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_bk1/command.txt Normal file
View File

@@ -0,0 +1 @@
neuronx-cc compile --framework=XLA model.MODULE_f53407701fa4882a24c0+55c11e15.hlo_module.pb --output model.MODULE_f53407701fa4882a24c0+55c11e15.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_bk1/compile_flags.MODULE_f53407701fa4882a24c0+55c11e15.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_bk1/log-neuron-cc.txt"]

590
token_generation_model/_tp0_bk1/global_metric_store.json Normal file
View File

@@ -0,0 +1,590 @@
{
"Average": {
"tensorizer": {
"StaticProfiler::AverageFractalPeUtilization": 97.62759399414063,
"StaticProfiler::AveragePartitionUtilization": 90.06017303466797,
"StaticProfiler::AveragePeUtilization": 82.14405822753906,
"StaticProfiler::LocalizationEfficiency": 160.6605224609375,
"StaticProfiler::LocalizationEfficiencyIgnoreNonlocal": 165.92486572265625,
"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": 5.160665512084961,
"AffinePredicateResolution": 0.7100803852081299,
"AliasDependencyElimination": 0.0032160282135009766,
"AliasDependencyInduction": 17.29242706298828,
"AliasDependencyReset": 18.28659439086914,
"BFComputeCutting": 0.24476861953735352,
"BirCodeGenLoop": 6.947810173034668,
"CCOpFusion": 0.6929187774658203,
"CanonicalizeConv": 2.9000000722589903e-05,
"CanonicalizeDAGForPGTiling": 0.21501421928405762,
"CanonicalizeForTensorizer": 0.0008399999933317304,
"CanonicalizeIR": 0.8826379776000977,
"Canonicalizer": 0.014088000170886517,
"CoalesceCCOp": 0.5657715797424316,
"CommuteConcat": 0.028485536575317383,
"DMALocalityOpt": 0.042168617248535156,
"DMAProfiler": 0.1519615650177002,
"DMATilingProfiler": 0.09857678413391113,
"DataLocalityOpt": 2.827857255935669,
"DataStreaming": 0.20627403259277344,
"DeConcat": 0.08719110488891602,
"DeadCodeElimination": 0.03139615058898926,
"DeadStoreElimination": 1.1847929954528809,
"DelinearIndices": 0.4130244255065918,
"Delinearization": 0.35245656967163086,
"DelinearizeSPMD": 0.4181056022644043,
"DoNothing": 0.00046539306640625,
"DramToDramTranspose": 0.30753135681152344,
"DumpGraphAndMetadata": 0.2011098861694336,
"EliminateDivs": 1.3804805278778076,
"ExpandBatchNorm": 1.1923854351043701,
"ExpandISAMacro": 0.08926725387573242,
"FactorizeBlkDims": 0.5252444744110107,
"FactorizeThreadAxesInFreeDims": 0.1739037036895752,
"FlattenMacroLoop": 0.07839345932006836,
"GenericAccessSimplifier": 0.025495052337646484,
"HoistCompute": 6.199999916134402e-05,
"IdentifyCrossPassTensors": 0.0005760000203736126,
"InferInitValue": 1.3857665061950684,
"InferIntrinsicOnCC": 0.646845817565918,
"InferNeuronTensor": 2.0493955612182617,
"InferNonlocalTensors": 6.630102634429932,
"InferPSumTensor": 1.2878854274749756,
"InferShardAxis": 10.588101387023926,
"InferSharedMemLoc": 0.10591363906860352,
"InlineNativeKernels": 0.0488896369934082,
"InsertCoreBarrier": 0.3818776607513428,
"InsertIOTransposes": 0.8546113967895508,
"InsertImplicitShardAxisBeforeISel": 0.36914730072021484,
"InsertLocalTransposes": 1.5642907619476318,
"InsertOffloadedTransposes": 0.12299966812133789,
"LICM": 0.12427496910095215,
"LateLegalizeInst": 0.4006483554840088,
