aws-neuron/Qwen3-1.7B-TP4-BS4-SEQ2048
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: aws-neuron/Qwen3-1.7B-TP4-BS4-SEQ2048
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-08-30 21:56:23 +08:00
commit fd57cdfd17
114 changed files with 237628 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
token_generation_model/_tp0_bk3/command.txt Normal file
View File

@@ -0,0 +1 @@
neuronx-cc compile --framework=XLA model.MODULE_02e08d5fdadfbcb8a569+a6f96e4f.hlo_module.pb --output model.MODULE_02e08d5fdadfbcb8a569+a6f96e4f.neff --target=trn2 --auto-cast=none --model-type=transformer '--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=1 --vectorize-strided-dma ' --lnc=2 -O2 --internal-hlo2tensorizer-options=--verify-hlo=true --logfile=log-neuron-cc.txt --verbose=35

1
token_generation_model/_tp0_bk3/compile_flags.MODULE_02e08d5fdadfbcb8a569+a6f96e4f.json Normal file
View File

@@ -0,0 +1 @@
["--target=trn2", "--auto-cast=none", "--model-type=transformer", "--tensorizer-options=--enable-ccop-compute-overlap --cc-pipeline-tiling-factor=1 --vectorize-strided-dma ", "--lnc=2", "-O2", "--internal-hlo2tensorizer-options=--verify-hlo=true", "--logfile=/models/Qwen3-1.7B-compiled/token_generation_model/_tp0_bk3/log-neuron-cc.txt"]

590
token_generation_model/_tp0_bk3/global_metric_store.json Normal file
View File

@@ -0,0 +1,590 @@
{
"Average": {
"tensorizer": {
"StaticProfiler::AverageFractalPeUtilization": 97.88961791992188,
"StaticProfiler::AveragePartitionUtilization": 89.79463958740234,
"StaticProfiler::AveragePeUtilization": 78.904052734375,
"StaticProfiler::LocalizationEfficiency": 146.6905975341797,
"StaticProfiler::LocalizationEfficiencyIgnoreNonlocal": 151.06674194335938,
"TilingProfiler::AveragePartitionUtilizationAfterTiling": 0,
"TilingProfiler::AveragePeUtilizationAfterTiling": 0
}
},
"Count": {
"tensorizer": {
"StaticProfiler::AverageFractalPeUtilization": 1,
"StaticProfiler::AveragePartitionUtilization": 1,
"StaticProfiler::AveragePeUtilization": 1,
"StaticProfiler::LocalizationEfficiency": 1,
"StaticProfiler::LocalizationEfficiencyIgnoreNonlocal": 1,
"TilingProfiler::AveragePartitionUtilizationAfterTiling": 1,
"TilingProfiler::AveragePeUtilizationAfterTiling": 1
}
},
"Sum": {
"compiletime": {
"AGOrderingAnalysisPass": 6.388453960418701,
"AffinePredicateResolution": 0.40378904342651367,
"AliasDependencyElimination": 0.016310930252075195,
"AliasDependencyInduction": 15.282819747924805,
"AliasDependencyReset": 16.618453979492188,
"BFComputeCutting": 0.1391315460205078,
"BirCodeGenLoop": 4.402980804443359,
"CCOpFusion": 0.8881416320800781,
"CanonicalizeConv": 4.999999873689376e-06,
"CanonicalizeDAGForPGTiling": 0.20721173286437988,
"CanonicalizeForTensorizer": 0.00076299998909235,
"CanonicalizeIR": 1.028043270111084,
"Canonicalizer": 0.08714199811220169,
"CoalesceCCOp": 0.2357621192932129,
"CommuteConcat": 0.0287783145904541,
"DMALocalityOpt": 0.07430553436279297,
"DMAProfiler": 0.08057951927185059,
"DMATilingProfiler": 0.09187006950378418,
"DataLocalityOpt": 2.820284605026245,
"DataStreaming": 0.4295690059661865,
"DeConcat": 0.06640625,
"DeadCodeElimination": 0.053946495056152344,
"DeadStoreElimination": 1.0540196895599365,
"DelinearIndices": 0.3954615592956543,
"Delinearization": 0.13945698738098145,
"DelinearizeSPMD": 0.17487859725952148,
"DoNothing": 0.0004494190216064453,
"DramToDramTranspose": 0.34574389457702637,
"DumpGraphAndMetadata": 0.23639750480651855,
"EliminateDivs": 1.4790117740631104,
"ExpandBatchNorm": 1.123610019683838,
"ExpandISAMacro": 0.12250304222106934,
"FactorizeBlkDims": 0.49267148971557617,
"FactorizeThreadAxesInFreeDims": 0.08273863792419434,
"FlattenMacroLoop": 0.08472108840942383,
"GenericAccessSimplifier": 0.025115966796875,
"HoistCompute": 0.0,
"IdentifyCrossPassTensors": 0.0003549999964889139,
"InferInitValue": 1.3054986000061035,
"InferIntrinsicOnCC": 0.7679235935211182,
"InferNeuronTensor": 1.671952724456787,
"InferNonlocalTensors": 7.110470771789551,
"InferPSumTensor": 4.2829790115356445,
"InferShardAxis": 12.786405563354492,
"InferSharedMemLoc": 0.16948676109313965,
