Ufdjrw

I am generating a .pb file from the piece of code given below

import tensorflow as tf
with tf.Session() as sess:
 gom = tf.train.import_meta_graph('C:\chhaya\CLITP\Tvs_graphs\job.ckpt-20.meta')
 gom.restore(sess,tf.train.latest_checkpoint('C:\chhaya\CLITP\Tvs_graphs'))
 graph = tf.get_default_graph()
 input_graph = graph.as_graph_def()
 output_node_name = "predictions"
 output_graph = tf.graph_util.convert_variables_to_constants(sess,input_graph,output_node_name.split(','))
 res_file = 'C:\chhaya\CLITP\Tvs_graphs\Savedmodel.pb'
 with tf.gfile.GFile(res_file,'wb') as f:
 f.write(output_graph.SerializeToString())

But while inferencing from the pb file it takes 5seconds for the first image and 3 seconds for others after that. The code for inferencing is given below.

import tensorflow as tf
import os
from tensorflow.python.platform import gfile
from PIL import Image
import numpy as np
import scipy
from scipy import misc
import matplotlib.pyplot as plt
import cv2
import time
from aug_tool import data_aug
frozen_graph = 'C:\chhaya\CLITP\Tvs_graphs\Savedmodel1.pb'
with tf.gfile.GFile(frozen_graph,'rb') as f:
 reco = tf.GraphDef()
 reco.ParseFromString(f.read())

with tf.Graph().as_default() as gre:
 tf.import_graph_def(reco,input_map=None,return_elements=None,name='')

 l_input = gre.get_tensor_by_name('input_image:0') # Input Tensor
 l_output = gre.get_tensor_by_name('predictions:0')
 files=os.listdir('C:\chhayaCLITP\Tvs Mysore\NQCOVERFRONTLR\')
 imageslst=np.zeros((len(files),224,224,3))
 i=0
 for file in files:
 image = scipy.misc.imread('C:\chhayaCLITP\Tvs Mysore\NQCOVERFRONTLR\'+file)
 image = image.astype(np.uint8)
 Input_image_shape=(224,224,3)
 resized_img=cv2.resize(image,(224,224))
 channels = image.shape[2]
 #print(channels)
 if channels == 4:
 resized_img = cv2.cvtColor(resized_img,cv2.COLOR_RGBA2RGB)
 #image = np.expand_dims(resized_img,axis=2)
 #print() 
 image=np.expand_dims(resized_img,axis=0)
 height,width,channels = Input_image_shape
 imageslst[i,:,:,:]=image
 i=i+1

 init = tf.global_variables_initializer()
 with tf.Session(graph=gre) as sess:
 sess.run(init)
 for i in range(4):
 t1=time.time()
 Session_out = sess.run((tf.nn.sigmoid(l_output)) , feed_dict = l_input : imageslst[:1] )
 print(Session_out.shape)
 t2=time.time()
 print('time:'+str(t2-t1))

I am using tensorflow-gpu 1.12 with windows 7 using anaconda python 3.5
I also tried assigning gpu and cpu to predictions something like this

with tf.device('/gpu:0'):
 for i in range(4):
 t1=time.time()
 Session_out = sess.run((tf.nn.sigmoid(l_output)) , feed_dict = l_input : imageslst[:1] )
 print(Session_out.shape)
 t2=time.time()
 print('time:'+str(t2-t1))

The thing I noted here is no matter what I assign cpu or gpu the time taken is always the same.The model is a transfer learning vgg16 model. Am i doing something wrong in the code? My gpu is Quadro 8GB

After a lot of trial and error what I noticed was if I remove the tf.nn.sigmoid from while inferencing then the time taken was less than 10 milliseconds exactly how the model should have performed.
Before we were using the below line to get predictions which resulted in high inference time:-

Session_out = sess.run((tf.nn.sigmoid(l_output)) , feed_dict = l_input : imageslst[:1] )

So instead of using the above line now we use this:-

Session_out = sess.run(l_output , feed_dict = l_input : imageslst[:1] )

This proves to give us the accurate results. Although I am pretty unsure as to why that happens.