我正在使用colabs教程中的这示例对模型进行微调,经过培训后,我希望使用以下方法保存模型并在本地计算机上加载:
ckpt_manager = tf.train.CheckpointManager(ckpt, directory="test_data/checkpoint/", max_to_keep=5)
...
...
print('Done fine-tuning!')
ckpt_manager.save()
print('Checkpoint saved!')但是在我的本地计算机上恢复之后,使用检查点文件不会检测到任何对象(分数太低)。
我也试过了
tf.saved_model.save(detection_model, '/content/new_model/')然后装上这个:
detection_model = tf.saved_model.load('/saved_model_20201226/')
input_tensor = tf.convert_to_tensor(image, dtype=tf.float32)
detections = detection_model(input_tensor)告诉我以下错误: TypeError:'_UserObject‘对象不可调用
保存和加载优化模型的正确方法是什么?
编辑1: --它正在等待保存新的管道配置,在那之后终于起作用了!这是我的回答:
# Save new pipeline config
new_pipeline_proto = config_util.create_pipeline_proto_from_configs(configs)
config_util.save_pipeline_config(new_pipeline_proto, '/content/new_config')
exported_ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)
ckpt_manager = tf.train.CheckpointManager(
exported_ckpt, directory="test_data/checkpoint/", max_to_keep=5)
...
...
print('Done fine-tuning!')
ckpt_manager.save()
print('Checkpoint saved!')发布于 2020-12-27 08:55:58
它正在等待保存新的管道配置,在此之后,终于起作用了!这是我的回答:
# Save new pipeline config
new_pipeline_proto = config_util.create_pipeline_proto_from_configs(configs)
config_util.save_pipeline_config(new_pipeline_proto, '/content/new_config')
exported_ckpt = tf.compat.v2.train.Checkpoint(model=detection_model)
ckpt_manager = tf.train.CheckpointManager(
exported_ckpt, directory="test_data/checkpoint/", max_to_keep=5)
...
...
print('Done fine-tuning!')
ckpt_manager.save()
print('Checkpoint saved!')https://stackoverflow.com/questions/65463236
复制相似问题