"InlineNativeKernels": 0.05566692352294922,
"InsertCoreBarrier": 0.15158963203430176,
"InsertIOTransposes": 2.3931386470794678,
"InsertImplicitShardAxisBeforeISel": 0.5525217056274414,
"InsertLocalTransposes": 0.7429313659667969,
"InsertOffloadedTransposes": 0.19022369384765625,
"LICM": 0.14770793914794922,
"LateLegalizeInst": 0.31015753746032715,
"LateLegalizePostSplit": 0.22165155410766602,
"LateLowerReshapeOp": 0.03522944450378418,
"LateLowerTensorOp": 6.813569068908691,
"LateNeuronInstComb": 1.2720398902893066,
"LayoutPreprocessing": 1.8736376762390137,
"LayoutPreprocessingAndAnalysis": 2.3256783485412598,
"LayoutRequirementAnalysis": 0.4448714256286621,
"LegalizeCCOpLayout": 1.317772388458252,
"LegalizeOpLevelAlias": 0.5106728076934814,
"LegalizePartitionReduce": 0.08073091506958008,
"LegalizeSundaAccess": 0.9463906288146973,
"LegalizeSundaMacro": 0.7451527118682861,
"LegalizeType": 0.13950800895690918,
"LocalLayoutOpt": 0.6608095169067383,
"LoopFusion": 0.3101928234100342,
"LoopSplitting": 0.036502838134765625,
"LowerBroadcast": 0.06515121459960938,
"LowerCCOpBlockAxis": 0.32977890968322754,
"LowerComplexBroadcast": 0.15041875839233398,
"LowerIntrinsics": 1.0637662410736084,
"LowerShardAxis": 0.4116206169128418,
"LowerTensorOp": 3.285423755645752,
"LowerToSendRecv": 0.25904369354248047,
"LowerTranspose": 0.5238301753997803,
"MacroGeneration": 2.1192569732666016,
"MaskPropagation": 0.14647436141967773,
"MemcastMotion": 0.0,
"MemcpyElimination": 29.3009090423584,
"MutateDataType": 0.037621498107910156,
"NeuronAliasDependencyInduction": 0.06313967704772949,
"NeuronAliasDependencyReset": 0.08403873443603516,
"NeuronInstComb": 0.3617970943450928,
"NeuronLICM": 0.2975320816040039,
"NeuronLoopFusion": 1.491703987121582,
"NeuronLoopInterchange": 0.06900668144226074,
"NeuronSimplifier": 0.5315425395965576,
"NeuronSimplifyPredicates": 0.2570762634277344,
"NeuronValueNumbering": 0.11507940292358398,
"OptimizeAliasedCopyChain": 0.6118557453155518,
"OptimizeNKIKernels": 1.0786433219909668,
"PAGLayoutOpt": 18.570398330688477,
"PComputeCutting": 0.6222004890441895,
"PGLayoutTilingPipeline": 55.62285614013672,
"PGTiling": 10.0595703125,
"PadElimination": 0.021188735961914063,
"ParAxesAnnotation": 17.818796157836914,
"PartialLoopFusion": 1.5088038444519043,
"PartialSimdFusion": 0.8573262691497803,
"PenguinizeFunctions": 0.000754999986384064,
"PerfectLoopNest": 0.07143402099609375,
"PruneFunctions": 0.0012410000199452043,
"RecognizeOpIdiom": 0.1291654109954834,
"Recompute": 0.007833480834960938,
"RelaxPredicates": 0.14508056640625,
"Rematerialization": 0.17157793045043945,
"RemoveOptimizationBarriers": 0.0005370000144466758,
"RemoveShardedPartitionAxes": 1.3566563129425049,
"ReshapeWeights": 0.022662878036499023,
"ResolveAccessConflict": 0.2033238410949707,
"ResolveComplicatePredicates": 0.49439358711242676,
"RewriteReplicationMatmul": 0.03938794136047363,
"RewriteWeights": 0.07809877395629883,
"SFKVectorizer": 10.350098609924316,
"ScatterMotion": 0.0,
"ShardingPropagationAnalysis": 0.7199399471282959,
"SimpleAllReduceTiling": 0.07370853424072266,
"Simplifier": 0.09367704391479492,
"SimplifyMacroPredicates": 0.28262877464294434,
"SimplifyNeuronTensor": 0.9568257331848145,
"SimplifySlice": 0.027447223663330078,
"SimplifyTensor": 0.2734675407409668,
"SpillPSum": 1.2486350536346436,
"SplitAPUnionSets": 0.8627588748931885,
"SplitAccGrp": 0.1022341251373291,
"StaticProfiler": 0.4753732681274414,
"StaticTransposeLocalTensor": 0.6612956523895264,
"SundaISel": 1.6735970973968506,
"TCTransform": 0.029105186462402344,
"TensorInitialization": 0.1944262981414795,
"TensorOpSimplifier": 3.3376035690307617,
"TensorOpTransform": 19.926769256591797,
"TensorizerLegalizationPass": 0.0006489999941550195,
"TileCCOps": 0.1856687068939209,
"TilingProfiler": 0.5146362781524658,
"TransformConvOp": 1.4806337356567383,
"TritiumFusion": 0.2828052043914795,
"ValueNumbering": 0.08587479591369629,
"VectorizeDMA": 0.5646088123321533,
"VectorizeMatMult": 0.08723974227905273,
"VerifySupportedOps": 0.0004870000120718032,
"WeightCoalescing": 0.06492829322814941,
