0.06334156799316407, 0.06272905731201171, 0.06370470428466797, 0.06331856155395507, 0.06311731338500977, 0.0627341423034668, 0.06273763275146485, 0.06359139251708984, 0.06443993377685547, 0.06490528106689453, 0.06442208099365235, 0.06388953781127929, 0.06334627151489258, 0.0631624641418457, 0.06402285003662109, 0.06332956695556641, 0.0637993278503418, 0.06358291244506836, 0.0633133773803711, 0.06330217742919922, 0.0638579216003418, 0.06359116744995118, 0.06416588592529297, 0.06381510543823242, 0.06354585647583008, 0.06401958465576171, 0.06356268692016602, 0.06339142227172852, 0.06431161499023437, 0.06392012786865234, 0.0636701774597168, 0.063587646484375, 0.06413507080078125, 0.06363958358764649, 0.06346428680419922, 0.06417203521728515, 0.06385868835449218, 0.06366412734985352, 0.06375804901123047, 0.06428294372558593, 0.06409552001953125, 0.06460867309570313, 0.0641416015625, 0.06375116729736328, 0.06361804962158203, 0.06415071868896484, 0.06364652633666992, 0.06332387161254883, 0.06406172943115235, 0.06379520034790039, 0.06346108627319336, 0.06395318222045898, 0.06386198425292969, 0.06359939193725586, 0.06357932662963867, 0.06406358337402343, 0.07445728302001953, 0.06475315093994141, 0.06353561782836914, 0.06343024063110352, 0.06279529571533203, 0.06280928039550782, 0.06386246490478516, 0.06349619293212891, 0.06315145492553711, 0.06297868728637696, 0.06260678482055664, 0.06386134338378906, 0.06337068939208984, 0.06317091369628906, 0.06400994873046875, 0.06505027008056641, 0.06404787445068359, 0.06393836975097657, 0.06415302276611329, 0.06372438430786133, 0.06333440017700195, 0.06293724822998047, 0.06370902252197265, 0.0632176628112793, 0.06301907348632813, 0.0629554557800293, 0.06402566528320312, 0.06337631988525391, 0.06329344177246093, 0.06374115371704102, 0.06409072113037109, 0.06431254577636719, 0.06407881927490235, 0.06416793823242188, 0.06366207885742188, 0.06313478469848632, 0.06464752197265625, 0.06399814224243164, 0.06348144149780273, 0.06342851257324218, 0.06331836700439453, 0.06376710510253907, 0.06335283279418945, 0.06342860794067383, 0.06381523132324218, 0.0637014389038086, 0.06395302581787109, 0.06361619186401367, 0.06457414245605468, 0.06394684982299804, 0.06339369583129882, 0.06410610961914062, 0.06351248168945313, 0.0634150733947754, 0.06509331512451172, 0.0639870719909668, 0.06342105484008789, 0.06325411224365235, 0.06406934356689453, 0.06366073608398437, 0.06332527923583985, 0.0642159652709961, 0.0636948471069336, 0.07531222534179688, 0.06462079620361329, 0.06379996871948242, 0.06337696075439453, 0.06303801727294922, 0.0636102409362793, 0.06324070358276367, 0.06300991821289062, 0.06269583892822266, 0.06371737670898438, 0.06325500869750976, 0.06313926315307618, 0.06258652877807618, 0.06288272094726563, 0.06478438568115234, 0.06436438751220704, 0.06378307342529296, 0.06399107360839844, 0.06361072158813477, 0.06359132766723632, 0.06384409713745118, 0.06353737640380859, 0.06320115280151367, 0.06280121612548828, 0.06375711822509765, 0.06329139328002929, 0.06425775909423828, 0.06379312133789063, 0.06327436828613281, 0.06305385589599609, 0.0638554573059082, 0.06383967971801757, 0.06440409851074219, 0.06389990234375, 0.06346867370605469, 0.06398425674438477, 0.06354102325439454, 0.0631764793395996, 0.06457929229736328, 0.06410518646240235, 0.0635945930480957, 0.06369612884521485, 0.06408448028564454, 0.06363561630249023, 0.06340563201904297, 0.06410886383056641, 0.06390182495117187, 0.06461974334716797, 0.06414415740966797, 0.06425395202636719, 0.06407373046875, 0.06354534530639648, 0.06346342468261719, 0.06413311767578125, 0.06381158447265625, 0.06408956909179687, 0.0638100814819336, 0.06348185729980468, 0.06382976150512695, 0.0641170883178711, 0.06343670272827148, 0.0639447021484375, 0.06460633850097657, 0.07571673583984374, 0.06435225677490235, 0.06355267333984375, 0.06304240036010743, 0.06290633773803711, 0.06353475189208985, 0.06309747314453125, 0.06297369766235351, 0.06274867248535156, 0.06370950317382812, 0.06312419128417969, 0.06291292953491211, 0.0626324462890625, 0.06366758346557617, 0.06419321441650391, 0.06503833770751953, 0.06456114959716797, 0.06379296112060547, 0.06356150436401367, 0.06330995178222656, 0.06420451354980469, 0.0638039665222168, 0.06333030319213867, 0.06318080139160157, 0.0638230094909668, 0.06338032150268555, 0.06447305297851562, 0.06386415863037109, 0.06330780792236328, 0.06332828903198243, 0.06437542724609376, 0.0640552978515625, 0.06458163452148437, 0.064052734375, 0.06363977432250977, 0.06350467300415039, 0.06399385452270508, 0.06348121643066407, 0.06449625396728516, 0.06398361587524413, 0.06338355255126953, 0.06317670440673828, 0.06400972747802734, 0.06347574234008789, 0.06323247909545898, 0.0639283218383789, 0.06360883331298828, 0.06386191940307617, 0.06383190536499024, 0.06400415802001953, 0.06360547256469727, 0.06441558074951172, 0.06400991821289062, 0.0636844482421875, 0.0638902702331543, 0.06449533081054687, 0.06392214584350586, 0.06343702316284179, 0.06347558212280273, 0.0640857925415039, 0.06349427032470703, 0.06471900939941407, 0.06391334533691406, 0.07597491455078124, 0.0642171859741211, 0.06341791915893555, 0.06361747360229492, 0.06329296112060546, 0.06308911895751954, 0.06257247924804688, 0.06369635009765626, 0.06357852935791015, 0.06301417541503906, 0.06384118270874023, 0.06354742431640625, 0.06318025588989258, 0.06273247909545898, 0.06388518524169921, 0.06417996978759766, 0.06533542633056641, 0.06374662399291992, 0.0634511375427246, 0.06325417709350586, 0.0636399040222168, 0.06313129425048829, 0.06416121673583984, 0.06388214492797852, 0.06344294357299805, 0.06315795135498047, 0.06286515045166016, 0.06378515243530274, 0.06464374542236329, 0.06358537673950196, 0.06363590240478516, 0.06431145477294922, 0.06379868698120117, 0.06421977233886719, 0.064036865234375, 0.06374604797363281, 0.06342361450195312, 0.06404390716552734, 0.06351052856445312, 0.06409830474853516, 0.06376652908325195, 0.06351052856445312, 0.0638416976928711, 0.06405795288085937, 0.06359827041625976, 0.06482793426513672, 0.06409756469726563, 0.06391609573364258, 0.06394265747070313, 0.06387302398681641, 0.06441004943847656, 0.0641370849609375, 0.06354940795898438, 0.06423772430419922, 0.0638914566040039, 0.06347980880737304, 0.0645406723022461, 0.0639733772277832, 0.0636129264831543, 0.06380710220336915, 0.06404061126708985, 0.06463970947265625, 0.06415360260009766, 0.07515340423583984, 0.06424278259277344, 0.06361385726928712, 0.06319862365722656, 0.0627677116394043, 0.06386073684692382, 0.06337075042724609, 0.06318540954589844, 0.06317232131958007, 0.06365008163452149, 0.06380748748779297, 0.06350848007202148, 0.0631541748046875, 0.06285619354248047, 0.06392863845825195, 0.06398396682739257, 0.06446115112304687, 0.06371532821655274, 0.06376144027709961, 0.06311040115356445, 0.06364499282836913, 0.06317712020874024, 0.06383606338500977, 0.06353724670410156, 0.06316851043701172, 0.06351804733276367, 0.0635338897705078, 0.06326051330566407, 0.06312691116333008, 0.06395967864990235, 0.06397132873535157, 0.06395296096801757, 0.0638463363647461, 0.0638683853149414, 0.06458188629150391, 0.06406294250488281, 0.0636956787109375, 0.06351811218261719, 0.06378147125244141, 0.06360678482055664, 0.06482125091552735, 0.06421094512939453, 0.06389980697631836, 0.06352646255493163, 0.06325385665893554, 0.06406444549560547, 0.06366617584228515, 0.06444457244873047, 0.06416681671142578, 0.06398867034912109, 0.06407513427734375, 0.06455270385742187, 0.06422335815429688, 0.06396355056762695, 0.06360303878784179, 0.0640401611328125, 0.06465740966796875, 0.06404521942138672, 0.0636646728515625, 0.06338713455200196, 0.0640722885131836, 0.06435635375976563, 0.0642718734741211]",tokens/s,15.659582820633176,, @@ -5324,7 +5324,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 102777 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 108896 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.3.1+cu121,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-0.5B,Qwen/Qwen1.5-0.5B,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.75552,Linux,x86_64,Linux-5.10.220-209.869.