"LateLegalizePostSplit": 0.09814929962158203,
"LateLowerReshapeOp": 0.0431976318359375,
"LateLowerTensorOp": 5.883858680725098,
"LateNeuronInstComb": 1.04233980178833,
"LayoutPreprocessing": 1.2663531303405762,
"LayoutPreprocessingAndAnalysis": 1.9893858432769775,
"LayoutRequirementAnalysis": 0.6988728046417236,
"LegalizeCCOpLayout": 1.499981164932251,
"LegalizeOpLevelAlias": 0.6842360496520996,
"LegalizePartitionReduce": 0.09090352058410645,
"LegalizeSundaAccess": 3.133746862411499,
"LegalizeSundaMacro": 0.6402661800384521,
"LegalizeType": 0.1407451629638672,
"LocalLayoutOpt": 0.7939743995666504,
"LoopFusion": 0.29726386070251465,
"LoopSplitting": 0.08333015441894531,
"LowerBroadcast": 0.06013798713684082,
"LowerCCOpBlockAxis": 1.4568629264831543,
"LowerComplexBroadcast": 0.06818914413452148,
"LowerIntrinsics": 0.9435675144195557,
"LowerShardAxis": 0.22619009017944336,
"LowerTensorOp": 3.0802407264709473,
"LowerToSendRecv": 0.19478487968444824,
"LowerTranspose": 0.5133018493652344,
"MacroGeneration": 4.587047100067139,
"MaskPropagation": 0.1672959327697754,
"MemcastMotion": 0.00015999999595806003,
"MemcpyElimination": 29.92277717590332,
"MutateDataType": 0.04070854187011719,
"NeuronAliasDependencyInduction": 0.018942594528198242,
"NeuronAliasDependencyReset": 0.02425408363342285,
"NeuronInstComb": 0.4308168888092041,
"NeuronLICM": 0.29584765434265137,
"NeuronLoopFusion": 1.117248296737671,
"NeuronLoopInterchange": 0.0636141300201416,
"NeuronSimplifier": 0.704033374786377,
"NeuronSimplifyPredicates": 0.2319803237915039,
"NeuronValueNumbering": 0.11443066596984863,
"OptimizeAliasedCopyChain": 0.2889397144317627,
"OptimizeNKIKernels": 1.2374975681304932,
"PAGLayoutOpt": 18.6531982421875,
"PComputeCutting": 0.5685915946960449,
"PGLayoutTilingPipeline": 53.59619903564453,
"PGTiling": 11.116703033447266,
"PadElimination": 0.012248039245605469,
"ParAxesAnnotation": 17.06294822692871,
"PartialLoopFusion": 1.3277881145477295,
"PartialSimdFusion": 0.7600483894348145,
"PenguinizeFunctions": 0.001062000053934753,
"PerfectLoopNest": 0.061771392822265625,
"PruneFunctions": 0.0004830000107176602,
"RecognizeOpIdiom": 0.12406373023986816,
"Recompute": 0.00886678695678711,
"RelaxPredicates": 0.11265873908996582,
"Rematerialization": 0.18548583984375,
"RemoveOptimizationBarriers": 0.009134000167250633,
"RemoveShardedPartitionAxes": 1.4166390895843506,
"ReshapeWeights": 0.02198624610900879,
"ResolveAccessConflict": 0.20142865180969238,
"ResolveComplicatePredicates": 0.6187732219696045,
"RewriteReplicationMatmul": 0.04161477088928223,
"RewriteWeights": 0.0676727294921875,
"SFKVectorizer": 11.220393180847168,
"ScatterMotion": 0.003527000080794096,
"ShardingPropagationAnalysis": 0.8252537250518799,
"SimpleAllReduceTiling": 0.1859297752380371,
"Simplifier": 0.09509730339050293,
"SimplifyMacroPredicates": 0.26697540283203125,
"SimplifyNeuronTensor": 0.41252875328063965,
"SimplifySlice": 0.02629375457763672,
"SimplifyTensor": 0.42005443572998047,
"SpillPSum": 0.5884366035461426,
"SplitAPUnionSets": 0.4571385383605957,
"SplitAccGrp": 0.052629709243774414,
"StaticProfiler": 0.12954020500183105,
