commit - /dev/null
commit + 7852cccb5dd074eef9f30057416d0eaa32f39cdf
blob - /dev/null
blob + 38e8c03a2fd1447d8925bc002028a464ad01d5b4 (mode 644)
--- /dev/null
+++ README.md
+# Human-Detection-Using-Drone-During-Disasters
+
+This project focuses on detecting humans in disaster scenarios using three different methods. The system is designed to send a message to a predefined number whenever a human is detected. The detection methods include:
+
+Live Camera Feed: Utilizes the laptop's live camera to detect humans in real-time.
+AI-Generated Dataset: Employs a custom dataset simulating people trapped in earthquake scenarios to train AI models for human detection.
+Drone Footage: Analyzes video recorded by a DJI Naza M Lite drone to identify humans in the footage.
+We used Vonage for sending messages and Visual Studio for project development. The objective of this project is to enhance disaster response times by promptly alerting rescuers to the presence of humans in disaster-stricken areas.
+
+In each scenario, the software sends an alert message to a predefined number upon detecting a human, ensuring a faster response in disaster situations.
+
+
+
blob - /dev/null
blob + cec1768c7a4023a89c124f011c41bfc6b8edf264 (mode 644)
--- /dev/null
+++ coco.names
+person
\ No newline at end of file
blob - /dev/null
blob + 759151b8b9dbb287ce5a9f9a3470681c07710708 (mode 644)
Binary files /dev/null and frozen_inference_graph.pb differ
blob - /dev/null
blob + 1504d776d9d0a6bee9d66b7f9c3c22ffdaa7f264 (mode 644)
--- /dev/null
+++ main.py
+import cv2
+import cvzone
+from collections import deque
+
+thres = 0.55 # Umbral de confianza
+nmsThres = 0.2 # Umbral de NMS (Non-Maxima Suppression)
+
+cap = cv2.VideoCapture('v1.mp4')
+cap.set(3, 640)
+cap.set(4, 480)
+
+# Carga los nombres de las clases
+classNames = []
+classFile = 'coco.names'
+with open(classFile, 'rt') as f:
+ classNames = f.read().strip().split('\n')
+
+# Cargar la configuración y los pesos de la red
+configPath = 'ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt'
+weightsPath = "frozen_inference_graph.pb"
+
+net = cv2.dnn_DetectionModel(weightsPath, configPath)
+net.setInputSize(320, 320)
+net.setInputScale(1.0 / 127.5)
+net.setInputMean((127.5, 127.5, 127.5))
+net.setInputSwapRB(True)
+
+# Inicializa el contador de personas y una lista de seguimiento
+person_count = 0
+trackers = [] # Lista de trackers activos
+max_distance = 50 # Distancia máxima entre dos detecciones para considerar que es la misma persona
+detections = deque(maxlen=20) # Cola para almacenar las últimas detecciones de personas
+
+while True:
+ success, img = cap.read()
+ if not success:
+ break # Sale del bucle si no se pueden leer más frames
+
+ # Detección de objetos
+ classIds, confs, bbox = net.detect(img, confThreshold=thres, nmsThreshold=nmsThres)
+
+ # Verifica si la detección no está vacía
+ if len(classIds) != 0:
+ for classId, conf, box in zip(classIds.flatten(), confs.flatten(), bbox):
+ if 0 < classId <= len(classNames):
+ if classNames[classId - 1].lower() == "person":
+ # Verifica si la persona ya fue contada en los últimos frames
+ x, y, w, h = box
+ detected = False
+
+ for prev_box in detections:
+ prev_x, prev_y, prev_w, prev_h = prev_box
+ distance = ((x - prev_x) ** 2 + (y - prev_y) ** 2) ** 0.5
+ if distance < max_distance:
+ detected = True
+ break
+
+ if not detected:
+ person_count += 1 # Incrementa el contador solo si no ha sido detectada recientemente
+ detections.append(box) # Agrega la nueva detección a la cola
+
+ # Dibuja el cuadro alrededor de la persona