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.43.4,,0.33.0,,,,1.21.2,,,,0.12.0,,,True,MB,1178.492928,1093.599232,0.0,698.351616,690.178048,s,1,9.307734375,9.307734375,0.0,9.307734375,9.307734375,9.307734375,9.307734375,[9.307734375],,kWh,4.813778019170816e-05,5.30268701156643e-06,1.66522355439791e-05,7.009270274725368e-05,,MB,1372.995584,1408.172032,0.0,1000.341504,957.775872,s,10,0.5886727676391602,0.058867276763916024,0.0008932273377649839,0.05856430435180664,0.05927543525695801,0.06033389415740967,0.061180661277771,"[0.05887075042724609, 0.058735103607177735, 0.05904022216796875, 0.05820880126953125, 0.058393505096435545, 0.06139235305786133, 0.05903468704223633, 0.05824940872192383, 0.05835657501220703, 0.058391361236572265]",tokens/s,4348.76580118822,kWh,1.8262286517710889e-06,2.013398846739338e-07,1.2138065265999281e-06,3.241375063044951e-06,tokens/kWh,78978826.89314988,MB,1387.06944,1416.56064,0.0,1008.730112,957.778432,s,10,25.123927734375,2.5123927734375004,0.007594485465332524,2.5096402587890623,2.52302841796875,2.52503056640625,2.52663228515625,"[2.5029521484375, 2.51033251953125, 2.505419677734375, 2.52703271484375, 2.508947998046875, 2.506647216796875, 2.50775732421875, 2.51946875, 2.52258349609375, 2.512785888671875]",tokens/s,25.075697027181892,kWh,7.273305940906234e-05,8.022381448986752e-06,3.13892751113997e-05,0.00011214471596944882,tokens/kWh,561774.12779005,,s,630,25.121339572906486,0.03987514217921666,0.0006522795463124235,0.03971884727478027,0.04030095443725586,0.040638852691650386,0.04335835014343263,"[0.039836639404296874, 0.040035518646240234, 0.040170177459716794, 0.039548095703125, 0.03970502471923828, 0.03939376068115234, 0.039524383544921875, 0.03946905517578125, 0.03934003067016602, 0.039572704315185545, 0.0394596176147461, 0.039615840911865235, 0.039416481018066406, 0.03941580963134766, 0.0394958381652832, 0.03980271911621094, 0.039354366302490236, 0.039370750427246096, 0.03942604827880859, 0.039667713165283204, 0.0397127685546875, 0.03951411056518555, 0.03952230453491211, 0.03984384155273438, 0.039569408416748046, 0.03940761566162109, 0.039896385192871094, 0.03985580825805664, 0.03975270462036133, 0.039616512298583983, 0.03965542221069336, 0.039772159576416014, 0.039712257385253906, 0.03989350509643555, 0.03991676712036133, 0.03953276824951172, 0.03972758483886719, 0.04024095916748047, 0.03947734451293945, 0.03963324737548828, 0.03955023956298828, 0.03963910293579102, 0.03971468734741211, 0.03976169586181641, 0.040372383117675784, 0.04014969635009766, 0.03961974334716797, 0.0395497932434082, 0.0397946891784668, 0.039736446380615235, 0.03977849578857422, 0.03951996612548828, 0.039363296508789065, 0.03947043228149414, 0.03971535873413086, 0.03972531127929688, 0.03955862426757813, 0.039553470611572265, 0.03953276824951172, 0.0396492805480957, 0.04370761489868164, 0.03965945434570312, 0.039697025299072264, 0.040117599487304687, 0.03997119903564453, 0.039870750427246096, 0.039686111450195315, 0.03997596740722656, 0.04007219314575195, 0.04203724670410156, 0.040457759857177734, 0.0402006721496582, 0.04028416061401367, 0.040091232299804686, 0.0399257926940918, 0.03949606323242188, 0.03956531143188476, 0.040037696838378906, 0.03982815933227539, 0.039873790740966794, 0.03938777542114258, 0.03946099090576172, 0.03996192169189453, 0.03932844924926758, 0.03942399978637695, 0.039354366302490236, 0.039454113006591796, 0.04000009536743164, 0.03957964706420898, 0.03951411056518555, 0.03977328109741211, 0.03963772964477539, 0.0396822395324707, 0.039556961059570316, 0.039622112274169924, 0.0396743049621582, 0.03964339065551758, 0.040551551818847655, 0.03957235336303711, 0.039519649505615234, 0.03975753784179688, 0.03965983963012695, 0.03967238235473633, 0.039666847229003904, 0.03996096038818359, 0.03991376113891602, 0.04015068817138672, 0.040065567016601564, 0.039760929107666015, 0.03962140655517578, 0.03955318450927734, 0.0401324462890625, 0.040065216064453124, 0.039763904571533205, 0.039848064422607424, 0.03967158508300781, 0.04012403106689453, 0.039811614990234376, 0.03960992050170899, 0.03953907012939453, 0.03991263961791992, 0.039580543518066405, 0.040938655853271486, 0.03960636901855469, 0.03985055923461914, 0.039642433166503906, 0.03991961669921875, 0.03984143829345703, 0.03998134231567383, 0.03970054244995117, 0.04029958343505859, 0.04053087997436523, 0.039739391326904294, 0.039733375549316406, 0.0395665283203125, 0.039766719818115234, 0.039593982696533206, 0.039707969665527344, 0.03966207885742187, 0.03940371322631836, 0.03962575912475586, 0.03967689514160156, 0.03980806350708008, 0.03967808151245117, 0.039791423797607424, 0.039626750946044925, 0.04024515151977539, 0.039772449493408205, 0.040640033721923825, 0.03969257736206055, 0.039935359954833986, 0.04027865600585938, 0.03996031951904297, 0.03981654357910156, 0.03975609588623047, 0.03976992034912109, 0.03989993667602539, 0.03959807968139648, 0.039600128173828124, 0.03952025604248047, 0.03954278564453125, 0.03958927917480469, 0.0397973747253418, 0.03982755279541016, 0.03944214248657227, 0.03952195358276367, 0.039517982482910156, 0.03969302368164063, 0.03957360076904297, 0.039800704956054686, 0.03949961471557617, 0.03981536102294922, 0.03966566467285156, 0.03950947189331055, 0.03954332733154297, 0.039858177185058595, 0.04044095993041992, 0.0397628173828125, 0.03968928146362305, 0.03945568084716797, 0.03986636734008789, 0.04024662399291992, 0.039481502532958984, 0.03952076721191406, 0.0396759033203125, 0.03980886459350586, 0.039704734802246094, 0.0395830078125, 0.03959667205810547, 0.04001308822631836, 0.03993439865112305, 0.039808734893798825, 0.03948303985595703, 0.04003049468994141, 0.039513694763183595, 0.04011212921142578, 0.03960108947753906, 0.039741439819335936, 0.03963286590576172, 0.03958560180664063, 0.039516223907470706, 0.03956137466430664, 0.039470111846923825, 0.039699424743652345, 0.041523200988769535, 0.04022902297973633, 0.03948118209838867, 0.03942604827880859, 0.0396409912109375, 0.039659614562988284, 0.039451744079589846, 0.03967001724243164, 0.03971958541870117, 0.04053606414794922, 0.04031206512451172, 0.03970943832397461, 0.04045401763916016, 0.03977836990356445, 0.04011148834228516, 0.03994489669799805, 0.040151039123535154, 0.04017561721801758, 0.04027996826171875, 0.039927391052246096, 0.03974588775634766, 0.03979411315917969, 0.0434752311706543, 0.03988320159912109, 0.03976806259155274, 0.04534272003173828, 0.04501913452148437, 0.0409169921875, 0.04099225616455078, 0.04020803070068359, 0.039664478302001954, 0.039876609802246096, 0.03980284881591797, 0.03959632110595703, 