"StaticTransposeLocalTensor": 0.36740827560424805,
"SundaISel": 1.4737660884857178,
"TCTransform": 0.031054019927978516,
"TensorInitialization": 0.18288874626159668,
"TensorOpSimplifier": 3.2880165576934814,
"TensorOpTransform": 18.079126358032227,
"TensorizerLegalizationPass": 0.0004990000161342323,
"TileCCOps": 0.17596793174743652,
"TilingProfiler": 0.3990769386291504,
"TransformConvOp": 1.2077322006225586,
"TritiumFusion": 0.2976958751678467,
"ValueNumbering": 0.10138130187988281,
"VectorizeDMA": 0.8864037990570068,
"VectorizeMatMult": 0.05278921127319336,
"VerifySupportedOps": 0.0004729999927803874,
"WeightCoalescing": 0.08338785171508789,
"ZeroSizeTensorElimination": 0.0006525516510009766,
"algsimp": 0.0014349999837577343,
"batchnorm_expander": 0.000818000000435859,
"boundary-marker-removal": 0.0003020000003743917,
"call-inliner": 0.00018099999579135329,
"canonicalize-boundary-marker": 0.0005639999872073531,
"collective-stream-id-checker": 0.002294000005349517,
"comparison-expander": 0.0012120000319555402,
"computation-deduplicator": 0.0011109999613836408,
"config-lowering": 0.00042299999040551484,
"constant_folding": 0.0001289999927394092,
"cse": 0.0007570000016130507,
"dce": 7.300000288523734e-05,
"dynamic-slice-transpose": 0.0002640000020619482,
"eliminate-redundant-compare": 0.0001250000059371814,
"emit-offloaded-dropout": 0.0006559999892488122,
"flatten-call-graph": 0.0003330000035930425,
"fuse-send-recv": 0.013015000149607658,
"hilo-conditional-to-select": 0.00016799999866634607,
"hilo::LegalizeAlias": 0.0036430000327527523,
"hilo::NeuronInstCombine": 1.1000000085914508e-05,
"hilo::NeuronOpFusion": 0.00011600000289035961,
"hilo::ReplaceTokenTypeWithU8Pass": 0.0003260000084992498,
"hilo::ScheduleFusion": 3.300000025774352e-05,
"hilo::SixtyFourHack": 0.0013909999979659915,
"hilo::VerifyAliasing": 0.00014600000577047467,
"hlo-mac-count": 0.019317999482154846,
"io-con-pipe-begin": 0.0002589999930933118,
"io-con-pipe-end": 9.999999974752427e-07,
"io-layout-normalization": 0.07799100130796432,
"legalize-ccops-for-tensorizer": 5.2999999752501026e-05,
"legalize-compare": 0.0006939999875612557,
"lower-argminmax-custom-call": 0.0002629999944474548,
"map-inline": 0.06807900220155716,
"metadata-naming": 0.07221200317144394,
"mlir::detail::OpToOpPassAdaptor": 0.007284999825060368,
"mlir::hlo::MhloToPyPenguin": 0.6235949993133545,
"mlir::mhlo::LowerComplexExtraPass": 0.00916799996048212,
"mlir::mhlo::LowerComplexPass": 0.0006799999973736703,
"native-to-custom-softmax": 0.0017719999887049198,
"native-to-custom-softmax-dx": 0.0017500000540167093,
"neuron-hlo-verifier": 0.2390509992837906,
"operand_upcaster": 0.070872001349926,
"post-par-pipe-begin": 9.999999974752427e-07,
"post-par-pipe-end": 9.999999747378752e-06,
"post-partition-simplification": 0.21146699786186218,
"pre-hlo-begin": 8.600000001024455e-05,
"pre-hlo-end": 9.999999974752427e-07,
"replace-minimum-constant": 0.00019700000120792538,
"reshape-mover": 5.400000009103678e-05,
"simplify-concat": 0.002297000028192997,
"simplify-while-loops": 4.600000102072954e-05,
"transform-variadic-reduce": 0.00037900000461377203,
