Almost 4X slower than an i7 chip

apple / tensorflow_macos

TensorFlow for macOS 11.0+ accelerated using Apple's ML Compute framework.

Other

3.67k stars 309 forks source link

Almost 4X slower than an i7 chip #54

Open wenmin-wu opened 3 years ago

wenmin-wu commented 3 years ago

I developed an "Embedding + MLP" neural network and read zipped TFRecords using tf.data.TFRecordDataset, the time consumed per step is about 75ms compared to my i7 chip MBP which only takes 22ms per step, it's almost 4X slower.

anna-tikhonova commented 3 years ago

Thank you for reporting this issue! Could you provide a reproducible test case, so we can investigate this locally?