"ZeroSizeTensorElimination": 0.0008580684661865234,
"algsimp": 0.0014750000555068254,
"batchnorm_expander": 0.0015930000226944685,
"boundary-marker-removal": 0.0007229999755509198,
"call-inliner": 0.00018899999849963933,
"canonicalize-boundary-marker": 0.0011119999689981341,
"collective-stream-id-checker": 8.900000102585182e-05,
"comparison-expander": 0.0007440000190399587,
"computation-deduplicator": 0.0037249999586492777,
"config-lowering": 0.06129100173711777,
"constant_folding": 0.00012700000661425292,
"cse": 0.0007800000021234155,
"dce": 5.199999941396527e-05,
"dynamic-slice-transpose": 0.0002530000056140125,
"eliminate-redundant-compare": 0.0001140000022132881,
"emit-offloaded-dropout": 0.0006549999816343188,
"flatten-call-graph": 0.0013830000534653664,
"fuse-send-recv": 0.07738199830055237,
"hilo-conditional-to-select": 0.00014400000509340316,
"hilo::LegalizeAlias": 0.004616000223904848,
"hilo::NeuronInstCombine": 0.0015979999443516135,
"hilo::NeuronOpFusion": 0.0001250000059371814,
"hilo::ReplaceTokenTypeWithU8Pass": 0.0016670000040903687,
"hilo::ScheduleFusion": 9.999999747378752e-06,
"hilo::SixtyFourHack": 0.0008529999759048223,
"hilo::VerifyAliasing": 0.000291000003926456,
"hlo-mac-count": 0.02539600059390068,
"io-con-pipe-begin": 7.999999979801942e-06,
"io-con-pipe-end": 0.0,
"io-layout-normalization": 0.0035389999393373728,
"legalize-ccops-for-tensorizer": 0.00014200000441633165,
"legalize-compare": 0.0007239999831654131,
"lower-argminmax-custom-call": 0.0003229999856557697,
"map-inline": 0.0010969999711960554,
"metadata-naming": 0.009553000330924988,
"mlir::detail::OpToOpPassAdaptor": 9.999999747378752e-05,
"mlir::hlo::MhloToPyPenguin": 0.6092489957809448,
"mlir::mhlo::LowerComplexExtraPass": 0.0049149999395012856,
"mlir::mhlo::LowerComplexPass": 0.008914999663829803,
"native-to-custom-softmax": 0.0010550000006332994,
"native-to-custom-softmax-dx": 0.001585999969393015,
"neuron-hlo-verifier": 0.22506099939346313,
"operand_upcaster": 0.003352999920025468,
"post-par-pipe-begin": 9.999999974752427e-07,
"post-par-pipe-end": 0.0,
"post-partition-simplification": 0.37791401147842407,
"pre-hlo-begin": 3.999999989900971e-06,
"pre-hlo-end": 9.999999974752427e-07,
"replace-minimum-constant": 0.00019299999985378236,
"reshape-mover": 5.2999999752501026e-05,
"simplify-concat": 0.06822799891233444,
"simplify-while-loops": 5.0999999075429514e-05,
"transform-variadic-reduce": 0.0008849999867379665,
"tuple-simplifier": 0.00012099999730708078,
"unpack-nested-aws-ntwsr": 0.000615999975707382,
"unroll-while-loop": 7.000000096013537e-06
},
"hilo": {
"HloMacCount": 1854726144.0,
"Traffic": 1252630656.0
},
"tensorizer": {
"DMATilingProfiler::TotalInstructionsAfterTiling": 51674,
"StaticProfiler::AifUb": 21.735599517822266,
"StaticProfiler::ArithmeticIntensityTensorizer": 31.884082794189453,
"StaticProfiler::AverageDmaLength": 2843.9130859375,
"StaticProfiler::DDRTransferBytes": 1013018224,
"StaticProfiler::InternalTransferBytes": 226623072,
"StaticProfiler::LoadExpanded": 281500,
"StaticProfiler::StoreExpanded": 4273,
"StaticProfiler::TotalDMAExpanded": 285773,
"StaticProfiler::TotalDynamicInstancesCount": 67989,
"StaticProfiler::TotalDynamicInstancesWithMmPackedCount": 57540,
"StaticProfiler::TotalLNCComm": 0,
"StaticProfiler::TotalLNCCommTransfer": 0,
"TilingProfiler::BatchnormInstructionsAfterTiling": 0,
"TilingProfiler::DmaInstructionsAfterTiling": 0,
"TilingProfiler::GenericInstructionsAfterTiling": 123,
"TilingProfiler::MatMultInstructionsAfterTiling": 35008,
"TilingProfiler::NumPfTransposes": 292,
"TilingProfiler::NumPfTransposesForIo": 30,
"TilingProfiler::NumPfTransposesForLocal": 142,
"TilingProfiler::NumPfTransposesForNonlocal": 120,
"TilingProfiler::PfTransposeInstructions": 10670,
"TilingProfiler::PfTransposeInstructionsForIo": 8360,
"TilingProfiler::PfTransposeInstructionsForLocal": 648,
"TilingProfiler::PfTransposeInstructionsForNonlocal": 1662,
"TilingProfiler::ReduceInstructionsAfterTiling": 283,
"TilingProfiler::SimdInstructionsAfterTiling": 2805,
"TilingProfiler::TotalInstructionsAfterTiling": 0,
"TransformConvOp::Conv1d_depthwise_bf01_oi01_bf01": 0,