0.04005043029785156, 0.03949977493286133, 0.03985539245605469, 0.0399832649230957, 0.03973791885375977, 0.03958169555664062, 0.04023875045776367, 0.039448287963867186, 0.03946793746948242, 0.03937251281738281, 0.039593982696533206, 0.03940332794189453, 0.04034515380859375, 0.0402213134765625, 0.040048641204833986, 0.04024863815307617, 0.039995166778564455, 0.03971510314941406, 0.0396596794128418, 0.03959142303466797, 0.0393691520690918, 0.03968582534790039, 0.03981923294067383, 0.039525249481201175, 0.039653377532958986, 0.03949772644042969, 0.0397209587097168, 0.039626750946044925, 0.03943587112426758, 0.03954524612426758, 0.03996057510375976, 0.03943987274169922, 0.03936307144165039, 0.039395328521728515, 0.04182160186767578, 0.039530303955078124, 0.039651424407958984, 0.03940652847290039, 0.03930495834350586, 0.039591934204101564, 0.039712158203125, 0.04006358337402344, 0.03972710418701172, 0.03937279891967774, 0.03968991851806641, 0.03944681549072265, 0.03948880004882813, 0.040428287506103514, 0.03995859146118164, 0.039712703704833985, 0.039624702453613284, 0.03979673767089844, 0.03970851135253906, 0.0396412467956543, 0.03982950210571289, 0.04016128158569336, 0.03989718246459961, 0.04000678253173828, 0.03962502288818359, 0.03987299346923828, 0.039446529388427735, 0.039708671569824217, 0.040054561614990235, 0.03965359878540039, 0.03996057510375976, 0.039593982696533206, 0.03953811264038086, 0.040581214904785154, 0.04224252700805664, 0.040243457794189454, 0.039487232208251954, 0.03961427307128906, 0.03993164825439453, 0.03973318481445313, 0.040753662109375, 0.04004044723510742, 0.039739391326904294, 0.04007516860961914, 0.04176454544067383, 0.040866111755371096, 0.04020038223266602, 0.03973625564575195, 0.0397628173828125, 0.039599777221679684, 0.03957180786132813, 0.039726207733154294, 0.03948556900024414, 0.03958217620849609, 0.03963523101806641, 0.0396308479309082, 0.039610366821289066, 0.03965542221069336, 0.03960128021240234, 0.03957372665405273, 0.039487583160400394, 0.039653663635253904, 0.03942745590209961, 0.03953961563110352, 0.03953216171264649, 0.039586177825927736, 0.039561344146728517, 0.04035747146606445, 0.0398111686706543, 0.04006057739257812, 0.039924030303955076, 0.039908641815185546, 0.03963385772705078, 0.04037628936767578, 0.039771617889404295, 0.039559520721435544, 0.04011644744873047, 0.03982745742797852, 0.03987804794311523, 0.04004627227783203, 0.03995536041259766, 0.039897087097167966, 0.03968384170532226, 0.03954473495483399, 0.039817569732666015, 0.03954687881469727, 0.039798080444335936, 0.03958428955078125, 0.03944195175170898, 0.03962944030761719, 0.03948044967651367, 0.03977510452270508, 0.04005260848999023, 0.03967942428588867, 0.03978720092773438, 0.03971891021728516, 0.039857471466064456, 0.03982124710083008, 0.039653182983398434, 0.03982227325439453, 0.039608001708984375, 0.03962809753417969, 0.039657726287841796, 0.039504638671875, 0.039607872009277345, 0.03970502471923828, 0.03976508712768555, 0.039790847778320315, 0.03988742446899414, 0.03944655990600586, 0.039849151611328126, 0.03964150238037109, 0.039992927551269535, 0.0396352653503418, 0.03942867279052734, 0.03975372695922851, 0.039976287841796875, 0.03974969482421875, 0.04045475387573242, 0.04043161773681641, 0.039810081481933594, 0.040637409210205075, 0.03978358459472656, 0.0399901123046875, 0.039339134216308594, 0.039506111145019535, 0.03944723129272461, 0.03944243240356445, 0.039458049774169925, 0.03936332702636719, 0.03952844619750977, 0.03935846328735351, 0.03957555389404297, 0.04008345413208008, 0.041207809448242184, 0.03980287933349609, 0.03967718505859375, 0.03950892639160156, 0.039352127075195316, 0.03958147048950195, 0.03967612838745117, 0.03975987243652344, 0.04035174560546875, 0.04057088088989258, 0.04015043258666992, 0.039696414947509764, 0.039929759979248046, 0.03959584045410156, 0.03978857421875, 0.039521087646484376, 0.03952844619750977, 0.039686145782470705, 0.03963068771362305, 0.040145057678222656, 0.04067327880859375, 0.0397918701171875, 0.039846656799316406, 0.03967712020874024, 0.03968083190917969, 0.03967974472045899, 0.03942819213867187, 0.03981327819824219, 0.04011539077758789, 0.03983443069458008, 0.03957084655761719, 0.03970547103881836, 0.03964694213867188, 0.040017280578613285, 0.03972774505615234, 0.04260752105712891, 0.03974467086791992, 0.03977481460571289, 0.03944473648071289, 0.039577598571777346, 0.039798782348632815, 0.03952435302734375, 0.04036403274536133, 0.041090847015380856, 0.03960444641113281, 0.039478687286376955, 0.03925872039794922, 0.03926425552368164, 0.039321952819824216, 0.039394977569580075, 0.0392540168762207, 0.03916595077514649, 0.04287692642211914, 0.03957964706420898, 0.03945792007446289, 0.03946912002563477, 0.039340480804443356, 0.04030054473876953, 0.03996710586547852, 0.03952777481079101, 0.03923011016845703, 0.03941580963134766, 0.03957555389404297, 0.039540542602539065, 0.03956252670288086, 0.03962358474731445, 0.03922467041015625, 0.039400096893310546, 0.045762016296386716, 0.04368809509277344, 0.03998348617553711, 0.03955686569213867, 0.03947135925292969, 0.039411006927490236, 0.03943078231811523, 0.039374912261962894, 0.039617599487304686, 0.039848896026611326, 0.039723007202148435, 0.03969785690307617, 0.03988127899169922, 0.039615840911865235, 0.04044252777099609, 0.04009983825683594, 0.04025139236450195, 0.040627552032470704, 0.039765758514404295, 0.039711647033691407, 0.03970240020751953, 0.03984944152832031, 0.04105599975585938, 0.041044097900390625, 0.040304641723632816, 0.039768863677978515, 0.03964675140380859, 0.03972073745727539, 0.0395880012512207, 0.039809215545654295, 0.04021868896484375, 0.04033331298828125, 0.04029417419433594, 0.039870014190673826, 0.040018592834472656, 0.03981926345825195, 0.04001567840576172, 0.039724224090576174, 0.040497920989990235, 0.04020780944824219, 0.040178497314453124, 0.03992473602294922, 0.03987148666381836, 0.039839134216308594, 0.03970134353637695, 0.039833343505859375, 0.03966566467285156, 0.03973241424560547, 0.03977913665771484, 0.03977974319458008, 0.0398526382446289, 0.03965542221069336, 0.03956067276000977, 0.039491710662841795, 0.039575809478759764, 0.039657505035400394, 0.039510143280029296, 0.040476673126220705, 0.03981721496582031, 0.03979788970947266, 0.03983798217773438, 0.03999599838256836, 0.03971891021728516, 0.04027088165283203, 0.04043670272827148, 0.04048896026611328, 0.040025215148925784, 0.040180606842041014, 0.03982131195068359, 0.03987446212768555, 0.03975084686279297, 0.039965599060058594, 0.03977830505371094, 0.03998720169067383, 0.04108451080322266, 0.03994460678100586, 0.039699775695800785, 0.043072193145751954, 0.039739391326904294, 0.03959574508666992, 0.039532833099365235, 0.03966323089599609, 0.03993452835083008, 0.04029420852661133, 0.03966880035400391, 0.04073507308959961, 0.041477825164794924, 0.03991024017333984, 0.04049862289428711, 0.041063201904296874, 0.03992876815795898, 0.03993632125854492, 0.0397031364440918, 0.040215679168701175, 0.0401396484375, 0.04008665466308594, 0.039680896759033205, 0.03965488052368164, 0.039980670928955075, 0.03959081649780274, 0.03964131164550781, 0.03989481735229492, 0.040046592712402344, 0.04051459121704101, 0.039771102905273435, 0.04009983825683594, 0.03953868865966797, 0.03971878433227539, 0.03982144165039062, 0.03963852691650391, 0.03949820709228516, 