"tuple-simplifier": 0.0001340000017080456,
"unpack-nested-aws-ntwsr": 0.00025400001322850585,
"unroll-while-loop": 7.000000096013537e-06
},
"hilo": {
"HloMacCount": 1766645760.0,
"Traffic": 1252618368.0
},
"tensorizer": {
"DMATilingProfiler::TotalInstructionsAfterTiling": 42842,
"StaticProfiler::AifUb": 17.765029907226563,
"StaticProfiler::ArithmeticIntensityTensorizer": 28.541391372680664,
"StaticProfiler::AverageDmaLength": 3512.794677734375,
"StaticProfiler::DDRTransferBytes": 924925552,
"StaticProfiler::InternalTransferBytes": 182108768,
"StaticProfiler::LoadExpanded": 181148,
"StaticProfiler::StoreExpanded": 4273,
"StaticProfiler::TotalDMAExpanded": 185421,
"StaticProfiler::TotalDynamicInstancesCount": 57693,
"StaticProfiler::TotalDynamicInstancesWithMmPackedCount": 50217,
"StaticProfiler::TotalLNCComm": 0,
"StaticProfiler::TotalLNCCommTransfer": 0,
"TilingProfiler::BatchnormInstructionsAfterTiling": 0,
"TilingProfiler::DmaInstructionsAfterTiling": 0,
"TilingProfiler::GenericInstructionsAfterTiling": 123,
"TilingProfiler::MatMultInstructionsAfterTiling": 30976,
"TilingProfiler::NumPfTransposes": 348,
"TilingProfiler::NumPfTransposesForIo": 30,
"TilingProfiler::NumPfTransposesForLocal": 198,
"TilingProfiler::NumPfTransposesForNonlocal": 120,
"TilingProfiler::PfTransposeInstructions": 7892,
"TilingProfiler::PfTransposeInstructionsForIo": 5666,
"TilingProfiler::PfTransposeInstructionsForLocal": 564,
"TilingProfiler::PfTransposeInstructionsForNonlocal": 1662,
"TilingProfiler::ReduceInstructionsAfterTiling": 115,
"TilingProfiler::SimdInstructionsAfterTiling": 2295,
"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": 2.9000000722589903e-05,
"CanonicalizeForTensorizer": 0.0008399999933317304,
"Canonicalizer": 0.014088000170886517,
"HoistCompute": 6.199999916134402e-05,
"IdentifyCrossPassTensors": 0.0005760000203736126,
"MemcastMotion": 0.00015999999595806003,
"PenguinizeFunctions": 0.001062000053934753,
"PruneFunctions": 0.0004830000107176602,
"RemoveOptimizationBarriers": 0.009134000167250633,
"ScatterMotion": 0.003527000080794096,
"TensorizerLegalizationPass": 0.0004990000161342323,
"VerifySupportedOps": 0.0004729999927803874,
"algsimp": 0.0014349999837577343,
"batchnorm_expander": 0.000818000000435859,
"boundary-marker-removal": 0.0003020000003743917,
"call-inliner": 0.00018099999579135329,
"canonicalize-boundary-marker": 0.0005639999872073531,
"collective-stream-id-checker": 0.002294000005349517,
"comparison-expander": 0.0012120000319555402,
"computation-deduplicator": 0.0011109999613836408,
"config-lowering": 0.00042299999040551484,
"constant_folding": 0.0001289999927394092,
"cse": 0.0007570000016130507,
"dce": 7.300000288523734e-05,
"dynamic-slice-transpose": 0.0002640000020619482,
"eliminate-redundant-compare": 0.0001250000059371814,
"emit-offloaded-dropout": 0.0006559999892488122,
"flatten-call-graph": 0.0003330000035930425,
"fuse-send-recv": 0.013015000149607658,
"hilo-conditional-to-select": 0.00016799999866634607,
"hilo::LegalizeAlias": 0.0036430000327527523,
"hilo::NeuronInstCombine": 1.1000000085914508e-05,
"hilo::NeuronOpFusion": 0.00011600000289035961,
