0

我正在使用以下脚本 predictor.py 以便从托管在 GCP AI Platform 中的 Keras 模型中获取预测。

import os
import pickle
import tensorflow as tf
import numpy as np
import logging

class MyPredictor(object):

    def __init__(self, model, bow_model):
        self._model = model
        self._bow_model = bow_model

    def predict(self, instances, **kwargs):
      
        vectors = self.embedding([instances])

        vectors = vectors.tolist()

        output = self._model.predict(vectors)

        return output

    def embedding(self, statement):
        vector = self._bow_model.transform(statement).toarray()
        #vector = vector.to_list()
        return vector


    @classmethod
    def from_path(cls, model_dir):

        model_path = os.path.join(model_dir, 'model.h5')
        model = tf.keras.models.load_model(model_path, compile = False)

        preprocessor_path = os.path.join(model_dir, 'bow.pkl')
        with open(preprocessor_path, 'rb') as f:
            bow_model = pickle.load(f)


        return cls(model, bow_model)

但是我得到

Prediction failed: Error when checking input: expected dense_input to have shape (2898,) but got array with shape (1,)

问题似乎是由于我在尝试进行实际预测时输入数据的维度，在行输出 = self._model.predict([vectors]) 中。该模型需要一个形状为 (2898, ) 的向量

我觉得这很奇怪......因为当我打印矢量的形状和尺寸时，我得到以下信息

This is the shape
(1, 2898)

This is the dim number
2

This is the vector 
[[0 0 0 ... 0 0 0]]

所以尺寸和形状都很好，它应该真的可以工作......

此外，我进行了测试以获取本地存储的模型的预测并且它工作正常。这是测试文件：

import os
import pickle
import tensorflow as tf
import numpy as np

class MyPredictor(object):

    def __init__(self, model, bow_model):
        self._model = model
        self._bow_model = bow_model

    def predict(self, instances, **kwargs):

        print("These are the instances ", instances)

        vector = self.embedding([instances])

        output = self._model.predict(vector)

        return output

    def embedding(self, statement):
        vector = self._bow_model.transform(statement).toarray()
        #vector = vector.to_list()
        return vector



model_path = 'model.h5'
model = tf.keras.models.load_model(model_path, compile = False)

preprocessor_path = 'bow.pkl'
with open(preprocessor_path, 'rb') as f:
    bow_model = pickle.load(f)


instances = 'test'

predictor = MyPredictor(model, bow_model)

outputs = predictor.predict(instances)

print(outputs)
4

1 回答 1

1

解决了！

就像在这行添加一组括号一样愚蠢output = self._model.predict([vectors])

之后，我收到了另一个关于预测输出不是 json 可序列化的错误。我只需将 .tolist() 添加到返回即可解决此问题return output.to_list()

import os
import pickle
import tensorflow as tf
import numpy as np
import logging

class MyPredictor(object):

    def __init__(self, model, bow_model):
        self._model = model
        self._bow_model = bow_model

    def predict(self, instances, **kwargs):
      
        vectors = self.embedding([instances])

        vectors = vectors.tolist()

        output = self._model.predict([vectors])

        return output.to_list()

    def embedding(self, statement):
        vector = self._bow_model.transform(statement).toarray()
        #vector = vector.to_list()
        return vector


    @classmethod
    def from_path(cls, model_dir):

        model_path = os.path.join(model_dir, 'model.h5')
        model = tf.keras.models.load_model(model_path, compile = False)

        preprocessor_path = os.path.join(model_dir, 'bow.pkl')
        with open(preprocessor_path, 'rb') as f:
            bow_model = pickle.load(f)


        return cls(model, bow_model)
于 2021-03-10T17:33:51.817 回答