我刚刚想到了一个实现这个的方法。

from collections import Counter
train_datagen = ImageDataGenerator()
train_generator = train_datagen.flow_from_directory(...)

counter = Counter(train_generator.classes)                          
max_val = float(max(counter.values()))       
class_weights = {class_id : max_val/num_images for class_id, num_images in counter.items()}                     

model.fit_generator(...,
                    class_weight=class_weights)

train_generator.classes是每个图像的类别列表。Counter(train_generator.classes)创建每个类别中图像数量的计数器。
请注意，这些权重可能不利于收敛，但您可以将其用作基于出现次数的其他类型权重的基础。

• 这个答案的灵感来自：https://github.com/fchollet/keras/issues/1875#issuecomment-273752868*

或者，您可以简单地执行以下操作：

from sklearn.utils import class_weight
import numpy as np

class_weights = class_weight.compute_class_weight(
               'balanced',
                np.unique(train_generator.classes), 
                train_generator.classes)

然后，您可以设置（按照上面的注解）：

model.fit_generator(..., class_weight=class_weights)

我尝试了这两种解决方案，sklearn.utils.class_weight的准确性更好，虽然我不知道为什么，但它们并不都产生相同的类权重。

正如文章here中所建议的，分配类权重的一个好方法是用途：

(1 / class_count) * (total_count/2)

因此，稍微修改Fábio Perez上面建议的方法：

counter = Counter(train_generator.classes)
total = float(sum(counter.values()))
class_weight = {class_id : (1/num_images)*(total)/2.0 for class_id, num_images in counter.items()}

Pasha Dembo建议的代码运行得很好。但是，在插入model_fit生成器之前，您应该在字典中对其进行转换：

from sklearn.utils import class_weight import numpy as np

class_weights = class_weight.compute_class_weight(
           'balanced',
            np.unique(train_generator.classes), 
            train_generator.classes)

train_class_weights = dict(enumerate(class_weights))
model.fit_generator(..., class_weight=train_class_weights)

或者，您可以简单地执行以下操作：

from sklearn.utils import class_weight import numpy as np
 
 class_weights = class_weight.compute_class_weight(
                'balanced',
                 np.unique(train_generator.classes), 
                 train_generator.classes) You can then set (as per comment above):
 
 model.fit_generator(..., class_weight=class_weights)

from sklearn.utils import class_weight
import numpy as np
class_weights = dict(zip(np.unique(traingen.classes),class_weight.compute_class_weight(
                        class_weight = 'balanced',
                        classes = np.unique(traingen.classes), 
                        y = traingen.classes)))

April 2023 version.结束使用此：

from sklearn.utils.class_weight import compute_class_weight

unique_classes = np.unique(ds_train.classes)
# "If ‘balanced’, class weights will be given by n_samples / (n_classes * np.bincount(y))."
class_weights = compute_class_weight("balanced", classes=unique_classes, y=ds_train.classes)
class_weight = {class_id: weight for class_id, weight in zip(unique_classes, class_weights)}

model.fit(..., class_weight=class_weight)