Tensorflow-TensorRT icon indicating copy to clipboard operation
Tensorflow-TensorRT copied to clipboard

Published 20 hours ago • verified publisher icon ardianumam

'TensorRT_model.pb' model's size is twice as large as the original pb model

Open Julyliying opened this issue 5 years ago • 5 comments

Hi, when i use tensorrt to generate trt model, i found the FP16 or INT8 trt model's size is twice as large as the original tensorflow pb model, i think it's wired, and do you know why? my running environment is: GPU 1080Ti, CUDA10.0, cudann7.5, Tensorflow1.13, TensorRT5.0 thanks

Julyliying avatar May 29 '19 08:05 Julyliying

Does anyone know what is wrong?

Julyliying avatar May 29 '19 09:05 Julyliying

my FP16 trt model is 653.5M ， I guess it‘s OK，because at GPU 970M the PFS from 10 to 20

ruyijidan avatar Jul 09 '19 08:07 ruyijidan

can you see the trt model fps？if you can please tell me the detail version ty!!
cudann7.5.?, Tensorflow1.13.?, TensorRT5.0.?

ruyijidan avatar Jul 09 '19 09:07 ruyijidan

GPU 1080Ti, CUDA10.0, cudann7.5, Tensorflow1.13, TensorRT5.0. @ruyijidan

Julyliying avatar Jul 09 '19 09:07 Julyliying

@Julyliying ty！but little obvious effect at 1080ti, cudnn7.5.1.10 Tensorflow1.13,.1 TensorRT5.1.5.0

ruyijidan avatar Jul 09 '19 11:07 ruyijidan