預測失敗：檢查輸入時出錯：預期dense_input 具有形狀（2898，）但得到的數組具有形狀（1，）

Question

我正在使用以下腳本 predictor.py 以從 GCP AI 平台中托管的 Keras model 獲取預測。

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,)

問題似乎是由於我在嘗試進行實際預測時輸入數據的維度，在 output = self._model.predict([vectors]) 行中。 model 需要一個形狀為 (2898, ) 的向量

我覺得這很奇怪......因為當我打印矢量的形狀和尺寸時，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]]

所以尺寸和形狀都很好，它應該真的可以工作......

此外，我進行了測試以獲取本地存儲的 model 的預測，它工作正常。 這是測試文件：

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)

Answer 1

解決了！

就像在這一行中添加一組括號一樣愚蠢output = self._model.predict([vectors])

之后，我收到另一個關於預測的 output 不是 json 可序列化的錯誤。 我通過將.tolist() 添加到返回return output.to_list() ) 來解決這個問題

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)