"TransformConvOp::Conv2d_dw_fb01_io01_01bf_rep_nhwc_Pcinh": 0,
"TransformConvOp::Conv2d_pbp_0f1b_0i1o_01fb_experimental_1": 0,
"TransformConvOp::Conv2d_pbp_fb01_io01_01bf_experimental_1": 0,
"TransformConvOp::conv2d_column_packing": 0,
"TransformConvOp::conv2d_column_packing_1": 0,
"TransformConvOp::conv2d_column_packing_io10": 0,
"TransformConvOp::conv2d_depthwise_f01b_o01i_bf01": 0
}
},
"all": {
"compiletime": {
"CanonicalizeConv": 4.999999873689376e-06,
"CanonicalizeForTensorizer": 0.00076299998909235,
"Canonicalizer": 0.08714199811220169,
"HoistCompute": 0.0,
"IdentifyCrossPassTensors": 0.0003549999964889139,
"MemcastMotion": 0.0,
"PenguinizeFunctions": 0.000754999986384064,
"PruneFunctions": 0.0012410000199452043,
"RemoveOptimizationBarriers": 0.0005370000144466758,
"ScatterMotion": 0.0,
"TensorizerLegalizationPass": 0.0006489999941550195,
"VerifySupportedOps": 0.0004870000120718032,
"algsimp": 0.0014750000555068254,
"batchnorm_expander": 0.0015930000226944685,
"boundary-marker-removal": 0.0007229999755509198,
"call-inliner": 0.00018899999849963933,
"canonicalize-boundary-marker": 0.0011119999689981341,
"collective-stream-id-checker": 8.900000102585182e-05,
"comparison-expander": 0.0007440000190399587,
"computation-deduplicator": 0.0037249999586492777,
"config-lowering": 0.06129100173711777,
"constant_folding": 0.00012700000661425292,
"cse": 0.0007800000021234155,
"dce": 5.199999941396527e-05,
"dynamic-slice-transpose": 0.0002530000056140125,
"eliminate-redundant-compare": 0.0001140000022132881,
"emit-offloaded-dropout": 0.0006549999816343188,
"flatten-call-graph": 0.0013830000534653664,
"fuse-send-recv": 0.07738199830055237,
"hilo-conditional-to-select": 0.00014400000509340316,
"hilo::LegalizeAlias": 0.004616000223904848,
"hilo::NeuronInstCombine": 0.0015979999443516135,
"hilo::NeuronOpFusion": 0.0001250000059371814,
"hilo::ReplaceTokenTypeWithU8Pass": 0.0016670000040903687,
"hilo::ScheduleFusion": 9.999999747378752e-06,
"hilo::SixtyFourHack": 0.0008529999759048223,
"hilo::VerifyAliasing": 0.000291000003926456,
"hlo-mac-count": 0.02539600059390068,
"io-con-pipe-begin": 7.999999979801942e-06,
"io-con-pipe-end": 0.0,
"io-layout-normalization": 0.0035389999393373728,
"legalize-ccops-for-tensorizer": 0.00014200000441633165,
"legalize-compare": 0.0007239999831654131,
"lower-argminmax-custom-call": 0.0003229999856557697,
"map-inline": 0.0010969999711960554,
"metadata-naming": 0.009553000330924988,
"mlir::detail::OpToOpPassAdaptor": 9.999999747378752e-05,
"mlir::hlo::MhloToPyPenguin": 0.6092489957809448,
"mlir::mhlo::LowerComplexExtraPass": 0.0049149999395012856,
"mlir::mhlo::LowerComplexPass": 0.008914999663829803,
"native-to-custom-softmax": 0.0010550000006332994,
"native-to-custom-softmax-dx": 0.001585999969393015,
"neuron-hlo-verifier": 0.22506099939346313,
"operand_upcaster": 0.003352999920025468,
"post-par-pipe-begin": 9.999999974752427e-07,
"post-par-pipe-end": 0.0,
"post-partition-simplification": 0.37791401147842407,
"pre-hlo-begin": 3.999999989900971e-06,
"pre-hlo-end": 9.999999974752427e-07,
"replace-minimum-constant": 0.00019299999985378236,
"reshape-mover": 5.2999999752501026e-05,
"simplify-concat": 0.06822799891233444,
"simplify-while-loops": 5.0999999075429514e-05,
"transform-variadic-reduce": 0.0008849999867379665,
"tuple-simplifier": 0.00012099999730708078,
"unpack-nested-aws-ntwsr": 0.000615999975707382,
"unroll-while-loop": 7.000000096013537e-06
}
},
"cumsum": {
"compiletime": {
"CoalesceCCOp": 0.00023365020751953125,
"DMALocalityOpt": 0.00019097328186035156,
"DMAProfiler": 0.0009503364562988281,
"DataStreaming": 0.00032329559326171875,
"DoNothing": 0.00015044212341308594,
"ExpandISAMacro": 0.0005333423614501953,
"FactorizeBlkDims": 0.0005185604095458984,
"InferPSumTensor": 0.0006201267242431641,
"InferSharedMemLoc": 0.0002682209014892578,
"InsertCoreBarrier": 0.0003275871276855469,
"LateLegalizeInst": 0.0004162788391113281,
"LateNeuronInstComb": 0.0005996227264404297,
"LegalizeSundaAccess": 0.0016045570373535156,
"LegalizeType": 0.0002541542053222656,
"LowerBroadcast": 0.00023365020751953125,
