从 TF 服务 api 获取<Response [400]>

Question

我已经训练了一个模型来使用 TF 和 TFX 对叶子进行分类。我是 TFServing 在本地环境中部署模型。虽然我能够部署模型但是当我向模型发出 POST 请求时，它返回 或

requests.exceptions.HTTPError: 400 Client Error: Bad Request for url: http://localhost:8501/v1/models/leaf_classification_native_keras_8/versions/1:predict

下面是部署模型的命令

docker run -p 8501:8501 --name tfserving_leaf_classification --mount  type=bind,source=/home/bhargavpatel/Desktop/Image_Classification_tfx/Image_Classification/tfx/pipelines/leaf_classification_native_keras_8/Pusher/pushed_model,target=/models/leaf_classification_native_keras_8 -e MODEL_NAME=leaf_classification_native_keras_8 -t tensorflow/serving &

发帖请求的Python代码

import base64
import io
import json

import numpy as np
from PIL import Image
import requests
import tensorflow as tf
import cv2
import pprint

SERVER_URL = "http://localhost:8501/v1/models/leaf_classification_native_keras_8:predict"

IMAGE_PATH = (    "/home/bhargavpatel/Desktop/Image_Classification_tfx/Image_Classification/raw_data/test/healthy/healthy_test.0.jpg"
)

def main():
    headers = {"content-type": "serving_default"}
    img = cv2.imread(IMAGE_PATH, 1)
    img = cv2.resize(img, (224, 224), interpolation=cv2.INTER_AREA)
    img = tf.keras.applications.mobilenet.preprocess_input(img)
    image_content = img.astype("uint8").tolist()
    data = json.dumps({"signature_name": "serving_default", "instances": image_content})
    for _ in range(3):
        response = requests.post(SERVER_URL, json=data, headers=headers)
        pprint.pprint(response)

    total_time = 0
    num_requests = 10
    for _ in range(num_requests):
        response = requests.post(SERVER_URL, json=data, headers=headers)
        response.raise_for_status()
        total_time += response.elapsed.total_seconds()
        prediction = response.json()["predictions"][0]

    print("Prediction class: {}, avg latency: {} ms".format(np.argmax(prediction), (total_time * 1000) / num_requests))


if __name__ == "__main__":
    main()

我使用下面提到的命令检查了是否可以访问 docker 容器

curl http://localhost:8501/v1/models/leaf_classification_native_keras_8

收到以下回复

{
 "model_version_status": [
  {
   "version": "215",
   "state": "AVAILABLE",
   "status": {
    "error_code": "OK",
    "error_message": ""
   }
  }
 ]
}

但是当我发布预测请求时，它会给出错误 400 响应。

Answer 1

您能否尝试将标头更改为

headers = {"content-type": "application/json"}

以定义 JSON 数据有效负载。 json数据中的“signature_name”是可选的，可以忽略。

img_content

需要更新如下。

image_content = img.astype('float32').to_list()