在Flask应用中加载Keras模型时遇tree模块无flatten属性错误

问题详情 #

在Flask应用中加载Keras模型时触发错误：

AttributeError: module 'tree' has no attribute 'flatten'

使用的TensorFlow 2版本应为2.16.1（原描述的12.16.1为笔误，TensorFlow 2系列版本号以2.x开头），加载模型的核心代码如下：

from flask import Flask, render_template, request
from tensorflow.keras.models import load_model
import numpy as np
from tensorflow.keras.preprocessing.image import load_img
from tensorflow.keras.applications.vgg16 import preprocess_input
import os
from tensorflow.keras.preprocessing import image

app = Flask(__name__)
model = load_model('mymodel.h5')

完整报错追踪：

Traceback (most recent call last):
File "C:\Users\ALTARON\Desktop\Project\model\app.py", line 10, in
model = load_model('mymodel.h5')

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\saving\saving_api.py", line 183, in load_model
return legacy_h5_format.load_model_from_hdf5(filepath)

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\legacy\saving\legacy_h5_format.py", line 133, in load_model_from_hdf5
model = saving_utils.model_from_config(

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\legacy\saving\saving_utils.py", line 85, in model_from_config
return serialization.deserialize_keras_object(

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\legacy\saving\serialization.py", line 495, in deserialize_keras_object
deserialized_obj = cls.from_config(

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\models\sequential.py", line 322, in from_config
model = cls(name=name)

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\models\model.py", line 144, in new
return super().new(cls)

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\layers\layer.py", line 217, in new
obj = super().new(cls, *args, **kwargs)

File "C:\Users\ALTARON\AppData\Local\Programs\Python\Python311\Lib\site-packages\keras\src\ops\operation.py", line 100, in new
flat_arg_values = tree.flatten(kwargs)

AttributeError: module 'tree' has no attribute 'flatten'

模型结构摘要：

Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 conv2d (Conv2D)             (None, 256, 256, 32)      896       
                                                                 
 max_pooling2d (MaxPooling2D) (None, 85, 85, 32)       0         
                                                                 
 conv2d_1 (Conv2D)           (None, 85, 85, 64)        18496     
                                                                 
 max_pooling2d_1 (MaxPooling2D) (None, 28, 28, 64)     0         
                                                                 
 conv2d_2 (Conv2D)           (None, 28, 28, 128)       73856     
                                                                 
 conv2d_3 (Conv2D)           (None, 28, 28, 128)       147584    
                                                                 
 max_pooling2d_2 (MaxPooling2D) (None, 9, 9, 128)      0         
                                                                 
 flatten (Flatten)           (None, 10368)             0         
                                                                 
 dense (Dense)               (None, 200)               2073800   
                                                                 
 dropout (Dropout)           (None, 200)               0         
                                                                 
 dense_1 (Dense)             (None, 38)                7638      
                                                                 
=================================================================
Total params: 2322270 (8.86 MB)
Trainable params: 2322270 (8.86 MB)
Non-trainable params: 0 (0.00 Byte)

解决方案 #

该错误源于第三方tree包与TensorFlow/Keras依赖的dm-tree模块冲突，或dm-tree未正确安装/版本不兼容，以下是具体修复方案：

方案1：替换冲突模块 #

卸载用于目录遍历的第三方tree包，安装Keras所需的dm-tree：

pip uninstall -y tree
pip install dm-tree

方案2：临时强制导入正确模块 #

若无法卸载tree包，可在加载模型前覆盖全局tree引用：

import sys
import tree as dm_tree
sys.modules['tree'] = dm_tree

# 后续正常导入并加载模型
from tensorflow.keras.models import load_model
model = load_model('mymodel.h5')

方案3：降级TensorFlow版本 #

若前两种方法无效，降级到依赖逻辑更稳定的TensorFlow 2.15.x版本：

pip install tensorflow==2.15.0

验证 #

完成修复后启动Flask应用，确认模型能正常加载，再测试图片预测流程，确保无报错。

内容的提问来源于stack exchange，提问作者Atul Anand