"LowerIntrinsics": 0.00023651123046875,
"LowerTranspose": 0.0002307891845703125,
"NeuronInstComb": 0.0006008148193359375,
"NeuronLICM": 0.00036597251892089844,
"NeuronSimplifyPredicates": 0.002484560012817383,
"NeuronValueNumbering": 0.0004477500915527344,
"SFKVectorizer": 0.002768278121948242,
"SimpleAllReduceTiling": 0.00021409988403320313,
"SimplifyNeuronTensor": 0.0005936622619628906,
"SpillPSum": 0.0005390644073486328,
"WeightCoalescing": 0.00023293495178222656
}
},
"sg00": {
"hilo": {
"ArithmeticIntensity": 2.961329698562622,
"HloMacCount": 1854726144.0,
"Traffic": 1252630656.0
}
},
"sg0000": {
"compiletime": {
"AGOrderingAnalysisPass": 6.388453960418701,
"AffinePredicateResolution": 0.40378904342651367,
"AliasDependencyElimination": 0.016310930252075195,
"AliasDependencyInduction": 15.282819747924805,
"AliasDependencyReset": 16.618453979492188,
"BFComputeCutting": 0.1391315460205078,
"BirCodeGenLoop": 4.402980804443359,
"CCOpFusion": 0.8881416320800781,
"CanonicalizeDAGForPGTiling": 0.20721173286437988,
"CanonicalizeIR": 1.028043270111084,
"CoalesceCCOp": 0.23186254501342773,
"CommuteConcat": 0.0287783145904541,
"DMALocalityOpt": 0.07128620147705078,
"DMAProfiler": 0.07605504989624023,
"DMATilingProfiler": 0.09187006950378418,
"DataLocalityOpt": 2.820284605026245,
"DataStreaming": 0.42258405685424805,
"DeConcat": 0.06640625,
"DeadCodeElimination": 0.053946495056152344,
"DeadStoreElimination": 1.0540196895599365,
"DelinearIndices": 0.3954615592956543,
"Delinearization": 0.13945698738098145,
"DelinearizeSPMD": 0.17487859725952148,
"DoNothing": 0.00014019012451171875,
"DramToDramTranspose": 0.34574389457702637,
"DumpGraphAndMetadata": 0.23639750480651855,
"EliminateDivs": 1.4790117740631104,
"ExpandBatchNorm": 1.123610019683838,
"ExpandISAMacro": 0.11814379692077637,
"FactorizeBlkDims": 0.4797940254211426,
"FactorizeThreadAxesInFreeDims": 0.08273863792419434,
"FlattenMacroLoop": 0.08472108840942383,
"GenericAccessSimplifier": 0.025115966796875,
"InferInitValue": 1.3054986000061035,
"InferIntrinsicOnCC": 0.7679235935211182,
"InferNeuronTensor": 1.671952724456787,
"InferNonlocalTensors": 7.110470771789551,
"InferPSumTensor": 4.2699432373046875,
"InferShardAxis": 12.786405563354492,
"InferSharedMemLoc": 0.16634631156921387,
"InlineNativeKernels": 0.05566692352294922,
"InsertCoreBarrier": 0.1479203701019287,
"InsertIOTransposes": 2.3931386470794678,
"InsertImplicitShardAxisBeforeISel": 0.5525217056274414,
"InsertLocalTransposes": 0.7429313659667969,
"InsertOffloadedTransposes": 0.19022369384765625,
"LICM": 0.14770793914794922,
"LateLegalizeInst": 0.30222225189208984,
"LateLegalizePostSplit": 0.22165155410766602,
"LateLowerReshapeOp": 0.03522944450378418,
"LateLowerTensorOp": 6.813569068908691,
"LateNeuronInstComb": 1.2631995677947998,
"LayoutPreprocessing": 1.8736376762390137,
"LayoutPreprocessingAndAnalysis": 2.3256783485412598,
"LayoutRequirementAnalysis": 0.4448714256286621,
"LegalizeCCOpLayout": 1.317772388458252,
"LegalizeOpLevelAlias": 0.5106728076934814,
"LegalizePartitionReduce": 0.08073091506958008,
"LegalizeSundaAccess": 0.9303104877471924,
"LegalizeSundaMacro": 0.7451527118682861,
"LegalizeType": 0.12957215309143066,
"LocalLayoutOpt": 0.6608095169067383,
"LoopFusion": 0.3101928234100342,
"LoopSplitting": 0.036502838134765625,
"LowerBroadcast": 0.06164741516113281,
"LowerCCOpBlockAxis": 0.32977890968322754,
"LowerComplexBroadcast": 0.15041875839233398,
"LowerIntrinsics": 1.0597736835479736,
"LowerShardAxis": 0.4116206169128418,
"LowerTensorOp": 3.285423755645752,
"LowerToSendRecv": 0.25904369354248047,
"LowerTranspose": 0.5202181339263916,
"MacroGeneration": 2.1192569732666016,
"MaskPropagation": 0.14647436141967773,
"MemcpyElimination": 29.3009090423584,
"MutateDataType": 0.037621498107910156,
"NeuronAliasDependencyInduction": 0.06313967704772949,
"NeuronAliasDependencyReset": 0.08403873443603516,
"NeuronInstComb": 0.3526570796966553,
"NeuronLICM": 0.28780031204223633,
"NeuronLoopFusion": 1.491703987121582,
"NeuronLoopInterchange": 0.06900668144226074,
"NeuronSimplifier": 0.5315425395965576,