+ cvzone.cornerRect(img, box)
+ cv2.putText(img, f'PERSON {round(conf * 100, 2)}%',
+ (box[0] + 10, box[1] + 30), cv2.FONT_HERSHEY_COMPLEX_SMALL,
+ 1, (0, 255, 0), 2)
+
+ cv2.imshow("Image", img)
+ cv2.waitKey(1)
+
+# Al finalizar el video, muestra el conteo de personas
+print(f'Cantidad total de personas detectadas: {person_count}')
blob - /dev/null
blob + 2fb4408bf540d9571945aca7fd11486987c3c429 (mode 644)
--- /dev/null
+++ on_dataset.py
+import os
+import cv2
+import numpy as np
+
+# Load YOLO
+net = cv2.dnn.readNet("yolov3.weights", "yolov3.cfg")
+classes = []
+with open("coco.names", "r") as f:
+ classes = [line.strip() for line in f.readlines()]
+
+layer_names = net.getLayerNames()
+output_layers = [layer_names[i[0] - 1] for i in net.getUnconnectedOutLayers()]
+
+# Load dataset
+dataset_path = r"YOUR_PATH"
+image_files = [file for file in os.listdir(dataset_path) if file.endswith((".jpg", ".jpeg"))]
+
+# Loop through images in the dataset
+for image_file in image_files:
+ # Read image
+ image = cv2.imread(os.path.join(dataset_path, image_file))
+ if image is None:
+ print(f"Error: Unable to read image file '{image_file}'. Skipping...")
+ continue
+ height, width, channels = image.shape
+
+ # Detecting objects
+ blob = cv2.dnn.blobFromImage(image, 0.00392, (416, 416), (0, 0, 0), True, crop=False)
+ net.setInput(blob)
+ outs = net.forward(output_layers)
+
+ # Showing information on the screen
+ class_ids = []
+ confidences = []
+ boxes = []
+ for out in outs:
+ for detection in out:
+ scores = detection[5:]
+ class_id = np.argmax(scores)
+ confidence = scores[class_id]
+ if confidence > 0.5: # Confidence threshold
+ # Object detected
+ center_x = int(detection[0] * width)
+ center_y = int(detection[1] * height)
+ w = int(detection[2] * width)
+ h = int(detection[3] * height)
+
+ # Rectangle coordinates
+ x = int(center_x - w / 2)
+ y = int(center_y - h / 2)
+
+ boxes.append([x, y, w, h])
+ confidences.append(float(confidence))
+ class_ids.append(class_id)
+
+ indexes = cv2.dnn.NMSBoxes(boxes, confidences, 0.5, 0.4) # Non-maximum suppression
+
+ font = cv2.FONT_HERSHEY_PLAIN
+ for i in range(len(boxes)):
+ if i in indexes:
+ x, y, w, h = boxes[i]
+ label = str(classes[class_ids[i]])
+ color = (255, 0, 0)
+ cv2.rectangle(image, (x, y), (x + w, y + h), color, 2)
+ cv2.putText(image, label, (x, y + 30), font, 3, color, 3)
+ print(f"Object detected: {label} (Confidence: {confidences[i]})")
+
+ # Show image
+ #cv2.imshow("Image", image)
+ #cv2.waitKey(0)
+ #cv2.destroyAllWindows()
blob - /dev/null
blob + 2e069e45aa9bda19ee64aefe964ec96466cd6dd2 (mode 644)
Binary files /dev/null and output.avi differ
blob - /dev/null
blob + 514860d693735bdc4f0ed2dd62a8fa9b38377e5b (mode 644)
--- /dev/null
+++ recorded vedio.py
+import cv2
+import cvzone
+from sms import send_msg
+
+thres = 0.55
+nmsThres = 0.2
+
+# Path to your video file
+video_path = r"YOUR_PATH"
+
+classNames = []
+classFile = 'coco.names'
+with open(classFile, 'rt') as f:
+ classNames = f.read().split('\n')
+
+configPath = 'ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt'
+weightsPath = "frozen_inference_graph.pb"
+
+net = cv2.dnn_DetectionModel(weightsPath, configPath)
+net.setInputSize(320, 320)
+net.setInputScale(1.0 / 127.5)
+net.setInputMean((127.5, 127.5, 127.5))
+net.setInputSwapRB(True)
+
+# Open the video file
+cap = cv2.VideoCapture(video_path)
+
+# Get the video's frame width and height
+frame_width = int(cap.get(3))
+frame_height = int(cap.get(4))
+
+# Specify the codec and create VideoWriter object
