adding initial changes to forked repo

Dockerfile

@@ -5,12 +5,13 @@ RUN apt-get update && \
     apt-get install -y --no-install-recommends apt-utils && \
     apt-get -y upgrade
 
-# clone the repository 
+# clone the repository
 RUN git clone --depth 1 https://github.com/tensorflow/models.git
 
 # Install object detection api dependencies
 RUN apt-get install -y protobuf-compiler python-pil python-lxml python-tk && \
     pip install Cython && \
+    pip install tf_slim && \
     pip install contextlib2 && \
     pip install jupyter && \
     pip install matplotlib && \
@@ -47,8 +48,14 @@ WORKDIR /Tensorflow-2-Object-Detection-API-Flask-Application
 # change here to download your pretrained model
 RUN mkdir models && \
     cd models/ && \
-    curl -O "http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03.tar.gz" && \
-    tar xzf ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03.tar.gz && \
-    rm ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03.tar.gz
+    #curl -O "http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03.tar.gz" && \
+    #tar xzf ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03.tar.gz && \
+    #rm ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03.tar.gz
+
+    curl -O "http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_coco_2018_01_28.tar.gz" && \
+    tar xzf ssd_mobilenet_v1_coco_2018_01_28.tar.gz && \
+    rm ssd_mobilenet_v1_coco_2018_01_28.tar.gz
 
-CMD ["python", "main.py"]
+
+
+CMD ["python", "main.py"]

client.py

@@ -4,6 +4,9 @@
 import cv2
 import argparse
 import os
+import time
+
+start_time = time.perf_counter()
 
 address = 'http://127.0.0.1:5000'
 img_path = 'images/girl_image.jpg'
@@ -25,7 +28,7 @@ def parse_args():
 
 def post_image(img_path, URL):
     """ post image and return the response """
- 
+
     img = open(img_path, 'rb').read()
     response = requests.post(URL, data=img, headers=headers)
     return response
@@ -50,5 +53,8 @@ def write_image(save_dir, parsed_response):
     print("detections are: ", parsed_response["detections"])
     print("the image size is: ", parsed_response["image size"])
 
-    write_image(args.output_dir, parsed_response)
-
+    # write_image(args.output_dir, parsed_response)
+
+    finish_time = time.perf_counter()
+    total_time = finish_time - start_time
+    print (total_time)

main.py

@@ -10,8 +10,11 @@
 DEFAULT_PORT = 5000
 DEFAULT_HOST = '0.0.0.0'
 
-model_path = "models/ssd_mobilenet_v1_fpn_shared_box_predictor_640x640_coco14_sync_2018_07_03/saved_model"
+# model_path = "models/ssd_mobilenet_v2_coco_2018_03_29/saved_model"
+model_path = "models/ssd_mobilenet_v1_coco_2018_01_28/saved_model"
+# model_path = "models/ssd_mobilenet_v1_ppn_shared_box_predictor_300x300_coco14_sync_2018_07_03/saved_model"
 labels_path = "data/mscoco_label_map.pbtxt"
+print (model_path)
 
 vis_threshold = 0.5
 max_boxes = 20
@@ -62,7 +65,7 @@ def infer():
 
     # build a response dict to send back to client
     response = parsed_output_dict
-    
+
     # encode response
     response_encoded = json.dumps(response, cls=NumpyArrayEncoder)
 

utilities.py

@@ -11,7 +11,11 @@
 
 from collections import defaultdict
 from io import StringIO
+
+import matplotlib
+matplotlib.use('TkAgg')
 from matplotlib import pyplot as plt
+
 import cv2
 
 from object_detection.utils import ops as utils_ops
@@ -50,7 +54,7 @@ def run_inference_for_single_image(model, image):
     # Convert to numpy arrays, and take index [0] to remove the batch dimension.
     # We're only interested in the first num_detections.
     num_detections = int(output_dict.pop('num_detections'))
-    output_dict = {key:value[0, :num_detections].numpy() 
+    output_dict = {key:value[0, :num_detections].numpy()
                     for key,value in output_dict.items()}
     output_dict['num_detections'] = num_detections
 
@@ -62,7 +66,7 @@ def run_inference_for_single_image(model, image):
         # Reframe the the bbox mask to the image size.
         detection_masks_reframed = utils_ops.reframe_box_masks_to_image_masks(
                     output_dict['detection_masks'], output_dict['detection_boxes'],
-                    image.shape[0], image.shape[1])      
+                    image.shape[0], image.shape[1])
         detection_masks_reframed = tf.cast(detection_masks_reframed > 0.5,
                                             tf.uint8)
         output_dict['detection_masks_reframed'] = detection_masks_reframed.numpy()