"NeuronSimplifyPredicates": 0.25080394744873047,
"NeuronValueNumbering": 0.11057853698730469,
"OptimizeAliasedCopyChain": 0.6118557453155518,
"OptimizeNKIKernels": 1.0786433219909668,
"PAGLayoutOpt": 18.570398330688477,
"PComputeCutting": 0.6222004890441895,
"PGLayoutTilingPipeline": 55.62285614013672,
"PGTiling": 10.0595703125,
"PadElimination": 0.021188735961914063,
"ParAxesAnnotation": 17.818796157836914,
"PartialLoopFusion": 1.5088038444519043,
"PartialSimdFusion": 0.8573262691497803,
"PerfectLoopNest": 0.07143402099609375,
"RecognizeOpIdiom": 0.1291654109954834,
"Recompute": 0.007833480834960938,
"RelaxPredicates": 0.14508056640625,
"Rematerialization": 0.17157793045043945,
"RemoveShardedPartitionAxes": 1.3566563129425049,
"ReshapeWeights": 0.022662878036499023,
"ResolveAccessConflict": 0.2033238410949707,
"ResolveComplicatePredicates": 0.49439358711242676,
"RewriteReplicationMatmul": 0.03938794136047363,
"RewriteWeights": 0.07809877395629883,
"SFKVectorizer": 10.313403129577637,
"ShardingPropagationAnalysis": 0.7199399471282959,
"SimpleAllReduceTiling": 0.06984925270080566,
"Simplifier": 0.09367704391479492,
"SimplifyMacroPredicates": 0.28262877464294434,
"SimplifyNeuronTensor": 0.9048190116882324,
"SimplifySlice": 0.027447223663330078,
"SimplifyTensor": 0.2734675407409668,
"SpillPSum": 1.2256271839141846,
"SplitAPUnionSets": 0.8627588748931885,
"SplitAccGrp": 0.1022341251373291,
"StaticProfiler": 0.4753732681274414,
"StaticTransposeLocalTensor": 0.6612956523895264,
"SundaISel": 1.6735970973968506,
"TCTransform": 0.029105186462402344,
"TensorInitialization": 0.1944262981414795,
"TensorOpSimplifier": 3.3376035690307617,
"TensorOpTransform": 19.926769256591797,
"TileCCOps": 0.1856687068939209,
"TilingProfiler": 0.5146362781524658,
"TransformConvOp": 1.4806337356567383,
"TritiumFusion": 0.2828052043914795,
"ValueNumbering": 0.08587479591369629,
"VectorizeDMA": 0.5646088123321533,
"VectorizeMatMult": 0.08723974227905273,
"WeightCoalescing": 0.06122708320617676,
"ZeroSizeTensorElimination": 0.0008580684661865234
},
"tensorizer": {
"DMATilingProfiler::TotalInstructionsAfterTiling": 51674,
"StaticProfiler::AifUb": 21.735599517822266,
"StaticProfiler::ArithmeticIntensityTensorizer": 31.884082794189453,
"StaticProfiler::AverageDmaLength": 2843.9130859375,
"StaticProfiler::AverageFractalPeUtilization": 97.88961791992188,
"StaticProfiler::AveragePartitionUtilization": 89.79463958740234,
"StaticProfiler::AveragePeUtilization": 78.904052734375,
"StaticProfiler::DDRTransferBytes": 1013018224,
"StaticProfiler::InternalTransferBytes": 226623072,
"StaticProfiler::LoadExpanded": 281500,
"StaticProfiler::LocalizationEfficiency": 146.6905975341797,
"StaticProfiler::LocalizationEfficiencyIgnoreNonlocal": 151.06674194335938,
"StaticProfiler::StoreExpanded": 4273,
"StaticProfiler::TotalDMAExpanded": 285773,
"StaticProfiler::TotalDynamicInstancesCount": 67989,
"StaticProfiler::TotalDynamicInstancesWithMmPackedCount": 57540,
"StaticProfiler::TotalLNCComm": 0,
"StaticProfiler::TotalLNCCommTransfer": 0,
"TilingProfiler::AveragePartitionUtilizationAfterTiling": 0,
"TilingProfiler::AveragePeUtilizationAfterTiling": 0,
"TilingProfiler::BatchnormInstructionsAfterTiling": 0,
"TilingProfiler::DmaInstructionsAfterTiling": 0,
"TilingProfiler::GenericInstructionsAfterTiling": 123,
"TilingProfiler::MatMultInstructionsAfterTiling": 35008,
"TilingProfiler::NumPfTransposes": 292,
"TilingProfiler::NumPfTransposesForIo": 30,
"TilingProfiler::NumPfTransposesForLocal": 142,
"TilingProfiler::NumPfTransposesForNonlocal": 120,
"TilingProfiler::PfTransposeInstructions": 10670,
"TilingProfiler::PfTransposeInstructionsForIo": 8360,
"TilingProfiler::PfTransposeInstructionsForLocal": 648,
"TilingProfiler::PfTransposeInstructionsForNonlocal": 1662,
"TilingProfiler::ReduceInstructionsAfterTiling": 283,
"TilingProfiler::SimdInstructionsAfterTiling": 2805,
"TilingProfiler::TotalInstructionsAfterTiling": 0,
"TransformConvOp::Conv1d_depthwise_bf01_oi01_bf01": 0,
"TransformConvOp::Conv2d_dw_fb01_io01_01bf_rep_nhwc_Pcinh": 0,
"TransformConvOp::Conv2d_pbp_0f1b_0i1o_01fb_experimental_1": 0,