+out = cv2.VideoWriter('output.avi', cv2.VideoWriter_fourcc(*'MJPG'), 10, (frame_width, frame_height))
+
+sms_sent = False # Flag variable to track SMS sent status
+
+while cap.isOpened():
+ ret, frame = cap.read()
+ if not ret:
+ break
+
+ classIds, confs, bbox = net.detect(frame, confThreshold=thres, nmsThreshold=nmsThres)
+
+ # Send SMS only once for the first frame
+ if not sms_sent:
+ try:
+ send_msg()
+ sms_sent = True # Set flag to True once SMS is sent
+ except Exception as e:
+ print("Error sending SMS:", e)
+
+ if len(classIds) != 0:
+ for classId, conf, box in zip(classIds.flatten(), confs.flatten(), bbox):
+ if classId == 0: # 0 corresponds to the class "person" in COCO dataset
+ cv2.rectangle(frame, (box[0], box[1]), (box[0] + box[2], box[1] + box[3]), (0, 255, 0), 2)
+ cv2.putText(frame, f'{classNames[classId - 1].upper()} {round(conf * 100, 2)}',
+ (box[0] + 10, box[1] + 30), cv2.FONT_HERSHEY_COMPLEX_SMALL,
+ 1, (0, 255, 0), 2)
+
+ # Write the frame into the output video
+ out.write(frame)
+
+ cv2.imshow("Video", frame)
+ if cv2.waitKey(1) & 0xFF == ord('q'):
+ break
+
+# Release everything if job is finished
+cap.release()
+out.release()
+cv2.destroyAllWindows()
blob - /dev/null
blob + da855960e27b80ee0856fca67a8b4b6fbbf1ba03 (mode 644)
--- /dev/null
+++ single_img.py
+#
+# sending sms only once per image in dataset
+
+
+import cv2
+import os
+import cvzone
+from sms import send_msg
+thres = 0.55
+nmsThres = 0.2
+
+# Path to your dataset directory containing images
+dataset_path = r"YOUR_PATH"
+
+classNames = []
+classFile = 'coco.names'
+with open(classFile, 'rt') as f:
+ classNames = f.read().split('\n')
+print(classNames)
+
+configPath = 'ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt'
+weightsPath = "frozen_inference_graph.pb"
+
+net = cv2.dnn_DetectionModel(weightsPath, configPath)
+net.setInputSize(320, 320)
+net.setInputScale(1.0 / 127.5)
+net.setInputMean((127.5, 127.5, 127.5))
+net.setInputSwapRB(True)
+
+# Get a list of image files in the dataset directory
+image_files = [file for file in os.listdir(dataset_path) if file.endswith((".jpg", ".jpeg"))]
+
+sms_sent = False # Flag variable to track SMS sent status
+
+for image_file in image_files:
+ img = cv2.imread(os.path.join(dataset_path, image_file))
+ classIds, confs, bbox = net.detect(img, confThreshold=thres, nmsThreshold=nmsThres)
+
+ # Send SMS only once for each image
+ if not sms_sent:
+ try:
+ send_msg()
+ sms_sent = True # Set flag to True once SMS is sent
+ except Exception as e:
+ print("Error sending SMS:", e)
+
+ for classId, conf, box in zip(classIds.flatten(), confs.flatten(), bbox):
+ cvzone.cornerRect(img, box)
+ cv2.putText(img, f'{classNames[classId - 1].upper()} {round(conf * 100, 2)}',
+ (box[0] + 10, box[1] + 30), cv2.FONT_HERSHEY_COMPLEX_SMALL,
+ 1, (0, 255, 0), 2)
+
+ cv2.imshow("Image", img)
+ cv2.waitKey(0)
+
+cv2.destroyAllWindows()
blob - /dev/null
blob + c983cda0a21f3a1244f6e5af2066e8b667ec84e2 (mode 644)
--- /dev/null
+++ sms.py
+import os
+
+import vonage
+
+
+def send_msg():
+ # Credenciales y número de destino desde variables de entorno, nunca en el código:
+ # VONAGE_API_KEY, VONAGE_API_SECRET, SMS_TO (ej. 52XXXXXXXXXX)
+ client = vonage.Client(key=os.environ["VONAGE_API_KEY"], secret=os.environ["VONAGE_API_SECRET"])
+ sms = vonage.Sms(client)
+ responseData = sms.send_message(
+ {
+ "from": "Vonage APIs",
+ "to": os.environ["SMS_TO"],
+ "text": "Se a detectado una persona",
+ }
+ )
+
+ if responseData["messages"][0]["status"] == "0":
+ print("Message sent successfully.")