0.039554080963134765, 0.040068096160888675, 0.039884159088134766, 0.04013324737548828, 0.03996623992919922, 0.04012393569946289, 0.0407336311340332, 0.04003430557250977, 0.04365107345581055, 0.040551551818847655, 0.039787296295166016, 0.04003440093994141, 0.040210430145263674, 0.03989049530029297, 0.03965087890625, 0.039655872344970707, 0.03954937744140625, 0.03959590530395508, 0.03952243041992187, 0.039587329864501954, 0.039580158233642575, 0.03950796890258789, 0.039567359924316405, 0.03953049468994141, 0.039462913513183595, 0.03943833541870117, 0.03982460784912109, 0.04038291168212891, 0.03996899032592773, 0.03971289443969726, 0.039644542694091796, 0.03960201644897461, 0.03953129577636719, 0.039812576293945315, 0.039615009307861326, 0.03956121444702149, 0.040887935638427735, 0.039696800231933595, 0.03989910507202148, 0.03963904190063477, 0.03965705490112305, 0.03959584045410156, 0.039524513244628905, 0.039537086486816406, 0.039618560791015625]",tokens/s,25.078280486262706,, @@ -5440,7 +5440,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.52 GiB is free. Process 112445 has 13.22 GiB memory in use. Of the allocated memory 13.10 GiB is allocated by PyTorch, and 6.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.52 GiB is free. Process 118559 has 13.22 GiB memory in use. Of the allocated memory 13.10 GiB is allocated by PyTorch, and 6.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,opt,facebook/opt-350m,facebook/opt-350m,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -7012,7 +7012,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 97234 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 103313 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-32B,Qwen/Qwen1.5-32B,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -7187,7 +7187,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 61883 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 62277 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-70b-hf,meta-llama/Llama-2-70b-hf,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -7421,7 +7421,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.42 GiB is free. Process 111878 has 13.32 GiB memory in use. Of the allocated memory 13.21 GiB is allocated by PyTorch, and 1.48 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.42 GiB is free. Process 117926 has 13.32 GiB memory in use. Of the allocated memory 13.21 GiB is allocated by PyTorch, and 1.48 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-eager,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,opt,facebook/opt-350m,facebook/opt-350m,cuda,0,42,,,True,True,,float16,True,False,,eager,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.223-212.873.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.4.0,,4.44.2,,0.34.2,,,,1.21.4,,,,0.12.0,,,,MB,1792.356352,790.495232,0.0,387.97312,373.788672,s,1,9.1069775390625,9.1069775390625,0.0,9.1069775390625,9.1069775390625,9.1069775390625,9.1069775390625,[9.1069775390625],,kWh,5.134861882498475e-05,5.657091890138377e-06,1.690612463599306e-05,7.391183535111618e-05,,MB,1844.371456,891.158528,0.0,467.664896,421.489664,s,10,0.7397996520996093,0.07397996520996093,0.0010280551395685884,0.07349257659912109,0.07558497161865235,0.07593565521240235,0.07621620208740235,"[0.0734228515625, 0.07317411041259765, 0.073263427734375, 0.0735179214477539, 0.07310777282714843, 0.07434508514404296, 0.07346723175048828, 0.07628633880615235, 0.07550704193115235, 0.07370787048339844]",tokens/s,3460.3963285661457,kWh,2.16543798030839e-06,2.3880994990924374e-07,8.490068520443321e-07,3.253254782261966e-06,tokens/kWh,78690424.55445342,MB,1848.655872,941.490176,0.0,517.996544,433.41312,s,10,45.4326416015625,4.54326416015625,0.0071401817434617754,4.542790771484375,4.55135830078125,4.555304150390625,4.558460830078125,"[4.5426181640625, 4.53997998046875, 4.54296337890625, 4.5368056640625, 4.5445048828125, 4.531197265625, 4.55925, 4.54297119140625, 4.54186962890625, 4.5504814453125]",tokens/s,13.866682142874415,kWh,0.00013234953523760985,1.4598485911666738e-05,4.482989080215702e-05,0.00019177791195143363,tokens/kWh,328504.9845362499,,s,630,45.425543342590295,0.07210403705173069,0.0007504465010481083,0.07193017578125,0.07262707443237305,0.07305258026123047,0.07548663970947266,"[0.07269197082519531, 0.07283574676513672, 0.07184284973144531, 0.07167401885986328, 0.07168259429931641, 0.0717619171142578, 0.07178828430175781, 0.07218531036376953, 0.07222541046142578, 0.07196688079833985, 0.0715736312866211, 0.07173423767089844, 0.07143014526367188, 0.07230934143066406, 0.07209142303466796, 0.0719850845336914, 0.07181372833251953, 0.07395935821533203, 0.0722957763671875, 0.07240137481689453, 0.07250771331787109, 0.07282249450683594, 0.07248713684082031, 0.07242352294921875, 0.07206697845458984, 0.07197468566894531, 0.07155558776855468, 0.07176367950439454, 0.07164518737792969, 0.07182364654541015, 0.07199078369140625, 0.07266905975341797, 0.07208930969238281, 0.07222771453857423, 0.07198092651367187, 0.0716523208618164, 0.07396620941162109, 0.07392905426025391, 0.07204278564453125, 0.07179974365234375, 0.07175651550292969, 0.07251148986816407, 0.0719232940673828, 0.07180754852294922, 0.07158972930908203, 0.07180841827392578, 0.07160291290283204, 0.07158930969238281, 0.07160262298583984, 0.07154390716552735, 0.07245616149902344, 0.07213152313232422, 0.0718704605102539, 0.07261090850830078, 0.07189199829101563, 0.07253916931152343, 0.0715863037109375, 0.07239920043945312, 0.07262329864501953, 0.07177641296386719, 0.07167247772216796, 0.07136460876464844, 0.07157574462890624, 0.07166445159912109, 0.07178569793701171, 0.0715532455444336, 0.07169062042236328, 0.07174553680419922, 0.07273267364501954, 0.07170835113525391, 0.07194866943359375, 0.07191712188720703, 0.07176230621337891, 0.07164415740966797, 0.07175196838378907, 0.07164611053466798, 0.07187564849853516, 0.07308159637451171, 0.07213260650634766, 0.0719359359741211, 0.07180921936035156, 0.0714750747680664, 0.07196214294433594, 0.07228463745117188, 0.07215513610839844, 0.07163442993164063, 0.07193350219726563, 0.0716333770751953, 0.07166000366210938, 0.07208060455322265, 0.07185897827148438, 0.07219548797607422, 0.07233500671386718, 0.07234860992431641, 0.0760442886352539, 0.07224867248535156, 0.07303030395507812, 0.07207475280761719, 0.07195699310302735, 0.07195648193359375, 0.07195852661132812, 0.07191942596435547, 0.07154707336425781, 0.0717110366821289, 0.07273439788818359, 0.07271218872070312, 0.07232022094726563, 0.07235670471191406, 0.07190742492675781, 0.07179964447021485, 0.0717927703857422, 0.07189183807373047, 0.07173017883300781, 0.07171174621582031, 0.07193389129638672, 0.07180422210693359, 0.07151248168945312, 0.07141766357421875, 0.07168899536132813, 0.07213750457763672, 0.07243641662597657, 0.072157470703125, 0.07181478118896484, 0.07245353698730468, 0.0721204833984375, 0.07244473266601563, 0.07152025604248047, 0.07195177459716796, 0.07220285034179688, 0.07171603393554687, 0.07682643127441406, 0.07211036682128906, 0.07194601440429688, 0.07236271667480469, 0.0720931167602539, 0.07211497497558594, 0.07182915496826171, 0.07170604705810547, 0.07201190185546875, 0.07206578826904297, 0.07189814758300782, 0.07190841674804688, 0.07174944305419922, 0.0716063995361328, 0.07154278564453125, 0.07179190063476562, 0.07191126251220703, 0.07212278747558594, 0.07193238067626953, 