"TransformConvOp::Conv2d_pbp_fb01_io01_01bf_experimental_1": 0,
"TransformConvOp::conv2d_column_packing": 0,
"TransformConvOp::conv2d_column_packing_1": 0,
"TransformConvOp::conv2d_column_packing_io10": 0,
"TransformConvOp::conv2d_depthwise_f01b_o01i_bf01": 0
}
},
"topk": {
"compiletime": {
"CoalesceCCOp": 0.003665924072265625,
"DMALocalityOpt": 0.002828359603881836,
"DMAProfiler": 0.0035741329193115234,
"DataStreaming": 0.006661653518676758,
"DoNothing": 0.00015878677368164063,
"ExpandISAMacro": 0.0038259029388427734,
"FactorizeBlkDims": 0.012358903884887695,
"InferPSumTensor": 0.01241612434387207,
"InferSharedMemLoc": 0.0028722286224365234,
"InsertCoreBarrier": 0.0033416748046875,
"LateLegalizeInst": 0.0075190067291259766,
"LateNeuronInstComb": 0.008240699768066406,
"LegalizeSundaAccess": 0.014475584030151367,
"LegalizeType": 0.00968170166015625,
"LowerBroadcast": 0.0032701492309570313,
"LowerIntrinsics": 0.0037560462951660156,
"LowerTranspose": 0.0033812522888183594,
"NeuronInstComb": 0.008539199829101563,
"NeuronLICM": 0.00936579704284668,
"NeuronSimplifyPredicates": 0.0037877559661865234,
"NeuronValueNumbering": 0.0040531158447265625,
"SFKVectorizer": 0.03392672538757324,
"SimpleAllReduceTiling": 0.003645181655883789,
"SimplifyNeuronTensor": 0.05141305923461914,
"SpillPSum": 0.02246880531311035,
"WeightCoalescing": 0.0034682750701904297
}
}
}

3
token_generation_model/_tp0_bk3/graph.neff Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0c81d96024b571f29492feb3e8f54a86bb85699475d770c8027b26b3b25d6d30
size 3605504

4471
token_generation_model/_tp0_bk3/log-neuron-cc.txt Normal file
View File

File diff suppressed because it is too large Load Diff

3
token_generation_model/_tp0_bk3/metaneff.pb Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:6f9340d294060545b7fd3affe72e9160857b035623ad5e43d5186704d48fc4cd
size 2464555

3
token_generation_model/_tp0_bk3/model.MODULE_02e08d5fdadfbcb8a569+a6f96e4f.hlo_module.pb Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:25287d0dc180171fa1e47e9880d832a32168762120f79d5891ed820ee9b4fcda
size 2550451

3
token_generation_model/_tp0_bk3/model.MODULE_02e08d5fdadfbcb8a569+a6f96e4f.neff Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:0c81d96024b571f29492feb3e8f54a86bb85699475d770c8027b26b3b25d6d30
size 3605504

224
token_generation_model/_tp0_bk3/neuron_config.json Normal file
View File

@@ -0,0 +1,224 @@
{
"_attn_implementation_autoset": false,
"_name_or_path": "/models/Qwen3-1.7B/",
"add_cross_attention": false,
"architectures": [
"Qwen3ForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"attribute_map": {},
"bad_words_ids": null,
"begin_suppress_tokens": null,
"bos_token_id": 151643,
"chunk_size_feed_forward": 0,
"cross_attention_hidden_size": null,
"decoder_start_token_id": null,
"diversity_penalty": 0.0,
"do_sample": false,
"early_stopping": false,
"encoder_no_repeat_ngram_size": 0,
"eos_token_id": 151645,
"exponential_decay_length_penalty": null,
"finetuning_task": null,
"forced_bos_token_id": null,
"forced_eos_token_id": null,
"fused_spec_config": null,
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 2048,
"id2label": {
"0": "LABEL_0",
"1": "LABEL_1"
},
"initializer_range": 0.02,
"intermediate_size": 6144,
"is_decoder": false,
"is_encoder_decoder": false,
"label2id": {
"LABEL_0": 0,
"LABEL_1": 1
},
"length_penalty": 1.0,
"max_length": 20,
"max_position_embeddings": 40960,
"max_window_layers": 28,
"metadata": null,
"min_length": 0,
"model_type": "qwen3",
"neuron_config": {
"activation_quantization_type": null,
"allow_input_truncation": false,
"apply_seq_ids_mask": false,
"async_mode": false,
"attention_dp_degree": 1,
"attention_dtype": null,
"attn_block_cte_nki_kernel_enabled": false,
"attn_block_tkg_nki_kernel_cache_update": false,
"attn_block_tkg_nki_kernel_cascaded_attention": false,
"attn_block_tkg_nki_kernel_enabled": false,
"attn_cls": {
"__module__": "neuronx_distributed_inference.models.qwen3.modeling_qwen3",
"__name__": "NeuronQwen3Attention"
},
"attn_kernel_enabled": null,
"attn_tkg_builtin_kernel_enabled": false,
"attn_tkg_nki_kernel_enabled": false,