+ else:
+ print(f"Message failed with error: {responseData['messages'][0]['error-text']}")
+if __name__=='__main__':
+ send_msg()
blob - /dev/null
blob + 64bedd739162d744f7973da2e6ed07eb14976bb3 (mode 644)
--- /dev/null
+++ ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt
+node {
+ name: "normalized_input_image_tensor"
+ op: "Placeholder"
+ attr {
+ key: "dtype"
+ value {
+ type: DT_FLOAT
+ }
+ }
+ attr {
+ key: "shape"
+ value {
+ shape {
+ dim {
+ size: 1
+ }
+ dim {
+ size: 320
+ }
+ dim {
+ size: 320
+ }
+ dim {
+ size: 3
+ }
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv/Conv2D"
+ op: "Conv2D"
+ input: "normalized_input_image_tensor"
+ input: "FeatureExtractor/MobilenetV3/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/Conv/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/Conv/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/Conv/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/Conv/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/Conv/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/Conv/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/Conv/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/Conv/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/Conv/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/Conv/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/Conv/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/Conv/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/Conv/hard_swish/mul_1"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/expand/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_1/output"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/project/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/output"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/expand/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_2/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_1/output"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_2/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/expand/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_3/output"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/project/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/output"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/expand/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_4/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_3/output"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_4/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/expand/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/depthwise/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_5/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/input"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_5/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_6/output"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/project/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/output"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_7/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_6/output"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_7/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_8/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/input"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_8/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_9/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/input"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_9/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_10/output"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/project/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/output"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_11/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_10/output"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_11/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "BoxPredictor_0/ClassPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/mul_1"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/depthwise"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/ClassPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_0/ClassPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_0/ClassPredictor_depthwise/Relu6"
+ input: "BoxPredictor_0/ClassPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/ClassPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_0/ClassPredictor/Conv2D"
+ input: "BoxPredictor_0/ClassPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/BoxEncodingPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/mul_1"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/depthwise"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/BoxEncodingPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_0/BoxEncodingPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_0/BoxEncodingPredictor_depthwise/Relu6"
+ input: "BoxPredictor_0/BoxEncodingPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "loc_pred_transposed"
+ value {
+ b: true
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/BoxEncodingPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_0/BoxEncodingPredictor/Conv2D"
+ input: "BoxPredictor_0/BoxEncodingPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_12/output"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/project/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/output"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_13/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_12/output"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/input"
+ op: "Identity"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_13/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/input"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/expand/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/mul_1/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Mean"
+ op: "Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Mean/reduction_indices"
+ attr {
+ key: "keep_dims"
