在我的训练和验证数据集上运行model.fit()方法,我不断收到此错误,而且我不确定这是怎么回事:
ValueError: Could not find matching function to call loaded from the SavedModel. Got:
Positional arguments (4 total):
* Tensor("inputs:0", shape=(224, 224, 3), dtype=float32)
* False
* False
* 0.99
Keyword arguments: {}
Expected these arguments to match one of the following 4 option(s):
hist = model.fit(x=EXAMPLES_TRAIN,
steps_per_epoch=steps_per_epoch,
validation_data=EXAMPLES_VALID,
validation_steps=validation_steps,
epochs=_NUM_EPOCHS,
callbacks=
tf.keras.callbacks.TensorBoard(s_dir), # log metrics
hp.KerasCallback(s_dir, s_hparams), # log hparams
EarlyStopping(monitor='val_loss', mode='min', baseline=_CHKPT_ES_BASELINE, patience=_CHKPT_ES_PATIENCE, verbose=1), # early stopping],
verbose=1).history
我不介意分享更多细节以提供有关我的问题的更多见解。关于我可能做错了什么的任何想法,因为我有相似的代码,在什么地方可以很好地工作,而不会出现此错误。
答案 0 :(得分:0)
在加载权重之前尝试对模型进行虚拟调用:
model = build_model(...) # Your model-building code
model(tf.zeros((224, 224, 3), tf.float32)
# load model weights