Image Classification using Mobilenet models

Install MLC following the installation page.

Mobilenet models are not official MLPerf models and so cannot be used for a Closed division MLPerf inference submission. But since they can be run with Imagenet dataset, we are allowed to use them for Open division submission. Only CPU runs are supported now.

TFLite Backend

Mobilenet-V1Mobilenet-V2Mobilenet-V2MobilenetsEfficientnet

Mobilenet V1

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_mobilenet-v1 \
 --adr.compiler.tags=gcc

Mobilenet V2

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_mobilenet-v2 \
 --adr.compiler.tags=gcc

Mobilenet V2

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_mobilenet-v2 \
 --adr.compiler.tags=gcc

Mobilenet V1,V2,V3

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_mobilenet \
 --adr.compiler.tags=gcc

Efficientnet

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_efficientnet \
 --adr.compiler.tags=gcc

ARMNN Backend

Mobilenet-V1Mobilenet-V2Mobilenet-V2MobilenetsEfficientnet

Mobilenet V1

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_armnn,_mobilenet-v1 \
 --adr.compiler.tags=gcc

Mobilenet V2

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_armnn,_mobilenet-v2 \
 --adr.compiler.tags=gcc

Mobilenet V2

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_armnn,_mobilenet-v2 \
 --adr.compiler.tags=gcc

Mobilenet V1,V2,V3

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_armnn,_mobilenet \
 --adr.compiler.tags=gcc

Efficientnet

pip install cmx4mlperf
cr run,mobilenet-models,_tflite,_armnn,_efficientnet \
 --adr.compiler.tags=gcc