+ value {
+ b: true
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/Relu"
+ op: "Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/BiasAdd"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv/Relu"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/BiasAdd"
+ op: "BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/BiasAdd"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/Relu6"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/mul/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/depthwise/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/Conv_1/mul"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/squeeze_excite/mul"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/expanded_conv_14/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/project/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/input"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv_1/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/expanded_conv_14/add"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv_1/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/add"
+ op: "AddV2"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/add/y"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/add"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/mul"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/BatchNorm/FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/Relu6"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/mul_1"
+ op: "Mul"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/mul"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/mul_1/y"
+}
+node {
+ name: "BoxPredictor_1/ClassPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/mul_1"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_1/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/depthwise"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_1/ClassPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_1/ClassPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_1/ClassPredictor_depthwise/Relu6"
+ input: "BoxPredictor_1/ClassPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_1/ClassPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_1/ClassPredictor/Conv2D"
+ input: "BoxPredictor_1/ClassPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "BoxPredictor_1/BoxEncodingPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/mul_1"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_1/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/depthwise"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_1/BoxEncodingPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_1/BoxEncodingPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_1/BoxEncodingPredictor_depthwise/Relu6"
+ input: "BoxPredictor_1/BoxEncodingPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "loc_pred_transposed"
+ value {
+ b: true
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_1/BoxEncodingPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_1/BoxEncodingPredictor/Conv2D"
+ input: "BoxPredictor_1/BoxEncodingPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/Conv_1/hard_swish/mul_1"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_2_1x1_256/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512_depthwise/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_2/ClassPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/Relu6"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_2/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/depthwise"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_2/ClassPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_2/ClassPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_2/ClassPredictor_depthwise/Relu6"
+ input: "BoxPredictor_2/ClassPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_2/ClassPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_2/ClassPredictor/Conv2D"
+ input: "BoxPredictor_2/ClassPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "BoxPredictor_2/BoxEncodingPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/Relu6"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_2/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/depthwise"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_2/BoxEncodingPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_2/BoxEncodingPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_2/BoxEncodingPredictor_depthwise/Relu6"
+ input: "BoxPredictor_2/BoxEncodingPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "loc_pred_transposed"
+ value {
+ b: true
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_2/BoxEncodingPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_2/BoxEncodingPredictor/Conv2D"
+ input: "BoxPredictor_2/BoxEncodingPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_2_3x3_s2_512/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_3_1x1_128/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256_depthwise/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_3/ClassPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/Relu6"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_3/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/depthwise"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_3/ClassPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_3/ClassPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_3/ClassPredictor_depthwise/Relu6"
+ input: "BoxPredictor_3/ClassPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_3/ClassPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_3/ClassPredictor/Conv2D"
+ input: "BoxPredictor_3/ClassPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "BoxPredictor_3/BoxEncodingPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/Relu6"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_3/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/depthwise"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_3/BoxEncodingPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_3/BoxEncodingPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_3/BoxEncodingPredictor_depthwise/Relu6"
