我在训练ML模型时出错

Question

我试图运行以下代码来训练模型:

history = model.fit(train_ds, 
steps_per_epoch=train_len/20,
validation_data=val_ds,
validation_steps=val_len/20,
epochs=50,
verbose=1,
callbacks=[cp_callback]
)

但是得到错误:

TypeError: Cannot convert 0.8 to EagerTensor of dtype int64

我无法训练模型。

Answer 1

尝试使用int函数映射step_per_epoch和validation_steps:

int(steps_per_epoch=train_len/20)
int(validation_steps=val_len/20)