"hilo::ReplaceTokenTypeWithU8Pass": 0.0003260000084992498,
"hilo::ScheduleFusion": 3.300000025774352e-05,
"hilo::SixtyFourHack": 0.0013909999979659915,
"hilo::VerifyAliasing": 0.00014600000577047467,
"hlo-mac-count": 0.019317999482154846,
"io-con-pipe-begin": 0.0002589999930933118,
"io-con-pipe-end": 9.999999974752427e-07,
"io-layout-normalization": 0.07799100130796432,
"legalize-ccops-for-tensorizer": 5.2999999752501026e-05,
"legalize-compare": 0.0006939999875612557,
"lower-argminmax-custom-call": 0.0002629999944474548,
"map-inline": 0.06807900220155716,
"metadata-naming": 0.07221200317144394,
"mlir::detail::OpToOpPassAdaptor": 0.007284999825060368,
"mlir::hlo::MhloToPyPenguin": 0.6235949993133545,
"mlir::mhlo::LowerComplexExtraPass": 0.00916799996048212,
"mlir::mhlo::LowerComplexPass": 0.0006799999973736703,
"native-to-custom-softmax": 0.0017719999887049198,
"native-to-custom-softmax-dx": 0.0017500000540167093,
"neuron-hlo-verifier": 0.2390509992837906,
"operand_upcaster": 0.070872001349926,
"post-par-pipe-begin": 9.999999974752427e-07,
"post-par-pipe-end": 9.999999747378752e-06,
"post-partition-simplification": 0.21146699786186218,
"pre-hlo-begin": 8.600000001024455e-05,
"pre-hlo-end": 9.999999974752427e-07,
"replace-minimum-constant": 0.00019700000120792538,
"reshape-mover": 5.400000009103678e-05,
"simplify-concat": 0.002297000028192997,
"simplify-while-loops": 4.600000102072954e-05,
"transform-variadic-reduce": 0.00037900000461377203,
"tuple-simplifier": 0.0001340000017080456,
"unpack-nested-aws-ntwsr": 0.00025400001322850585,
"unroll-while-loop": 7.000000096013537e-06
}
},
"cumsum": {
"compiletime": {
"CoalesceCCOp": 0.0002148151397705078,
"DMALocalityOpt": 0.0001659393310546875,
"DMAProfiler": 0.0008566379547119141,
"DataStreaming": 0.00026869773864746094,
"DoNothing": 0.00015687942504882813,
"ExpandISAMacro": 0.0004825592041015625,
"FactorizeBlkDims": 0.0004360675811767578,
"InferPSumTensor": 0.0005650520324707031,
"InferSharedMemLoc": 0.0002880096435546875,
"InsertCoreBarrier": 0.0002799034118652344,
"LateLegalizeInst": 0.0003802776336669922,
"LateNeuronInstComb": 0.0006039142608642578,
"LegalizeSundaAccess": 0.001421213150024414,
"LegalizeType": 0.0002644062042236328,
"LowerBroadcast": 0.0002384185791015625,
"LowerIntrinsics": 0.0002446174621582031,
"LowerTranspose": 0.00022077560424804688,
"NeuronInstComb": 0.0005753040313720703,
"NeuronLICM": 0.00037217140197753906,
"NeuronSimplifyPredicates": 0.0022978782653808594,
"NeuronValueNumbering": 0.0004074573516845703,
"SFKVectorizer": 0.002466440200805664,
"SimpleAllReduceTiling": 0.0002262592315673828,
"SimplifyNeuronTensor": 0.0005323886871337891,
"SpillPSum": 0.00045490264892578125,
"WeightCoalescing": 0.0002465248107910156
}
},
"sg00": {
"hilo": {
"ArithmeticIntensity": 2.8207247257232666,
"HloMacCount": 1766645760.0,
"Traffic": 1252618368.0
}
},
"sg0000": {
"compiletime": {
"AGOrderingAnalysisPass": 5.160665512084961,
"AffinePredicateResolution": 0.7100803852081299,
"AliasDependencyElimination": 0.0032160282135009766,
"AliasDependencyInduction": 17.29242706298828,
"AliasDependencyReset": 18.28659439086914,
