One: Checklist for release 1.7.0

Created on 1 Jul 2020  路  11Comments  路  Source: Samsung/ONE

Checklist

typrelease

Most helpful comment

Compiler supports more operations

  • MatrixDiag, MatrixSetDiag, ReverseSequence, ReverseV2, SegmentSum, SelectV2, SparseToDense, Where

New command line interface for user interface consistancy

  • one-import: imports conventional models files to circle

    • one-import-tf: imports TensorFlow model to circle

    • one-import-tflite: imports TensorFlow lite model to circle

  • one-optimize: circle optimize command
  • one-quantize: circle quantize command

    • supports float32 to uint8, layer wise (for Conv series)

  • one-pack: package command
  • one-prepare-venv: prepares python virtual environment for import TensorFlow model
  • one-codegen: backend(if available) code generator

All 11 comments

Please add release note content.

Compiler @seanshpark @llFreetimell
Runtime @chunseoklee @hseok-oh @wateret

Compiler supports more operations

  • MatrixDiag, MatrixSetDiag, ReverseSequence, ReverseV2, SegmentSum, SelectV2, SparseToDense, Where

New command line interface for user interface consistancy

  • one-import: imports conventional models files to circle

    • one-import-tf: imports TensorFlow model to circle

    • one-import-tflite: imports TensorFlow lite model to circle

  • one-optimize: circle optimize command
  • one-quantize: circle quantize command

    • supports float32 to uint8, layer wise (for Conv series)

  • one-pack: package command
  • one-prepare-venv: prepares python virtual environment for import TensorFlow model
  • one-codegen: backend(if available) code generator
  • Nothing special with BCQ works.
  • Minor bug fix

Runtime CPU backend operation support

  • LogSoftmax, SpaceToBatchND

Runtime CPU backend quant8 operation support

  • Logistic, Mul, Tanh, SpaceToBatchND, Transpose, Sub, Max, Min, Less, Greater, GreaterEqual, LessEqual, Equal, NotEqual

Added links to SAM and TCM reports to the check list.

Added runtime binary for android (aar) to the check list

Runtime WG1

  • NNFW API

    • Introduce basic asynchronous execution API

  • Optimization

    • Remove dynamic tensor overhead from static models

nncc-1.7.0.tar.gz at http://npuci.mooo.com:8080/job/nnfw/job/release_1.7/job/daily-publish-package/lastSuccessfulBuild/artifact/ works OK

AM part updated

TCM result for runtime updated. Will prepare the result for compiler soon.

... Will prepare the result for compiler soon.

TCM result for compiler updated.

@seanshpark , 27 TCs with No Assertion were detected. PTAL. It should be fixed in the 1.8.0 release.

Test | Test Case | Checker | File | Line
-- | -- | -- | -- | --
NN Compiler | FormattedTensorShapeTest.BracketFormat | GTEST.NO_ASSERTION | /compiler/locop/src/FormattedTensorShape.test.cpp | 25
NN Compiler | CircleIfTestDeath.invalid_output_count_NEG | GTEST.NO_ASSERTION | /compiler/luci/lang/src/Nodes/CircleIf.test.cpp | 46
NN Compiler | CircleIfTestDeath.invalid_arity_NEG | GTEST.NO_ASSERTION | /compiler/luci/lang/src/Nodes/CircleIf.test.cpp | 41
NN Compiler | L2NormalizeTest.Uint8Quantized | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/L2Normalize.test.cpp | 48
NN Compiler | GetNthPercentileTest.FloatWithNegative | GTEST.NO_ASSERTION | /compiler/record-minmax/tests/RecordFunction.test.cpp | 66
NN Compiler | GetNthPercentileTest.SigleElement | GTEST.NO_ASSERTION | /compiler/record-minmax/tests/RecordFunction.test.cpp | 80
NN Compiler | GetNthPercentileTest.Simple | GTEST.NO_ASSERTION | /compiler/record-minmax/tests/RecordFunction.test.cpp | 37
NN Compiler | GetNthPercentileTest.Float | GTEST.NO_ASSERTION | /compiler/record-minmax/tests/RecordFunction.test.cpp | 52
NN Compiler | GetNthPercentileTest.Edge | GTEST.NO_ASSERTION | /compiler/record-minmax/tests/RecordFunction.test.cpp | 29
NN Compiler | LogisticTest.Uint8 | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/Logistic.test.cpp | 52
NN Compiler | ModuleTest.consturctor | GTEST.NO_ASSERTION | /compiler/luci/lang/src/Module.test.cpp | 21
NN Compiler | LeakReluTest.Uint8Simple | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/LeakyRelu.test.cpp | 66
NN Compiler | LeakReluTest.FloatSimple | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/LeakyRelu.test.cpp | 51
NN Compiler | NodeExecution_BiasEncode.f32 | GTEST.NO_ASSERTION | /compiler/locomotiv/src/Node/BiasEncode.test.cpp | 95
NN Compiler | NodeExecution_BiasEncode.s32 | GTEST.NO_ASSERTION | /compiler/locomotiv/src/Node/BiasEncode.test.cpp | 93
NN Compiler | CircleWhileTestDeath.invalid_arity_NEG | GTEST.NO_ASSERTION | /compiler/luci/lang/src/Nodes/CircleWhile.test.cpp | 41
NN Compiler | CircleWhileTestDeath.invalid_output_count_NEG | GTEST.NO_ASSERTION | /compiler/luci/lang/src/Nodes/CircleWhile.test.cpp | 46
NN Compiler | NodeExecution_MatMul.s32_4x2_2x6 | GTEST.NO_ASSERTION | /compiler/locomotiv/src/Node/MatMul.test.cpp | 159
NN Compiler | NodeExecution_MatMul.f32_2x3_3x3 | GTEST.NO_ASSERTION | /compiler/locomotiv/src/Node/MatMul.test.cpp | 121
NN Compiler | TensorFlowLiteImport.Dummy | GTEST.NO_ASSERTION | /compiler/luci/import/src/Importer.test.cpp | 23
NN Compiler | TransposeConvTest.Uint8FiltersTest | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/TransposeConv.test.cpp | 93
NN Compiler | TransposeConvTest.Uint8Simple | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/TransposeConv.test.cpp | 87
NN Compiler | TransposeConvTest.FloatTwoFiltersTest | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/TransposeConv.test.cpp | 73
NN Compiler | TransposeConvTest.FloatSimple | GTEST.NO_ASSERTION | /compiler/luci-interpreter/src/kernels/TransposeConv.test.cpp | 61
NN Compiler | CircleCustomTest.constructor_NEG | GTEST.NO_ASSERTION | /compiler/luci/lang/src/Nodes/CircleCustom.test.cpp | 38
NN Compiler | LinearV1FormatterTest.simple | GTEST.NO_ASSERTION | /compiler/locop/src/FormattedGraph.test.cpp | 24
NN Compiler | HermesTest.simple_usecase | GTEST.NO_ASSERTION | /compiler/hermes/src/hermes.test.cpp | 21

Was this page helpful?
0 / 5 - 0 ratings