第一步是确保您导出的图形具有可以接受 JPEG 数据的占位符和操作。请注意,CloudML 假设您要发送一批图像。我们必须使用 atf.map_fn
来解码和调整一批图像的大小。根据模型的不同,可能需要对数据进行额外的预处理以规范化数据等。如下所示:
# Number of channels in the input image
CHANNELS = 3
# Dimensions of resized images (input to the neural net)
HEIGHT = 200
WIDTH = 200
# A placeholder for a batch of images
images_placeholder = tf.placeholder(dtype=tf.string, shape=(None,))
# The CloudML Prediction API always "feeds" the Tensorflow graph with
# dynamic batch sizes e.g. (?,). decode_jpeg only processes scalar
# strings because it cannot guarantee a batch of images would have
# the same output size. We use tf.map_fn to give decode_jpeg a scalar
# string from dynamic batches.
def decode_and_resize(image_str_tensor):
"""Decodes jpeg string, resizes it and returns a uint8 tensor."""
image = tf.image.decode_jpeg(image_str_tensor, channels=CHANNELS)
# Note resize expects a batch_size, but tf_map supresses that index,
# thus we have to expand then squeeze. Resize returns float32 in the
# range [0, uint8_max]
image = tf.expand_dims(image, 0)
image = tf.image.resize_bilinear(
image, [HEIGHT, WIDTH], align_corners=False)
image = tf.squeeze(image, squeeze_dims=[0])
image = tf.cast(image, dtype=tf.uint8)
return image
decoded_images = tf.map_fn(
decode_and_resize, images_placeholder, back_prop=False, dtype=tf.uint8)
# convert_image_dtype, also scales [0, uint8_max] -> [0, 1).
images = tf.image.convert_image_dtype(decoded_images, dtype=tf.float32)
# Then shift images to [-1, 1) (useful for some models such as Inception)
images = tf.sub(images, 0.5)
images = tf.mul(images, 2.0)
# ...
此外,我们需要确保正确标记输入,在这种情况下,输入的名称(映射中的键)必须以_bytes
. 在发送 base64 编码数据时,它会让 CloudML 预测服务知道它需要对数据进行解码:
inputs = {"image_bytes": images_placeholder.name}
tf.add_to_collection("inputs", json.dumps(inputs))
gcloud 命令预期的数据格式为:
{"image_bytes": {"b64": "dGVzdAo="}}
(注意,如果image_bytes
是模型的唯一输入,您可以简化为{"b64": "dGVzdAo="}
)。
例如,要从磁盘上的文件创建它,您可以尝试以下操作:
echo "{\"image_bytes\": {\"b64\": \"`base64 image.jpg`\"}}" > instances
然后将其发送到服务,如下所示:
gcloud beta ml predict --instances=instances --model=my_model
请注意,当直接向服务发送数据时,您发送的请求正文需要包装在“实例”列表中。所以上面的 gcloud 命令实际上在 HTTP 请求的正文中向服务发送了以下内容:
{"instances" : [{"image_bytes": {"b64": "dGVzdAo="}}]}