"BFComputeCutting": 0.24476861953735352,
"BirCodeGenLoop": 6.947810173034668,
"CCOpFusion": 0.6929187774658203,
"CanonicalizeDAGForPGTiling": 0.21501421928405762,
"CanonicalizeIR": 0.8826379776000977,
"CoalesceCCOp": 0.5618836879730225,
"CommuteConcat": 0.028485536575317383,
"DMALocalityOpt": 0.03933858871459961,
"DMAProfiler": 0.14742827415466309,
"DMATilingProfiler": 0.09857678413391113,
"DataLocalityOpt": 2.827857255935669,
"DataStreaming": 0.19969892501831055,
"DeConcat": 0.08719110488891602,
"DeadCodeElimination": 0.03139615058898926,
"DeadStoreElimination": 1.1847929954528809,
"DelinearIndices": 0.4130244255065918,
"Delinearization": 0.35245656967163086,
"DelinearizeSPMD": 0.4181056022644043,
"DoNothing": 0.0001385211944580078,
"DramToDramTranspose": 0.30753135681152344,
"DumpGraphAndMetadata": 0.2011098861694336,
"EliminateDivs": 1.3804805278778076,
"ExpandBatchNorm": 1.1923854351043701,
"ExpandISAMacro": 0.08506417274475098,
"FactorizeBlkDims": 0.5093107223510742,
"FactorizeThreadAxesInFreeDims": 0.1739037036895752,
"FlattenMacroLoop": 0.07839345932006836,
"GenericAccessSimplifier": 0.025495052337646484,
"InferInitValue": 1.3857665061950684,
"InferIntrinsicOnCC": 0.646845817565918,
"InferNeuronTensor": 2.0493955612182617,
"InferNonlocalTensors": 6.630102634429932,
"InferPSumTensor": 1.275925636291504,
"InferShardAxis": 10.588101387023926,
"InferSharedMemLoc": 0.102783203125,
"InlineNativeKernels": 0.0488896369934082,
"InsertCoreBarrier": 0.3783996105194092,
"InsertIOTransposes": 0.8546113967895508,
"InsertImplicitShardAxisBeforeISel": 0.36914730072021484,
"InsertLocalTransposes": 1.5642907619476318,
"InsertOffloadedTransposes": 0.12299966812133789,
"LICM": 0.12427496910095215,
"LateLegalizeInst": 0.3925168514251709,
"LateLegalizePostSplit": 0.09814929962158203,
"LateLowerReshapeOp": 0.0431976318359375,
"LateLowerTensorOp": 5.883858680725098,
"LateNeuronInstComb": 1.0332238674163818,
"LayoutPreprocessing": 1.2663531303405762,
"LayoutPreprocessingAndAnalysis": 1.9893858432769775,
"LayoutRequirementAnalysis": 0.6988728046417236,
"LegalizeCCOpLayout": 1.499981164932251,
"LegalizeOpLevelAlias": 0.6842360496520996,
"LegalizePartitionReduce": 0.09090352058410645,
"LegalizeSundaAccess": 3.1170387268066406,
"LegalizeSundaMacro": 0.6402661800384521,
"LegalizeType": 0.1309196949005127,
"LocalLayoutOpt": 0.7939743995666504,
"LoopFusion": 0.29726386070251465,
"LoopSplitting": 0.08333015441894531,
"LowerBroadcast": 0.05664563179016113,
"LowerCCOpBlockAxis": 1.4568629264831543,
"LowerComplexBroadcast": 0.06818914413452148,
"LowerIntrinsics": 0.9396772384643555,
"LowerShardAxis": 0.22619009017944336,
"LowerTensorOp": 3.0802407264709473,
"LowerToSendRecv": 0.19478487968444824,
"LowerTranspose": 0.5097734928131104,
"MacroGeneration": 4.587047100067139,
"MaskPropagation": 0.1672959327697754,
"MemcpyElimination": 29.92277717590332,
"MutateDataType": 0.04070854187011719,
"NeuronAliasDependencyInduction": 0.018942594528198242,
"NeuronAliasDependencyReset": 0.02425408363342285,
"NeuronInstComb": 0.42151784896850586,
"NeuronLICM": 0.2860753536224365,
"NeuronLoopFusion": 1.117248296737671,
