search

AI-performance / embedded-ai.bench

benchmark for embededded-ai deep learning inference engines, such as NCNN / TNN / MNN / TensorFlow Lite etc.
https://www.ai-performance.com
Other
202 stars 29 forks source link

[ENHANCE] 需要增加SoC型号，CPU、GPU型号在对应的结果bench中展示 #21

Closed ysh329 closed 4 years ago

ysh329 commented 4 years ago

需要增加SoC型号，CPU、GPU型号在对应的结果bench中展示，在对应的device.py里增加对应方法获取这些信息

ysh329 commented 4 years ago

https://github.com/AI-performance/embedded-ai.bench/commit/a6468ff707d995b85f4711127e62aec030b08d5c

ysh329 commented 4 years ago

32785cf

ysh329 commented 4 years ago

短期来看，这个字典还需要补充./utils/device.py，长期来看，需要C++的profiler获取手机soc信息。

ysh329 commented 4 years ago
framework branch commit_id model_name platform soc_code soc_name cpu gpu npu product power_mode backend cpu_thread_num avg max min std_dev battery_level system_version repeats warmup imei
mnn master d7fb0ed caffe_mobilenetv1 android-armv8 kona SD865 Kryo585:1xA76@2.84+3xA77@2.42+4XA55@1.8 Adreno-650 TODO V1981A big_cores ARM 4 10.131 10.743 10.044 0.077 100 10 100 2 A00000D3EB320E
mnn master d7fb0ed caffe_mobilenetv2 android-armv8 kona SD865 Kryo585:1xA76@2.84+3xA77@2.42+4XA55@1.8 Adreno-650 TODO V1981A big_cores ARM 4 8.529 8.686 8.467 0.041 100 10 100 2 A00000D3EB320E
mnn master d7fb0ed tf_mobilenetv1 android-armv8 kona SD865 Kryo585:1xA76@2.84+3xA77@2.42+4XA55@1.8 Adreno-650 TODO V1981A big_cores ARM 4 10.122 10.298 10.046 0.047 100 10 100 2 A00000D3EB320E
mnn master d7fb0ed tf_mobilenetv2 android-armv8 kona SD865 Kryo585:1xA76@2.84+3xA77@2.42+4XA55@1.8 Adreno-650 TODO V1981A big_cores ARM 4 6.565 6.765 6.491 0.049 100 10 100 2 A00000D3EB320E