search

ctuning / ck-mlperf

This repository is outdated! Join the open MLPerf workgroup to participate in the development of the next generation of automation workflows for MLPerf benchmarks:
https://bit.ly/mlperf-edu-wg
BSD 3-Clause "New" or "Revised" License
32 stars 23 forks source link

Edgetpu Pull Request #55

Closed slahiruk closed 4 years ago

slahiruk commented 4 years ago

https://github.com/ctuning/ck-mlperf/issues/47 <-- Addressing this

Installation is working fine.

float is working for all tflite and armnn-tflite versions uint8 works for tflite while for armnn this happens:

Loading graph...                                                                                                             
terminate called after throwing an instance of 'armnn::InvalidArgumentException'                                             
  what():  Invalid dimension index: 1 (number of dimensions is 1) at function CheckDimensionIndex [/home/lahiru/CK-TOOLS/lib-
armnn-20.05-gcc-8.3.0-neon-onnx-rel.20.05-tf-tflite-linux-32/src/src/armnn/Tensor.cpp:129]                                   
./tmp-HYjZs5.sh: line 33: 14607 Aborted                 ./classification                                                     
  (post processing via CK (/home/lahiru/CK/ck-mlperf/script/image-classification, postprocess)

int8 gives this error

Loading graph...
ERROR: Unsupported type of graph's input: 8. Supported types are: Float32 (1), UInt8 (2)

  (post processing via CK (/home/lahiru/CK/ck-mlperf/script/image-classification, postprocess)