"NeuronLoopInterchange": 0.0636141300201416,
"NeuronSimplifier": 0.704033374786377,
"NeuronSimplifyPredicates": 0.22606325149536133,
"NeuronValueNumbering": 0.11009359359741211,
"OptimizeAliasedCopyChain": 0.2889397144317627,
"OptimizeNKIKernels": 1.2374975681304932,
"PAGLayoutOpt": 18.6531982421875,
"PComputeCutting": 0.5685915946960449,
"PGLayoutTilingPipeline": 53.59619903564453,
"PGTiling": 11.116703033447266,
"PadElimination": 0.012248039245605469,
"ParAxesAnnotation": 17.06294822692871,
"PartialLoopFusion": 1.3277881145477295,
"PartialSimdFusion": 0.7600483894348145,
"PerfectLoopNest": 0.061771392822265625,
"RecognizeOpIdiom": 0.12406373023986816,
"Recompute": 0.00886678695678711,
"RelaxPredicates": 0.11265873908996582,
"Rematerialization": 0.18548583984375,
"RemoveShardedPartitionAxes": 1.4166390895843506,
"ReshapeWeights": 0.02198624610900879,
"ResolveAccessConflict": 0.20142865180969238,
"ResolveComplicatePredicates": 0.6187732219696045,
"RewriteReplicationMatmul": 0.04161477088928223,
"RewriteWeights": 0.0676727294921875,
"SFKVectorizer": 11.18409252166748,
"ShardingPropagationAnalysis": 0.8252537250518799,
"SimpleAllReduceTiling": 0.18207907676696777,
"Simplifier": 0.09509730339050293,
"SimplifyMacroPredicates": 0.26697540283203125,
"SimplifyNeuronTensor": 0.3594348430633545,
"SimplifySlice": 0.02629375457763672,
"SimplifyTensor": 0.42005443572998047,
"SpillPSum": 0.5635168552398682,
"SplitAPUnionSets": 0.4571385383605957,
"SplitAccGrp": 0.052629709243774414,
"StaticProfiler": 0.12954020500183105,
"StaticTransposeLocalTensor": 0.36740827560424805,
"SundaISel": 1.4737660884857178,
"TCTransform": 0.031054019927978516,
"TensorInitialization": 0.18288874626159668,
"TensorOpSimplifier": 3.2880165576934814,
"TensorOpTransform": 18.079126358032227,
"TileCCOps": 0.17596793174743652,
"TilingProfiler": 0.3990769386291504,
"TransformConvOp": 1.2077322006225586,
"TritiumFusion": 0.2976958751678467,
"ValueNumbering": 0.10138130187988281,
"VectorizeDMA": 0.8864037990570068,
"VectorizeMatMult": 0.05278921127319336,
"WeightCoalescing": 0.07968640327453613,
"ZeroSizeTensorElimination": 0.0006525516510009766
},
"tensorizer": {
"DMATilingProfiler::TotalInstructionsAfterTiling": 42842,
"StaticProfiler::AifUb": 17.765029907226563,
"StaticProfiler::ArithmeticIntensityTensorizer": 28.541391372680664,
"StaticProfiler::AverageDmaLength": 3512.794677734375,
"StaticProfiler::AverageFractalPeUtilization": 97.62759399414063,
"StaticProfiler::AveragePartitionUtilization": 90.06017303466797,
"StaticProfiler::AveragePeUtilization": 82.14405822753906,
"StaticProfiler::DDRTransferBytes": 924925552,
"StaticProfiler::InternalTransferBytes": 182108768,
"StaticProfiler::LoadExpanded": 181148,
"StaticProfiler::LocalizationEfficiency": 160.6605224609375,
"StaticProfiler::LocalizationEfficiencyIgnoreNonlocal": 165.92486572265625,
"StaticProfiler::StoreExpanded": 4273,
"StaticProfiler::TotalDMAExpanded": 185421,
"StaticProfiler::TotalDynamicInstancesCount": 57693,
"StaticProfiler::TotalDynamicInstancesWithMmPackedCount": 50217,
"StaticProfiler::TotalLNCComm": 0,
"StaticProfiler::TotalLNCCommTransfer": 0,