0.07191069030761718, 0.0723397445678711, 0.07157926177978516, 0.07170934295654297, 0.07148592376708984, 0.0731355209350586, 0.071604736328125, 0.071932861328125, 0.071678466796875, 0.07158204650878906, 0.07171868896484375, 0.07233110046386719, 0.07216575622558594, 0.0720211181640625, 0.07222169494628906, 0.0722103042602539, 0.0717619171142578, 0.07187843322753906, 0.0715820770263672, 0.07259699249267579, 0.07175811004638671, 0.0717410888671875, 0.07162716674804688, 0.07160655975341797, 0.07306211090087891, 0.07337165069580077, 0.07235145568847656, 0.07162726593017578, 0.07214498901367188, 0.07275897979736329, 0.0718449935913086, 0.07178739166259765, 0.07152639770507813, 0.07618329620361328, 0.07196192169189453, 0.07171990203857422, 0.07166130828857421, 0.071563232421875, 0.07171711730957031, 0.07187881469726562, 0.072067138671875, 0.0720132827758789, 0.07235395050048828, 0.07232752227783203, 0.07194016265869141, 0.0718431396484375, 0.07174214172363282, 0.07194624328613282, 0.07177625274658203, 0.0718807373046875, 0.07187042999267577, 0.07171481323242188, 0.07173693084716797, 0.07228617858886718, 0.07165555572509766, 0.07203206634521485, 0.07280390167236328, 0.07191238403320313, 0.07218390655517579, 0.071810302734375, 0.07171520233154297, 0.07170035552978515, 0.07209935760498047, 0.07182015991210937, 0.0719834213256836, 0.07190290832519532, 0.07183987426757812, 0.07218982696533204, 0.07203401947021484, 0.07156764984130859, 0.07219744110107422, 0.07180563354492188, 0.0720169906616211, 0.07186319732666016, 0.07208566284179688, 0.0719502716064453, 0.07176316833496094, 0.07170486450195312, 0.07192822265625, 0.07173929595947266, 0.07257459259033203, 0.07216995239257812, 0.07392256164550781, 0.07292518615722657, 0.07184989166259766, 0.0723436508178711, 0.07170051574707031, 0.07174889373779297, 0.07169529724121093, 0.07165484619140625, 0.07177772521972656, 0.07193312072753906, 0.07198880004882813, 0.07182166290283203, 0.07194537353515625, 0.07287359619140625, 0.07187967681884766, 0.07186678314208984, 0.0718124771118164, 0.07251999664306641, 0.07159571075439453, 0.0718333740234375, 0.07188857269287109, 0.07138854217529297, 0.0717239990234375, 0.07171993255615235, 0.07171366119384766, 0.07182268524169921, 0.07226390075683593, 0.07167852783203126, 0.07349043273925782, 0.0718006362915039, 0.07218003082275391, 0.07184156799316406, 0.07164118194580078, 0.07162262725830078, 0.07236393737792969, 0.07158735656738281, 0.07542540740966797, 0.07282173156738281, 0.07178768157958984, 0.07169084930419922, 0.07227439880371093, 0.0720557098388672, 0.0716522216796875, 0.07214598083496093, 0.07193901062011719, 0.07204601287841797, 0.07156956481933593, 0.07158563232421875, 0.07163750457763672, 0.07186051177978516, 0.07212163543701172, 0.07170499420166015, 0.07139315032958984, 0.07164848327636719, 0.07232361602783204, 0.07173506927490235, 0.07215952301025391, 0.07153727722167968, 0.0719046401977539, 0.07332937622070312, 0.07261357116699219, 0.0717392349243164, 0.07185324859619141, 0.07164002990722657, 0.07228211212158203, 0.07196025848388672, 0.0725524444580078, 0.07202137756347657, 0.07262509155273437, 0.07197420501708984, 0.07213536071777343, 0.07176214599609375, 0.07180675506591797, 0.07304093170166015, 0.0721561279296875, 0.07174348449707031, 0.07168390655517579, 0.07177190399169922, 0.07188838195800781, 0.07408252716064453, 0.07265353393554687, 0.07243929290771485, 0.07254041290283203, 0.07360499572753906, 0.07190828704833985, 0.07256297302246094, 0.07213648223876953, 0.07269347381591797, 0.07194633483886718, 0.07176806640625, 0.07161212921142578, 0.07287359619140625, 0.071783203125, 0.07179264068603515, 0.072042236328125, 0.07261824035644532, 0.07149568176269532, 0.07199942779541016, 0.0713605728149414, 0.07143424224853516, 0.07151411437988281, 0.07211007690429687, 0.07147846221923829, 0.07144073486328124, 0.07169712066650391, 0.07174944305419922, 0.07182972717285156, 0.0718226547241211, 0.07176003265380859, 0.07219181060791016, 0.07232518768310547, 0.07172953796386719, 0.07160006713867187, 0.07162876892089844, 0.07154051208496094, 0.0720095977783203, 0.07174400329589843, 0.07168508911132812, 0.07155513763427734, 0.07267151641845702, 0.0718629150390625, 0.0715489273071289, 0.07191065979003906, 0.07206169891357422, 0.0720805435180664, 0.07149388885498047, 0.07156591796875, 0.07210428619384765, 0.07172262573242187, 0.0716731185913086, 0.07197779083251953, 0.07186335754394531, 0.07163510131835937, 0.07213664245605468, 0.07143103790283203, 0.07160348510742187, 0.07249369812011719, 0.07321981048583984, 0.07342899322509766, 0.07156684875488281, 0.07204940795898437, 0.07186434936523438, 0.07171891021728516, 0.07158758544921875, 0.07206025695800782, 0.07184003448486329, 0.07189161682128906, 0.07213670349121094, 0.07200109100341796, 0.07184783935546875, 0.07251203155517578, 0.07214105224609375, 0.07198047637939453, 0.07238880157470703, 0.08147379302978516, 0.07240185546875, 0.07186262512207031, 0.07262633514404297, 0.07197071838378906, 0.07180902099609375, 0.07181267547607421, 0.07193289947509765, 0.07168032073974609, 0.07184761810302734, 0.0720479965209961, 0.07258790588378906, 0.07205801391601563, 0.07200764465332031, 0.07198194885253906, 0.07181311798095703, 0.07221043395996093, 0.07192556762695312, 0.07236943817138672, 0.07575440216064454, 0.07207974243164063, 0.07151577758789063, 0.0716339874267578, 0.07150918579101563, 0.07196646118164063, 0.07169599914550781, 0.07173990631103516, 0.07197196960449219, 0.07186016082763672, 0.07176239776611328, 0.071723388671875, 0.07166767883300781, 0.07186656188964843, 0.07208284759521484, 0.07193436431884766, 0.07175782775878906, 0.07227391815185547, 0.07226383972167968, 0.07235791778564453, 0.07266079711914063, 0.07172303771972656, 0.07235292816162109, 0.07551165008544922, 0.07276233673095703, 0.07261593627929687, 0.07238262176513671, 0.07252073669433594, 0.07251238250732422, 0.07244588470458985, 0.07213993835449219, 0.07268256378173828, 0.07248365020751953, 0.0724480972290039, 0.07211408233642579, 0.07192070770263671, 0.07241209411621094, 0.07239580535888672, 0.07203068542480469, 0.07196031951904297, 0.07193062591552735, 0.07494831848144531, 0.07420531463623047, 0.07192511749267579, 0.07235807800292969, 0.0717968978881836, 0.07257337951660156, 0.07210131072998047, 0.072229248046875, 0.07191567993164062, 0.07155535888671875, 0.07147084808349609, 0.07138304138183593, 0.07145881652832031, 0.0716053466796875, 0.07221340942382813, 0.071884765625, 0.07165487670898438, 0.07266716766357421, 0.07156822204589844, 0.07245404815673828, 0.07209859466552734, 0.07197100830078125, 0.0726529312133789, 0.07205372619628907, 0.07175721740722656, 0.07164985656738282, 0.07176953887939454, 0.07193122863769531, 0.07203068542480469, 0.07179110717773438, 0.07173085021972657, 0.07203376007080078, 0.07166655731201171, 0.07263372802734375, 0.07162124633789063, 0.0720404510498047, 0.07155228424072266, 0.07179952239990234, 0.0715120620727539, 0.07166770935058593, 0.07186431884765625, 0.07232835388183594, 0.07531613159179687, 0.07220134735107422, 0.07226156616210938, 0.07168294525146485, 0.0718194580078125, 0.07264620971679688, 0.07259571075439453, 0.07227597045898437, 0.0718356170654297, 0.07194815826416015, 0.07164534759521485, 0.07183679962158203, 0.07182015991210937, 0.07200966644287109, 0.0718351058959961, 0.07217231750488282, 0.07192572784423829, 