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

Question

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

Julyliying opened this issue 6 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

Answer 1 · 2019-05-29T09:04:46.000Z

Does anyone know what is wrong?

Answer 2 · 2019-07-09T08:26:29.000Z

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

Answer 3 · 2019-07-09T09:52:36.000Z

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

Answer 4 · 2019-07-09T09:57:22.000Z

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

Answer 5 · 2019-07-09T11:01:40.000Z

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