"TilingProfiler::AveragePartitionUtilizationAfterTiling": 0,
"TilingProfiler::AveragePeUtilizationAfterTiling": 0,
"TilingProfiler::BatchnormInstructionsAfterTiling": 0,
"TilingProfiler::DmaInstructionsAfterTiling": 0,
"TilingProfiler::GenericInstructionsAfterTiling": 123,
"TilingProfiler::MatMultInstructionsAfterTiling": 30976,
"TilingProfiler::NumPfTransposes": 348,
"TilingProfiler::NumPfTransposesForIo": 30,
"TilingProfiler::NumPfTransposesForLocal": 198,
"TilingProfiler::NumPfTransposesForNonlocal": 120,
"TilingProfiler::PfTransposeInstructions": 7892,
"TilingProfiler::PfTransposeInstructionsForIo": 5666,
"TilingProfiler::PfTransposeInstructionsForLocal": 564,
"TilingProfiler::PfTransposeInstructionsForNonlocal": 1662,
"TilingProfiler::ReduceInstructionsAfterTiling": 115,
"TilingProfiler::SimdInstructionsAfterTiling": 2295,
"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": 0.003673076629638672,
"DMALocalityOpt": 0.0026640892028808594,
"DMAProfiler": 0.0036766529083251953,
"DataStreaming": 0.00630640983581543,
"DoNothing": 0.00016999244689941406,
"ExpandISAMacro": 0.003720521926879883,
"FactorizeBlkDims": 0.015497684478759766,
"InferPSumTensor": 0.011394739151000977,
"InferSharedMemLoc": 0.002842426300048828,
"InsertCoreBarrier": 0.0031981468200683594,
"LateLegalizeInst": 0.0077512264251708984,
"LateNeuronInstComb": 0.008512020111083984,
"LegalizeSundaAccess": 0.015286922454833984,
"LegalizeType": 0.00956106185913086,
"LowerBroadcast": 0.003253936767578125,
"LowerIntrinsics": 0.003645658493041992,
"LowerTranspose": 0.0033075809478759766,
"NeuronInstComb": 0.008723735809326172,
"NeuronLICM": 0.009400129318237305,
"NeuronSimplifyPredicates": 0.0036191940307617188,
"NeuronValueNumbering": 0.003929615020751953,
"SFKVectorizer": 0.033834218978881836,
"SimpleAllReduceTiling": 0.003624439239501953,
"SimplifyNeuronTensor": 0.05256152153015137,
"SpillPSum": 0.024464845657348633,
"WeightCoalescing": 0.003454923629760742
}
}
}

3
token_generation_model/_tp0_bk1/graph.neff Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:cda6f9bc52718ce93e1fb58a7aa253d606cf545bda03b396e531716e6a9ce678
size 3154944

4469
token_generation_model/_tp0_bk1/log-neuron-cc.txt Normal file
View File

File diff suppressed because it is too large Load Diff

3
token_generation_model/_tp0_bk1/metaneff.pb Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:fd8b7f01387ce1b80997167a938d7a05e712ade8a8271b5da3b01c446894cbd8
size 2464555

3
token_generation_model/_tp0_bk1/model.MODULE_f53407701fa4882a24c0+55c11e15.hlo_module.pb Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:d9364e7ab828d2cbc2c983e088e9c293cdaddc1137d948ee2d3c625e9eab02f7
size 2550451

3
token_generation_model/_tp0_bk1/model.MODULE_f53407701fa4882a24c0+55c11e15.neff Normal file
View File

@@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:cda6f9bc52718ce93e1fb58a7aa253d606cf545bda03b396e531716e6a9ce678
size 3154944

224
token_generation_model/_tp0_bk1/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": [
256
],
"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": [
256
],
"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
}