+ input: "BoxPredictor_3/BoxEncodingPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "loc_pred_transposed"
+ value {
+ b: true
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_3/BoxEncodingPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_3/BoxEncodingPredictor/Conv2D"
+ input: "BoxPredictor_3/BoxEncodingPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_3_3x3_s2_256/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_4_1x1_128/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256_depthwise/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_4/ClassPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/Relu6"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_4/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/depthwise"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_4/ClassPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_4/ClassPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_4/ClassPredictor_depthwise/Relu6"
+ input: "BoxPredictor_4/ClassPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_4/ClassPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_4/ClassPredictor/Conv2D"
+ input: "BoxPredictor_4/ClassPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "BoxPredictor_4/BoxEncodingPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/Relu6"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_4/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/depthwise"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_4/BoxEncodingPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_4/BoxEncodingPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_4/BoxEncodingPredictor_depthwise/Relu6"
+ input: "BoxPredictor_4/BoxEncodingPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "loc_pred_transposed"
+ value {
+ b: true
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_4/BoxEncodingPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_4/BoxEncodingPredictor/Conv2D"
+ input: "BoxPredictor_4/BoxEncodingPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_4_3x3_s2_256/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_1_Conv2d_5_1x1_64/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 2
+ i: 2
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/depthwise"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/Conv2D"
+ op: "Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128_depthwise/Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/Conv2D"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/BatchNorm/gamma"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/BatchNorm/beta"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/BatchNorm/moving_mean"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/Relu6"
+ op: "Relu6"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_5/ClassPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/Relu6"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_5/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/depthwise"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_5/ClassPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_5/ClassPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_5/ClassPredictor_depthwise/Relu6"
+ input: "BoxPredictor_5/ClassPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_5/ClassPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_5/ClassPredictor/Conv2D"
+ input: "BoxPredictor_5/ClassPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "BoxPredictor_5/BoxEncodingPredictor_depthwise/depthwise"
+ op: "DepthwiseConv2dNative"
+ input: "FeatureExtractor/MobilenetV3/layer_17_2_Conv2d_5_3x3_s2_128/Relu6"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/depthwise_weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_5/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+ op: "FusedBatchNormV3"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/depthwise"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/BatchNorm/gamma"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/BatchNorm/beta"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/BatchNorm/moving_mean"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/BatchNorm/moving_variance"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "epsilon"
+ value {
+ f: 0.001
+ }
+ }
+ attr {
+ key: "U"
+ value {
+ type: DT_FLOAT
+ }
+ }
+}
+node {
+ name: "BoxPredictor_5/BoxEncodingPredictor_depthwise/Relu6"
+ op: "Relu6"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/BatchNorm/FusedBatchNormV3"
+}
+node {
+ name: "BoxPredictor_5/BoxEncodingPredictor/Conv2D"
+ op: "Conv2D"
+ input: "BoxPredictor_5/BoxEncodingPredictor_depthwise/Relu6"
+ input: "BoxPredictor_5/BoxEncodingPredictor/weights"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+ attr {
+ key: "dilations"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+ attr {
+ key: "explicit_paddings"
+ value {
+ list {
+ }
+ }
+ }
+ attr {
+ key: "loc_pred_transposed"
+ value {
+ b: true
+ }
+ }
+ attr {
+ key: "padding"
+ value {
+ s: "SAME"
+ }
+ }
+ attr {
+ key: "strides"
+ value {
+ list {
+ i: 1
+ i: 1
+ i: 1
+ i: 1
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_5/BoxEncodingPredictor/BiasAdd"
+ op: "BiasAdd"
+ input: "BoxPredictor_5/BoxEncodingPredictor/Conv2D"
+ input: "BoxPredictor_5/BoxEncodingPredictor/biases"
+ attr {
+ key: "data_format"
+ value {
+ s: "NHWC"
+ }
+ }
+}
+node {
+ name: "concat/axis_flatten"
+ op: "Const"
+ attr {
+ key: "value"
+ value {
+ tensor {
+ dtype: DT_INT32
+ int_val: -1
+ tensor_shape {
+ dim {
+ size: 1