"batch_size": 4,
"bucket_n_active_tokens": false,
"buckets": [
1024
],
"cast_type": "config",
"cc_pipeline_tiling_factor": 1,
"chunked_prefill_config": null,
"context_encoding_buckets": null,
"cp_degree": 1,
"ctx_batch_size": 1,
"disable_kv_cache_tiling": false,
"draft_model_modules_to_not_convert": null,
"enable_bucketing": true,
"enable_cte_modular_flow": false,
"enable_eagle_draft_input_norm": false,
"enable_eagle_speculation": false,
"enable_fused_speculation": false,
"enable_long_context_mode": false,
"enable_output_completion_notifications": false,
"enable_spill_reload_dge": false,
"enable_token_tree": false,
"ep_degree": 1,
"expert_mlp_nki_kernel_enabled": null,
"flash_decoding_enabled": false,
"fused_qkv": false,
"fused_rmsnorm_skip_gamma": false,
"is_block_kv_layout": null,
"is_chunked_prefill": false,
"is_continuous_batching": true,
"is_eagle_draft": false,
"is_medusa": false,
"is_prefill_stage": false,
"is_prefix_caching": false,
"k_cache_transposed": false,
"kv_cache_batch_size": 4,
"kv_cache_padding_size": 0,
"kv_cache_quant": false,
"kv_cache_tiling": false,
"layer_boundary_markers": false,
"lm_head_pad": true,
"lm_head_pad_alignment_size": 1,
"local_ranks_size": 4,
"logical_nc_config": 2,
"lora_config": null,
"max_batch_size": 4,
"max_context_length": 2048,
"max_length": 2048,
"max_new_tokens": null,
"medusa_speculation_length": 0,
"medusa_tree": null,
"mlp_kernel_enabled": false,
"mlp_kernel_fuse_residual_add": false,
"modules_to_not_convert": null,
"moe_fused_nki_kernel_enabled": null,
"n_active_tokens": 1,
"n_positions": 2048,
"num_medusa_heads": 0,
"on_cpu": false,
"on_device_sampling_config": {
"deterministic": false,
"do_sample": false,
"dynamic": true,
"global_topk": 256,
"on_device_sampling_config": true,
"temperature": 1.0,
"top_k": 1,
"top_k_kernel_enabled": false,
"top_p": 1.0
},
"output_logits": false,
"overrides_torch_dtype": true,
"pa_block_size": 2048,
"pa_num_blocks": 4,
"padding_side": "right",
"pp_degree": 1,
"prefix_buckets": null,
"qk_layernorm": false,
"qkv_kernel_enabled": false,
"qkv_kernel_fuse_residual_add": false,
"qkv_kernel_nbsd_layout": false,
"quantization_dtype": "int8",
"quantization_type": "per_tensor_symmetric",
"quantize_clamp_bound": Infinity,
"quantized": false,
"quantized_checkpoints_path": null,
"quantized_mlp_kernel_enabled": false,
"rmsnorm_quantize_kernel_enabled": false,
"router_topk_nki_kernel_enabled": null,
"rpl_reduce_dtype": null,
"save_sharded_checkpoint": true,
"scratchpad_page_size": null,
"seq_len": 2048,
"seq_len_threshold_for_cc_tiling": 16384,
"sequence_parallel_enabled": false,
"shared_mlp_nki_kernel_enabled": null,
"skip_sharding": false,
"skip_warmup": false,
"spec_batch_size": 4,
"speculation_length": 0,
"start_rank_id": 0,
"strided_context_parallel_kernel_enabled": false,
"target": null,
"tensor_capture_config": null,
"tile_cc": false,
"tkg_batch_size": 4,
"token_generation_buckets": [
1024
],
"token_tree_config": null,
"torch_dtype": "bfloat16",
"tp_degree": 4,
"vocab_parallel": false,
"weight_gather_seq_len_threshold": 32768,
"weights_to_skip_layout_optimization": [],
"world_size": 4
},
"no_repeat_ngram_size": 0,
"num_attention_heads": 16,
"num_beam_groups": 1,
"num_beams": 1,
"num_cores_per_group": 1,
"num_hidden_layers": 28,
"num_key_value_heads": 8,
"num_return_sequences": 1,
"output_attentions": false,
"output_hidden_states": false,
"output_scores": false,
"pad_token_id": 0,
"prefix": null,
"problem_type": null,
"pruned_heads": {},
"remove_invalid_values": false,
"repetition_penalty": 1.0,
"return_dict": true,
"return_dict_in_generate": false,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"rope_theta": 1000000,
"sep_token_id": null,
"sliding_window": null,
"suppress_tokens": null,
"task_specific_params": null,
"temperature": 1.0,
"tf_legacy_loss": false,
"tie_encoder_decoder": false,
"tie_word_embeddings": true,
"tokenizer_class": null,
"top_k": 50,
"top_p": 1.0,
"torchscript": false,
"transformers_version": "4.51.0",
"typical_p": 1.0,
"use_bfloat16": false,
"use_cache": true,
"use_sliding_window": false,
"vocab_size": 151936
}