0.07210345458984375, 0.07204940795898437, 0.07154611206054687, 0.07234841918945313, 0.07176576232910156, 0.07193567657470704, 0.07167884826660156, 0.07177552032470703, 0.07185008239746093, 0.0718382110595703, 0.07178800201416016, 0.07223929595947266, 0.07213481903076172, 0.07171686553955078, 0.07174348449707031, 0.0718397445678711, 0.07160355377197265, 0.07372252655029297, 0.07152432250976562, 0.07194617462158204, 0.07182959747314453, 0.07208771514892579, 0.07161980438232422, 0.07257254028320312, 0.07164415740966797, 0.07178377532958985, 0.07258121490478515, 0.07219638061523438, 0.07275360107421874, 0.07240892791748046, 0.07244598388671875, 0.07252579498291016, 0.07236720275878906, 0.07199833679199219, 0.07226131439208984, 0.07216681671142579, 0.07186502075195313, 0.07457817840576172, 0.07187782287597656, 0.07179933166503906, 0.07205935668945312, 0.07203171539306641, 0.0717907485961914, 0.07208911895751953, 0.07184774780273437, 0.07189180755615235, 0.07163494110107421, 0.07137891387939453, 0.07168386840820312, 0.07170492553710937, 0.07138294219970703, 0.0719257583618164, 0.07146470642089844, 0.07179692840576171, 0.07197100830078125, 0.07199295806884766, 0.07232310485839843, 0.07214102172851562, 0.07188684844970702, 0.07276246643066406, 0.0747491226196289, 0.07219602966308594, 0.07191846466064453, 0.07159264373779296, 0.07250739288330078, 0.07182166290283203, 0.0718289566040039, 0.07209184265136719, 0.0722896957397461, 0.071802978515625, 0.07249766540527344, 0.07236204528808594, 0.07209363555908203, 0.07274838256835937, 0.07209334564208984, 0.07184652709960937, 0.07163123321533203, 0.07185788726806641, 0.0747872314453125, 0.0720341796875, 0.07196393585205078, 0.07184867095947266, 0.07209568023681641, 0.07266925048828125, 0.07205632019042969, 0.07234162902832031, 0.07176025390625, 0.07626547241210938, 0.07185820770263672, 0.07190729522705078, 0.07185964965820313, 0.07206707000732422, 0.07238832092285157, 0.07203517150878906, 0.07348223876953125, 0.07288813018798829, 0.07242140960693359, 0.07210924530029297, 0.07241149139404297, 0.07192972564697266, 0.07159884643554687, 0.07201074981689454, 0.07162982177734376, 0.07167123413085938, 0.07151385498046875, 0.07174412536621094, 0.0718333740234375, 0.07156678771972656, 0.0716953582763672, 0.07181436920166015, 0.07293955230712891, 0.07322112274169922, 0.07208505249023438, 0.07215119934082032, 0.07146905517578125, 0.07202957153320312, 0.0718424301147461, 0.071729248046875, 0.07211414337158203, 0.07197430419921876, 0.07257142639160157, 0.07223091125488282, 0.07326512145996093, 0.07213222503662109, 0.0722745590209961, 0.07241811370849609]",tokens/s,13.868848970031387,, @@ -7656,7 +7656,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpopscd5st/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpdfqbrymo/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-7b-hf,meta-llama/Llama-2-7b-hf,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -7762,7 +7762,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpribbtf_j/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp0aewst31/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-4B,Qwen/Qwen1.5-4B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -8184,7 +8184,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp6mz_xx4e/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp43a_pfrk/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gptj,EleutherAI/gpt-j-6b,EleutherAI/gpt-j-6b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -8509,7 +8509,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 100133 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 106196 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-32B,Qwen/Qwen1.5-32B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -8632,7 +8632,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpjtgd_ott/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpd203gwcp/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen,Qwen/Qwen-14B,Qwen/Qwen-14B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -9216,7 +9216,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp1darq03_/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpqiw4fumo/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,xglm,facebook/xglm-7.5B,facebook/xglm-7.5B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -9251,7 +9251,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmprktvbds4/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp1prygjxw/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-14B,Qwen/Qwen1.5-14B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -9292,7 +9292,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 76295 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 65087 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-70b-hf,meta-llama/Llama-2-70b-hf,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -9508,7 +9508,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpmtrqz4wh/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpg7kek4pf/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt2,openai-community/gpt2-large,openai-community/gpt2-large,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -9845,7 +9845,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp6gbhgwfe/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp_qqc59t3/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,EleutherAI/gpt-neox-20b,EleutherAI/gpt-neox-20b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -10759,7 +10759,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpp9fr4y0y/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpxyn341kr/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gemma,google/gemma-7b,google/gemma-7b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -10942,7 +10942,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.42 GiB is free. Process 113930 has 13.32 GiB memory in use. Of the allocated memory 13.21 GiB is allocated by PyTorch, and 1.48 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.42 GiB is free. Process 119981 has 13.32 GiB memory in use. Of the allocated memory 13.21 GiB is allocated by PyTorch, and 1.48 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,opt,facebook/opt-350m,facebook/opt-350m,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -11261,7 +11261,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpkid7zve_/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpr8hd82mp/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,EleutherAI/pythia-410m,EleutherAI/pythia-410m,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -14231,7 +14231,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 162.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 58.12 MiB is free. Process 130966 has 14.68 GiB memory in use. Of the allocated memory 14.19 GiB is allocated by PyTorch, and 384.43 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 162.