Could you give this file about config in the serve.py? about matterport-maskrcnn-with-tensorflow-serving

[huyhoang17]

@LymanLiuChina
st like this 😄

import os


# ==============
# === COMMON ===
# ==============
ROOT_FD = "."

# ==========================
# === SEGMENTATION MODEL ===
# ==========================
# Make it True if you want to use the provided coco weights
IS_COCO = False

# keras model directory path
MODEL_FD = os.path.join(ROOT_FD, "models")

# keras model file path
H5_WEIGHT_PATH = os.path.join(
    MODEL_FD,
    ""
)

# Path where the Frozen PB will be save
PATH_TO_SAVE_FROZEN_PB = os.path.join(
    MODEL_FD,
    "frozen_model"
)

# Name for the Frozen PB name
FROZEN_NAME = 'mask_frozen_graph.pb'

# PATH where to save serving model
PATH_TO_SAVE_TENSORFLOW_SERVING_MODEL = \
    os.path.join(ROOT_FD, "serving_model")

# Version of the serving model
VERSION_NUMBER = 1
OD_VERSION_NUMBER = 1

# Number of classes that you have trained your model
NUMBER_OF_CLASSES = 1

# Host-port gRPC
HOST = "http://..."
gRPC_PORT = 8500
REST_PORT = 8501
GRPC_TIMEOUT = 5
MODEL_SIG_NAME = ""
MODEL_SPEC_NAME = ""

REST_MODE = "predict"

# REST-API meta-data
REST_URL_FMT = "{}:{}/v1/models/{}:{}"
REST_URL = REST_URL_FMT.format(
    HOST,
    REST_PORT,
    MODEL_SPEC_NAME,
    REST_MODE
)

# Input meta-data
IMAGE_SIZE = 640
IN_TENSOR_IMAGE = "input_image"
IN_TENSOR_IMAGE_META = "input_image_meta"
IN_TENSOR_ANCHORS = "input_anchors"
IN_TENSOR_DTYPE = "float32"

# Output meta-data
OUT_TENSOR_DETECTION = "mrcnn_detection/Reshape_1"
OUT_TENSOR_MASK = "mrcnn_mask/Reshape_1"
OUT_DETECTION_SHAPE = (6,)
OUT_MASK_SHAPE = (28, 28, NUMBER_OF_CLASSES + 1)

# Class meta-data
CLASSES = []

[LymanLiuChina]

what is this from api.helpers import utils as api_utils ？please，could you give your complete program code?

[LymanLiuChina]

thanks,I got it!

[issue-label-bot]

Issue-Label Bot is automatically applying the label question to this issue, with a confidence of 0.83. Please mark this comment with 👍 or 👎 to give our bot feedback!

Links: app homepage, dashboard and code for this bot.

[huyhoang17]

@LymanLiuChina
Did you read README.md? 😃

I just took some functions in mrcnn/model.py, for example: api_utils.get_anchors & api_utils.unmold_detections