Checklist for ONE 1.21.0 Release

[chunseoklee]

Checklist

Release Action

• [x] Release Note #9674
• [x] Update debian package changelog ( before branch out ) #9675
• [x] After Branch out
  • [x] master branch version up #9677
    • [x] tizen spec file
    • [x] nnfw_version.h
    • [x] document version in /docs/conf.py
    • [x] onert aar version
• [x] #9676

Release Package

• Runtime
  • [x] arm64-android (tar.gz) for Android 10
  • [x] arm64-android (aar) for Android 10
  • [x] arm-tizen (rpm) for Tizen 6.0 M2
  • [x] arm-linux (tar.gz) for Ubuntu 18.04
  • [x] x64-linux (tar.gz) for Ubuntu 18.04
  • [x] nnfw-devel
  • [x] nnfw-plugin-devel
• Compiler
  • [x] x64-linux (tar.gz) for Ubuntu 18.04
    • [x] test with public models
    • [x] test with in-hose models

Release Criteria

• Runtime : TEM (>4.0), TCM(New Negitive / New Positive > 1), Svace & Coverity (critical=0), SAM (>4.0, DC > 4.0)

  • [x] TEM coverage link : 4.05 link
  • [x] TCM link : (248 / 468, Negative / Total TC) link
  • [x] Coverity link : link
  • [x] Svace link : link
  • [x] SAM link : 4.48 link
  • [x] License Verification : one_1.21.0 Report.zip

• Compiler : TC line coverage (>80%), TCM(New Negitive / New Positive > 1), Svace & Coverity (critical=0), SAM (>4.0)

  • [x] TEM coverage link : 4.09 link
  • [x] TCM link : (1014 / 2021) link
  • [x] Svace link : link
  • [x] Coverity link : link
  • [x] SAM link : 4.48 link
  • [x] License Verification : one_1.21.0 Report.zip

Miscellaneous

• [x] nncc preset update if needed
• [x] Security
  • Runtime :
    • [x] Svace_security : link
  • Compiler :
    • [x] Svace_security : link

[chunseoklee]

@seanshpark Could you please post compiler's release candidate as comment ?

[seanshpark]

Compiler frontend

• Support unrolling of LSTM and RNN Ops in one-import-onnx tool
• Introduced new tools one-infer, circle-operator, circle-interpreter
• Introduced Workflow(WIP) in one-cmds
• New option quant_config in one-quantize
• New option fake_quantize in one-quantize
• More Ops supported: Densify
• More Ops for quantization: ReduceMax
• More Ops for mixed-precision quantization (MPQ): LeakyRelu, Neg, Relu6, Squeeze
• More Ops for convert_nchw_to_nhwc option: LogSoftmax, ReduceMax, SplitV, Softmax
• New optimization options in one-optimize: replace_non_const_fc_with_bmm, resolve_customop_splitv, fold_densify
• Improved reshape elimination in convert_nchw_to_nhwc option.
• Support fusion of Channel-wise Add + Relu with TConv
• Support negative axis in ArgMin/Max
• Show errors for unrecognized options in one-optimize
• Fix shape inference for StridedSlice
• Fix FuseBatchNormWithTConvPass to support TConv with bias
• Deprecate --O1 option in circle2circle
• Support gcc-11
• Support limited Float16 for kernels constants with dequantization to Float32

NOTE (not necessary to add in comment)

• gathered these items from PRs from May

[seanshpark]

@jinevening , please take a look and edit https://github.com/Samsung/ONE/issues/9672#issuecomment-1237554355 if necessary :)

[jinevening]

I slightly updated the contents as follows. PTAL :)

1. Rephrase some items
2. Add some missing items
3. Remove new Op support in luci-interpreter (it seems duplicate with support for quantization).

[seanshpark]

@mhs4670go , if you have anything to change, please update.

[seanshpark]

@chunseoklee , I think we can go with compiler frontend comments

[chunseoklee]

@Samsung/nncc_committers

nncc preset update if needed

not need in this release ? Then, I will branch out soon

[chunseoklee]

@Samsung/nncc_committers

nncc preset update if needed

not need in this release ? Then, I will branch out soon

@mhs4670go confirmed that not needed. Thus I will make a new branch release/1.21.0