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 58.12 MiB is free. Process 136643 has 14.68 GiB memory in use. Of the allocated memory 14.19 GiB is allocated by PyTorch, and 384.43 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,stabilityai/stablelm-base-alpha-7b,stabilityai/stablelm-base-alpha-7b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -14416,7 +14416,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpd5cm698y/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpnl4e2k2j/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-7b-hf,meta-llama/Llama-2-7b-hf,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -14522,7 +14522,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp1ocwmqao/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpt04_zmv2/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-4B,Qwen/Qwen1.5-4B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -14944,7 +14944,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpx9is7iy8/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp07hf8wjo/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gptj,EleutherAI/gpt-j-6b,EleutherAI/gpt-j-6b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -15346,7 +15346,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 68.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 68.12 MiB is free. Process 80546 has 14.67 GiB memory in use. Of the allocated memory 14.27 GiB is allocated by PyTorch, and 293.36 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 68.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 68.12 MiB is free. Process 69413 has 14.67 GiB memory in use. Of the allocated memory 14.27 GiB is allocated by PyTorch, and 293.36 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neo,EleutherAI/gpt-neo-125m,EleutherAI/gpt-neo-125m,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -15460,7 +15460,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmptbxvghzz/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp0o9lt4x6/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen,Qwen/Qwen-14B,Qwen/Qwen-14B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -15613,7 +15613,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 422.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 328.12 MiB is free. Process 206612 has 14.42 GiB memory in use. Of the allocated memory 14.22 GiB is allocated by PyTorch, and 91.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 422.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 328.12 MiB is free. Process 212294 has 14.42 GiB memory in use. Of the allocated memory 14.22 GiB is allocated by PyTorch, and 91.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neo,EleutherAI/gpt-neo-1.3B,EleutherAI/gpt-neo-1.3B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -15810,7 +15810,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 28.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 4.12 MiB is free. Process 181447 has 14.73 GiB memory in use. Of the allocated memory 14.49 GiB is allocated by PyTorch, and 137.12 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 28.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 4.12 MiB is free. Process 187362 has 14.73 GiB memory in use. Of the allocated memory 14.49 GiB is allocated by PyTorch, and 137.12 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen,Qwen/Qwen-72B,Qwen/Qwen-72B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -15892,7 +15892,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 98.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 80.12 MiB is free. Process 126960 has 14.66 GiB memory in use. Of the allocated memory 14.13 GiB is allocated by PyTorch, and 428.13 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 98.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 80.12 MiB is free. Process 132620 has 14.66 GiB memory in use. Of the allocated memory 14.13 GiB is allocated by PyTorch, and 428.13 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2_moe,Qwen/Qwen1.5-MoE-A2.7B,Qwen/Qwen1.5-MoE-A2.7B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16150,7 +16150,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 280.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 232.12 MiB is free. Process 23491 has 14.51 GiB memory in use. Of the allocated memory 14.07 GiB is allocated by PyTorch, and 332.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 280.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 232.12 MiB is free. Process 24573 has 14.51 GiB memory in use. Of the allocated memory 14.07 GiB is allocated by PyTorch, and 332.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,EleutherAI/pythia-12b,EleutherAI/pythia-12b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16256,7 +16256,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpuctu4e2z/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp6_1ipt4z/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,xglm,facebook/xglm-7.5B,facebook/xglm-7.5B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16291,7 +16291,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpusmw0nve/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: XGLMForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpx2t161jx/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-14B,Qwen/Qwen1.5-14B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16409,7 +16409,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 110.12 MiB is free. Process 166388 has 14.63 GiB memory in use. Of the allocated memory 14.47 GiB is allocated by PyTorch, and 51.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 112.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 110.12 MiB is free. Process 172403 has 14.63 GiB memory in use. Of the allocated memory 14.47 GiB is allocated by PyTorch, and 51.96 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,Deci/DeciCoder-1b,Deci/DeciCoder-1b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16497,7 +16497,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 448.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 436.12 MiB is free. Process 172804 has 14.31 GiB memory in use. Of the allocated memory 14.14 GiB is allocated by PyTorch, and 61.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 448.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 436.12 MiB is free. Process 178717 has 14.31 GiB memory in use. Of the allocated memory 14.14 GiB is allocated by PyTorch, and 61.06 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,togethercomputer/RedPajama-INCITE-Base-7B-v0.1,togethercomputer/RedPajama-INCITE-Base-7B-v0.1,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16613,7 +16613,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 85456 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 74486 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,internlm,internlm/internlm-20b,internlm/internlm-20b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16691,7 +16691,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 34.12 MiB is free. Process 90324 has 14.71 GiB memory in use. Of the allocated memory 14.56 GiB is allocated by PyTorch, and 28.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 192.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 34.12 MiB is free. Process 79286 has 14.71 GiB memory in use. Of the allocated memory 14.56 GiB is allocated by PyTorch, and 28.11 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,codegen,Salesforce/codegen-16B-nl,Salesforce/codegen-16B-nl,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16726,7 +16726,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpnll99x97/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpsre_8wg7/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt2,openai-community/gpt2-large,openai-community/gpt2-large,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16842,7 +16842,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 302.12 MiB is free. Process 158176 has 14.44 GiB memory in use. Of the allocated memory 14.20 GiB is allocated by PyTorch, and 132.89 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 344.