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "BoxPredictor_0/ClassPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_0/ClassPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_1/ClassPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_1/ClassPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_2/ClassPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_2/ClassPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_3/ClassPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_3/ClassPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_4/ClassPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_4/ClassPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_5/ClassPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_5/ClassPredictor/BiasAdd"
+}
+node {
+ name: "ClassPredictor/concat"
+ op: "ConcatV2"
+ input: "BoxPredictor_0/ClassPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_1/ClassPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_2/ClassPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_3/ClassPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_4/ClassPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_5/ClassPredictor/BiasAdd/Flatten"
+ input: "concat/axis_flatten"
+}
+node {
+ name: "BoxPredictor_0/BoxEncodingPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_0/BoxEncodingPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_1/BoxEncodingPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_1/BoxEncodingPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_2/BoxEncodingPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_2/BoxEncodingPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_3/BoxEncodingPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_3/BoxEncodingPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_4/BoxEncodingPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_4/BoxEncodingPredictor/BiasAdd"
+}
+node {
+ name: "BoxPredictor_5/BoxEncodingPredictor/BiasAdd/Flatten"
+ op: "Flatten"
+ input: "BoxPredictor_5/BoxEncodingPredictor/BiasAdd"
+}
+node {
+ name: "BoxEncodingPredictor/concat"
+ op: "ConcatV2"
+ input: "BoxPredictor_0/BoxEncodingPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_1/BoxEncodingPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_2/BoxEncodingPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_3/BoxEncodingPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_4/BoxEncodingPredictor/BiasAdd/Flatten"
+ input: "BoxPredictor_5/BoxEncodingPredictor/BiasAdd/Flatten"
+ input: "concat/axis_flatten"
+}
+node {
+ name: "PriorBox_0"
+ op: "PriorBox"
+ input: "BoxPredictor_0/BoxEncodingPredictor/BiasAdd"
+ input: "normalized_input_image_tensor"
+ attr {
+ key: "clip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "flip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "height"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 32.0
+ float_val: 45.2548339959
+ float_val: 90.5096679919
+ tensor_shape {
+ dim {
+ size: 3
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "variance"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 0.1
+ float_val: 0.1
+ float_val: 0.2
+ float_val: 0.2
+ tensor_shape {
+ dim {
+ size: 4
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "width"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 32.0
+ float_val: 90.5096679919
+ float_val: 45.2548339959
+ tensor_shape {
+ dim {
+ size: 3
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "PriorBox_1"
+ op: "PriorBox"
+ input: "BoxPredictor_1/BoxEncodingPredictor/BiasAdd"
+ input: "normalized_input_image_tensor"
+ attr {
+ key: "clip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "flip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "height"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 112.0
+ float_val: 79.1959594929
+ float_val: 158.391918986
+ float_val: 64.6632301492
+ float_val: 193.99939066
+ float_val: 133.865604245
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "variance"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 0.1
+ float_val: 0.1
+ float_val: 0.2
+ float_val: 0.2
+ tensor_shape {
+ dim {
+ size: 4
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "width"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 112.0
+ float_val: 158.391918986
+ float_val: 79.1959594929
+ float_val: 193.989690448
+ float_val: 64.6599969069
+ float_val: 133.865604245
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "PriorBox_2"
+ op: "PriorBox"
+ input: "BoxPredictor_2/BoxEncodingPredictor/BiasAdd"
+ input: "normalized_input_image_tensor"
+ attr {
+ key: "clip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "flip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "height"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 160.0
+ float_val: 113.13708499
+ float_val: 226.27416998
+ float_val: 92.3760430703
+ float_val: 277.141986657
+ float_val: 182.428068016
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "variance"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 0.1
+ float_val: 0.1
+ float_val: 0.2
+ float_val: 0.2
+ tensor_shape {
+ dim {
+ size: 4
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "width"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 160.0
+ float_val: 226.27416998
+ float_val: 113.13708499
+ float_val: 277.128129211
+ float_val: 92.3714241527
+ float_val: 182.428068016
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "PriorBox_3"
+ op: "PriorBox"
+ input: "BoxPredictor_3/BoxEncodingPredictor/BiasAdd"
+ input: "normalized_input_image_tensor"
+ attr {
+ key: "clip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "flip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "height"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 208.0
+ float_val: 147.078210487
+ float_val: 294.156420974
+ float_val: 120.088855991
+ float_val: 360.284582654
+ float_val: 230.75528163
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "variance"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 0.1
+ float_val: 0.1
+ float_val: 0.2
+ float_val: 0.2
+ tensor_shape {
+ dim {
+ size: 4
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "width"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 208.0
+ float_val: 294.156420974
+ float_val: 147.078210487
+ float_val: 360.266567974
+ float_val: 120.082851399
