使用 InceptionV3 got ValueError: Shapes (None, 9) and (None, 13, 7, 2048) is incompatible

Question

我正在尝试使用 InceptionV3 来训练我的图像。但还是得到了这个ValueError: Shapes (None, 9) and (None, 13, 7, 2048) are incompatible

输入图像尺寸（RGB）：480 x 270（宽 x 高）

输出标签：[0,0,0,0,0,0,0,0,1]（9个输出）

包：

• 张量流GPU 2.2.0
• numpy 1.19.1

这是我的代码：

import numpy as np
from grabscreen import grab_screen
import cv2
import time
import pandas as pd
from random import shuffle
import tensorflow as tf 
from tensorflow.keras.applications.xception import Xception
from tensorflow.keras.applications.inception_v3 import InceptionV3
from tensorflow.keras import optimizers


FILE_I_END = 201

WIDTH = 480
HEIGHT = 270
LR = 1e-3
EPOCHS = 1

MODEL_NAME = 'model.h5'
PREV_MODEL = ''

LOAD_MODEL = False

wl = 0
sl = 0
al = 0
dl = 0

wal = 0
wdl = 0
sal = 0
sdl = 0
nkl = 0

w = [1,0,0,0,0,0,0,0,0]
s = [0,1,0,0,0,0,0,0,0]
a = [0,0,1,0,0,0,0,0,0]
d = [0,0,0,1,0,0,0,0,0]
wa = [0,0,0,0,1,0,0,0,0]
wd = [0,0,0,0,0,1,0,0,0]
sa = [0,0,0,0,0,0,1,0,0]
sd = [0,0,0,0,0,0,0,1,0]
nk = [0,0,0,0,0,0,0,0,1]

model = InceptionV3(include_top=False, weights=None,input_shape=(WIDTH,HEIGHT,3), classes=9,classifier_activation='softmax')

if LOAD_MODEL:
    model.load(PREV_MODEL)
    print('We have loaded a previous model!!!!')


# iterates through the training files


for e in range(EPOCHS):
    #data_order = [i for i in range(1,FILE_I_END+1)]
    data_order = [i for i in range(1,FILE_I_END+1)]
    shuffle(data_order)
    for count,i in enumerate(data_order):

        try:
            file_name = 'training_data-{}.npy'.format(i)
            # full file info
            train_data = np.load(file_name,allow_pickle=True)
            print('training_data-{}.npy'.format(i),len(train_data))
            
            train = train_data[:-50]
            test = train_data[-50:]

            X = np.array([i[0] for i in train]).reshape(-1,WIDTH,HEIGHT,3)
            Y = np.array([i[1] for i in train])

            test_x = np.array([i[0] for i in test]).reshape(-1,WIDTH,HEIGHT,3)
            test_y = np.array([i[1] for i in test])

            learning_rate = 0.001
            opt = optimizers.Adam(lr=LR, decay=1e-5)

            model.compile(loss='categorical_crossentropy',
                          optimizer=opt,
                          metrics=['accuracy'])

            model.fit(x=X, y=Y, epochs=1, verbose=1, validation_data=(test_x,test_y), batch_size=None)


            if count%10 == 0:
                print('SAVING MODEL!')
                model.save(MODEL_NAME)

        except Exception as e:
            print(str(e))

我收到了这个错误：

    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\engine\training.py:806 train_function  *
        return step_function(self, iterator)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\engine\training.py:796 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:1211 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:2585 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:2945 _call_for_each_replica
        return fn(*args, **kwargs)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\engine\training.py:789 run_step  **
        outputs = model.train_step(data)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\engine\training.py:749 train_step
        y, y_pred, sample_weight, regularization_losses=self.losses)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\engine\compile_utils.py:204 __call__
        loss_value = loss_obj(y_t, y_p, sample_weight=sw)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\losses.py:149 __call__
        losses = ag_call(y_true, y_pred)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\losses.py:253 call  **
        return ag_fn(y_true, y_pred, **self._fn_kwargs)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\util\dispatch.py:201 wrapper
        return target(*args, **kwargs)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\losses.py:1535 categorical_crossentropy
        return K.categorical_crossentropy(y_true, y_pred, from_logits=from_logits)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\util\dispatch.py:201 wrapper
        return target(*args, **kwargs)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\keras\backend.py:4687 categorical_crossentropy
        target.shape.assert_is_compatible_with(output.shape)
    C:\Users\codingan\Anaconda3\envs\gta5\lib\site-packages\tensorflow\python\framework\tensor_shape.py:1134 assert_is_compatible_with
        raise ValueError("Shapes %s and %s are incompatible" % (self, other))

    ValueError: Shapes (None, 9) and (None, 13, 7, 2048) are incompatible

你有什么解决办法吗？

非常感谢你。

Answer 1

如果设置include_top=False，模型的输出将为4D（source）：

池化：当 include_top 为 False 时，用于特征提取的可选池化模式。 无（默认）表示模型的输出将是最后一个卷积块的 4D 张量输出。 avg 表示全局平均池化将应用于最后一个卷积块的输出，因此模型的输出将是一个 2D 张量。 max 表示将应用全局最大池化。

您指定了类的数量，但 classes 仅在 include_top=True 时才有效：

classes：可选的图像分类数，仅在 include_top 为 True 且未指定 weights 参数时指定。默认为 1000。

classifier_activation 也一样：

classifier_activation：str 或可调用。在“顶层”层上使用的激活函数。除非 include_top=True，否则忽略。设置 classifier_activation=None 以返回“顶层”层的 logits。

tl;dr如果你设置include_top=True应该没问题