当将tf.estimator
与warm_start_from
和 model_dir
一起使用时,warm_start_from
目录和model_dir
目录都包含有效的检查点,实际上将还原哪个检查点?
为了提供一些背景信息,我的估算器代码如下所示
est = tf.estimator.Estimator(
model_fn=model_fn,
model_dir=model_dir,
warm_start_from=warm_start_dir)
for epoch in range(num_epochs):
est.train(input_fn=train_input_fn)
est.evaluate(input_fn=eval_input_fn)
(输入函数使用一击迭代器。)
因此,在第一次迭代中,当
model_dir
为空时,我希望加载热启动检查点,但是在下一个时期,我希望从model_dir
的上一次迭代中获取中间经过微调的检查点。但是至少从日志中看来,warm_start_dir
仍在加载。我可能会在下一次迭代中覆盖我的估算器,但我想知道是否应该以某种方式在估算器中构建它。
最佳答案
我有一个类似的问题,我通过提供一个在 session 开始时运行的初始化 Hook 并使用了tf.estimator.train_and_evaluate
解决了这一问题(尽管我无法为整个解决方案效劳,因为我看到了另一个解决方案的相似之处)其他目的):
class InitHook(tf.train.SessionRunHook):
"""initializes model from a checkpoint_path
args:
modelPath: full path to checkpoint
"""
def __init__(self, checkpoint_dir):
self.modelPath = checkpoint_dir
self.initialized = False
def begin(self):
"""
Restore encoder parameters if a pre-trained encoder model is available and we haven't trained previously
"""
if not self.initialized:
log = logging.getLogger('tensorflow')
checkpoint = tf.train.latest_checkpoint(self.modelPath)
if checkpoint is None:
log.info('No pre-trained model is available, training from scratch.')
else:
log.info('Pre-trained model {0} found in {1} - warmstarting.'.format(checkpoint, self.modelPath))
tf.train.warm_start(checkpoint)
self.initialized = True
然后,进行培训:
initHook = InitHook(checkpoint_dir = warm_start_dir)
trainSpec = tf.estimator.TrainSpec(
input_fn = train_input_fn,
max_steps = N_STEPS,
hooks = [initHook]
)
evalSpec = tf.estimator.EvalSpec(
input_fn = eval_input_fn,
steps = None,
name = 'eval',
throttle_secs = 3600
)
tf.estimator.train_and_evaluate(estimator, trainSpec, evalSpec)
它从头开始运行一次,以从
warm_start_dir
初始化变量。稍后,当估计器model_dir
中有新的检查点时,它将从此处继续warm_start。关于Tensorflow估算器-warm_start_from和model_dir,我们在Stack Overflow上找到一个类似的问题:https://stackoverflow.com/questions/49846207/