+ float_val: 230.75528163
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "PriorBox_4"
+ op: "PriorBox"
+ input: "BoxPredictor_4/BoxEncodingPredictor/BiasAdd"
+ input: "normalized_input_image_tensor"
+ attr {
+ key: "clip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "flip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "height"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 256.0
+ float_val: 181.019335984
+ float_val: 362.038671968
+ float_val: 147.801668913
+ float_val: 443.427178651
+ float_val: 278.969532387
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "variance"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 0.1
+ float_val: 0.1
+ float_val: 0.2
+ float_val: 0.2
+ tensor_shape {
+ dim {
+ size: 4
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "width"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 256.0
+ float_val: 362.038671968
+ float_val: 181.019335984
+ float_val: 443.405006738
+ float_val: 147.794278644
+ float_val: 278.969532387
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "PriorBox_5"
+ op: "PriorBox"
+ input: "BoxPredictor_5/BoxEncodingPredictor/BiasAdd"
+ input: "normalized_input_image_tensor"
+ attr {
+ key: "clip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "flip"
+ value {
+ b: false
+ }
+ }
+ attr {
+ key: "height"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 304.0
+ float_val: 214.960461481
+ float_val: 429.920922961
+ float_val: 175.514481834
+ float_val: 526.569774648
+ float_val: 311.897419034
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "variance"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 0.1
+ float_val: 0.1
+ float_val: 0.2
+ float_val: 0.2
+ tensor_shape {
+ dim {
+ size: 4
+ }
+ }
+ }
+ }
+ }
+ attr {
+ key: "width"
+ value {
+ tensor {
+ dtype: DT_FLOAT
+ float_val: 304.0
+ float_val: 429.920922961
+ float_val: 214.960461481
+ float_val: 526.543445501
+ float_val: 175.50570589
+ float_val: 311.897419034
+ tensor_shape {
+ dim {
+ size: 6
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "PriorBox/concat"
+ op: "ConcatV2"
+ input: "PriorBox_0"
+ input: "PriorBox_1"
+ input: "PriorBox_2"
+ input: "PriorBox_3"
+ input: "PriorBox_4"
+ input: "PriorBox_5"
+ input: "concat/axis_flatten"
+}
+node {
+ name: "ClassPredictor/concat3d/shape"
+ op: "Const"
+ attr {
+ key: "value"
+ value {
+ tensor {
+ dtype: DT_INT32
+ int_val: 0
+ int_val: -1
+ int_val: 91
+ tensor_shape {
+ dim {
+ size: 3
+ }
+ }
+ }
+ }
+ }
+}
+node {
+ name: "ClassPredictor/concat3d"
+ op: "Reshape"
+ input: "ClassPredictor/concat"
+ input: "ClassPredictor/concat3d/shape"
+}
+node {
+ name: "ClassPredictor/concat/sigmoid"
+ op: "Sigmoid"
+ input: "ClassPredictor/concat3d"
+}
+node {
+ name: "ClassPredictor/concat/sigmoid/Flatten"
+ op: "Flatten"
+ input: "ClassPredictor/concat/sigmoid"
+}
+node {
+ name: "detection_out"
+ op: "DetectionOutput"
+ input: "BoxEncodingPredictor/concat"
+ input: "ClassPredictor/concat/sigmoid/Flatten"
+ input: "PriorBox/concat"
+ attr {
+ key: "background_label_id"
+ value {
+ i: 0
+ }
+ }
+ attr {
+ key: "code_type"
+ value {
+ s: "CENTER_SIZE"
+ }
+ }
+ attr {
+ key: "confidence_threshold"
+ value {
+ f: 1e-08
+ }
+ }
+ attr {
+ key: "keep_top_k"
+ value {
+ i: 100
+ }
+ }
+ attr {
+ key: "nms_threshold"
+ value {
+ f: 0.6
+ }
+ }
+ attr {
+ key: "num_classes"
+ value {
+ i: 91
+ }
+ }
+ attr {
+ key: "share_location"
+ value {
+ b: true
+ }
+ }
+ attr {
+ key: "top_k"
+ value {
+ i: 100
+ }
+ }
+}
blob - /dev/null
blob + 3a52305187ed1dfff0fdbe909d0ac2252e72d116 (mode 644)
Binary files /dev/null and v1.mp4 differ
blob - /dev/null
blob + 1f3f3b8107eda44d374ab76886468d779e91ef9f (mode 644)
--- /dev/null
+++ yolov3.cfg
+[net]
+# Testing
+# batch=1
+# subdivisions=1
+# Training
+batch=64
+subdivisions=16
+width=608
+height=608
+channels=3
+momentum=0.9
+decay=0.0005
+angle=0
+saturation = 1.5
+exposure = 1.5
+hue=.1
+
+learning_rate=0.001
+burn_in=1000
+max_batches = 500200
+policy=steps
+steps=400000,450000
+scales=.1,.1
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=3
+stride=1
+pad=1
+activation=leaky
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=2
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=32
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=2
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=64
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=2
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=2
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+# Downsample
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=2
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=1024
+size=3
+stride=1
+pad=1
+activation=leaky
+
+[shortcut]
+from=-3
+activation=linear
+
+######################
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=512
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=1024
+activation=leaky
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+
+[yolo]
+mask = 6,7,8
+anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+
+
+[route]
+layers = -4
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[upsample]
+stride=2
+
+[route]
+layers = -1, 61
+
+
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=256
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=512
+activation=leaky
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+
+[yolo]
+mask = 3,4,5
+anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
+
+
+
+[route]
+layers = -4
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[upsample]
+stride=2
+
+[route]
+layers = -1, 36
+
+
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+filters=128
+size=1
+stride=1
+pad=1
+activation=leaky
+
+[convolutional]
+batch_normalize=1
+size=3
+stride=1
+pad=1
+filters=256
+activation=leaky
+
+[convolutional]
+size=1
+stride=1
+pad=1
+filters=255
+activation=linear
+
+
+[yolo]
+mask = 0,1,2
+anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
+classes=80
+num=9
+jitter=.3
+ignore_thresh = .7
+truth_thresh = 1
+random=1