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 302.12 MiB is free. Process 164080 has 14.44 GiB memory in use. Of the allocated memory 14.20 GiB is allocated by PyTorch, and 132.89 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,huggyllama/llama-30b,huggyllama/llama-30b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -16889,7 +16889,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 54.12 MiB is free. Process 153268 has 14.69 GiB memory in use. Of the allocated memory 14.37 GiB is allocated by PyTorch, and 203.94 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 54.12 MiB is free. Process 159161 has 14.69 GiB memory in use. Of the allocated memory 14.37 GiB is allocated by PyTorch, and 203.94 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-13b-hf,meta-llama/Llama-2-13b-hf,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -17137,7 +17137,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpcbmocn0c/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: RecurrentGemmaForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp_y90vgww/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,EleutherAI/gpt-neox-20b,EleutherAI/gpt-neox-20b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -17255,7 +17255,7 @@ ChildProcessError: Traceback (most recent call last): w_4bit, quant_state = bnb.functional.quantize_4bit( File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/functional.py"", line 1196, in quantize_4bit out = torch.zeros(((n + 1) // mod, 1), dtype=quant_storage, device=A.device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 38.12 MiB is free. Process 186689 has 14.70 GiB memory in use. Of the allocated memory 14.42 GiB is allocated by PyTorch, and 176.08 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 48.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 38.12 MiB is free. Process 192679 has 14.70 GiB memory in use. Of the allocated memory 14.42 GiB is allocated by PyTorch, and 176.08 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,01-ai/Yi-6B,01-ai/Yi-6B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -17868,7 +17868,7 @@ ChildProcessError: Traceback (most recent call last): return self._quantize(device) File ""/usr/local/lib/python3.10/dist-packages/bitsandbytes/nn/modules.py"", line 296, in _quantize w = self.data.contiguous().to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 104389 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 384.00 MiB. GPU 0 has a total capacity of 14.74 GiB of which 272.12 MiB is free. Process 110531 has 14.47 GiB memory in use. Of the allocated memory 14.35 GiB is allocated by PyTorch, and 13.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-0.5B,Qwen/Qwen1.5-0.5B,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -18187,7 +18187,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpscdauctr/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: FalconForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmpn_vssxhf/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gemma,google/gemma-7b,google/gemma-7b,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -18370,7 +18370,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.52 GiB is free. Process 113485 has 13.22 GiB memory in use. Of the allocated memory 13.10 GiB is allocated by PyTorch, and 6.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.52 GiB is free. Process 119493 has 13.22 GiB memory in use. Of the allocated memory 13.10 GiB is allocated by PyTorch, and 6.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,opt,facebook/opt-350m,facebook/opt-350m,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -18689,7 +18689,7 @@ ChildProcessError: Traceback (most recent call last): cls._check_and_enable_flash_attn_2( File ""/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py"", line 1688, in _check_and_enable_flash_attn_2 raise ValueError( -ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp8ul1y7ue/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new +ValueError: CodeGenForCausalLM does not support Flash Attention 2.0 yet. Please request to add support where the model is hosted, on its model hub page: https://huggingface.co//tmp/tmp01hx8n6z/no_weights_model/discussions/new or in the Transformers GitHub repo: https://github.com/huggingface/transformers/issues/new ",True,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 4bit-bnb-flash_attention_2,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,gpt_neox,EleutherAI/pythia-410m,EleutherAI/pythia-410m,cuda,0,42,,,True,True,,float16,True,False,,flash_attention_2,,False,,False,forward,bnb,0.0,,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -19302,7 +19302,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 98370 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 104446 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,qwen2,Qwen/Qwen1.5-32B,Qwen/Qwen1.5-32B,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -19673,7 +19673,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 62938 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.45 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.13 GiB is free. Process 63312 has 13.61 GiB memory in use. Of the allocated memory 13.36 GiB is allocated by PyTorch, and 136.95 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,llama,meta-llama/Llama-2-70b-hf,meta-llama/Llama-2-70b-hf,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last): @@ -20095,7 +20095,7 @@ ChildProcessError: Traceback (most recent call last): set_module_tensor_to_device(module, name, self.execution_device, tied_params_map=self.tied_params_map) File ""/usr/local/lib/python3.10/dist-packages/accelerate/utils/modeling.py"", line 408, in set_module_tensor_to_device new_value = old_value.to(device) -torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.42 GiB is free. Process 112908 has 13.32 GiB memory in use. Of the allocated memory 13.21 GiB is allocated by PyTorch, and 1.48 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) +torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 1.97 GiB. GPU 0 has a total capacity of 14.74 GiB of which 1.42 GiB is free. Process 118948 has 13.32 GiB memory in use. Of the allocated memory 13.21 GiB is allocated by PyTorch, and 1.48 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) ",,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 8bit-bnb-sdpa,pytorch,2.4.1+cu124,optimum_benchmark.backends.pytorch.backend.PyTorchBackend,text-generation,transformers,opt,facebook/opt-350m,facebook/opt-350m,cuda,0,42,,,True,True,,float16,True,False,,sdpa,,False,,False,forward,bnb,0.0,True,False,,inference,optimum_benchmark.scenarios.inference.scenario.InferenceScenario,10,10,10,1,2,256,,True,True,True,64,64,process,optimum_benchmark.launchers.process.launcher.ProcessLauncher,True,kill,False,spawn, Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz,8,33163.759616,Linux,x86_64,Linux-5.10.225-213.878.amzn2.x86_64-x86_64-with-glibc2.35,x86_64,3.10.12,['Tesla T4'],1,16106127360,0.5.0,,4.45.1,,0.34.2,,,,1.22.0,,,,0.13.0,,"Traceback (most recent call last):