预测失败：检查输入时出错：预期dense_input 具有形状(2898,) 但得到形状为(1,) 的数组

问题描述

我正在使用以下脚本 predictor.py 以从 GCP AI Platform 中托管的 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)

但是我明白

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，行输出 = self._model.predict([vectors])。模型期望形状为 (2898,)

我觉得这很奇怪...因为当我打印矢量的形状和尺寸时，我得到以下内容

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，我做了一个测试来获取本地存储的模型的预测，并且运行良好。这是测试文件：

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

class MyPredictor(object):

    def __init__(self,**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)

解决方法

解决了！

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

中添加一组括号一样愚蠢

在那之后，我遇到了另一个错误，即预测的输出不是 json 可序列化的。我只是通过将 .tolist() 添加到返回 return output.to_list()

中解决了这个问题

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)