为了实现finetune有如下两种解决方案：
model_fn里面定义好模型之后直接赋值
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
|
def model_fn(features, labels, mode, params):# .....# finetuneif params.checkpoint_path and (not tf.train.latest_checkpoint(params.model_dir)):checkpoint_path = Noneif tf.gfile.IsDirectory(params.checkpoint_path): checkpoint_path = tf.train.latest_checkpoint(params.checkpoint_path)else: checkpoint_path = params.checkpoint_pathtf.train.init_from_checkpoint( ckpt_dir_or_file=checkpoint_path, assignment_map={params.checkpoint_scope: params.checkpoint_scope} # 'OptimizeLoss/':'OptimizeLoss/') |
使用钩子 hooks。
可以在定义tf.contrib.learn.Experiment的时候通过train_monitors参数指定
|
1
2
3
4
5
6
7
8
9
10
11
|
# Define the experimentexperiment = tf.contrib.learn.Experiment(estimator=estimator, # Estimatortrain_input_fn=train_input_fn, # First-class functioneval_input_fn=eval_input_fn, # First-class functiontrain_steps=params.train_steps, # Minibatch stepsmin_eval_frequency=params.eval_min_frequency, # Eval frequency# train_monitors=[], # Hooks for training# eval_hooks=[eval_input_hook], # Hooks for evaluationeval_steps=params.eval_steps # Use evaluation feeder until its empty) |
也可以在定义tf.estimator.EstimatorSpec 的时候通过training_chief_hooks参数指定。
不过个人觉得最好还是在estimator中定义，让experiment只专注于控制实验的模式（训练次数,验证次数等等）。
|
1
2
3
4
5
6
7
8
9
10
11
12
13
|
def model_fn(features, labels, mode, params): # .... return tf.estimator.EstimatorSpec( mode=mode, predictions=predictions, loss=loss, train_op=train_op, eval_metric_ops=eval_metric_ops, # scaffold=get_scaffold(), # training_chief_hooks=None ) |
这里顺便解释以下tf.estimator.EstimatorSpec对像的作用。该对象描述来一个模型的方方面面。包括：
当前的模式：
mode: A ModeKeys. Specifies if this is training, evaluation or prediction.
计算图
predictions: Predictions Tensor or dict of Tensor.
loss: Training loss Tensor. Must be either scalar, or with shape [1].
train_op: Op for the training step.
eval_metric_ops: Dict of metric results keyed by name. The values of the dict are the results of calling a metric function, namely a (metric_tensor, update_op) tuple. metric_tensor should be evaluated without any impact on state (typically is a pure computation results based on variables.). For example, it should not trigger the update_op or requires any input fetching.
导出策略
export_outputs: Describes the output signatures to be exported to
SavedModel and used during serving. A dict {name: output} where:
name: An arbitrary name for this output.
output: an ExportOutput object such as ClassificationOutput, RegressionOutput, or PredictOutput. Single-headed models only need to specify one entry in this dictionary. Multi-headed models should specify one entry for each head, one of which must be named using signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY.
chief钩子 训练时的模型保存策略钩子CheckpointSaverHook， 模型恢复等
training_chief_hooks: Iterable of tf.train.SessionRunHook objects to run on the chief worker during training.
worker钩子 训练时的监控策略钩子如： NanTensorHook LoggingTensorHook 等
training_hooks: Iterable of tf.train.SessionRunHook objects to run on all workers during training.
指定初始化和saver
scaffold: A tf.train.Scaffold object that can be used to set initialization, saver, and more to be used in training.
evaluation钩子
evaluation_hooks: Iterable of tf.train.SessionRunHook objects to run during evaluation.
自定义的钩子如下：
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
|
class RestoreCheckpointHook(tf.train.SessionRunHook): def __init__(self, checkpoint_path, exclude_scope_patterns, include_scope_patterns ): tf.logging.info("Create RestoreCheckpointHook.") #super(IteratorInitializerHook, self).__init__() self.checkpoint_path = checkpoint_path self.exclude_scope_patterns = None if (not exclude_scope_patterns) else exclude_scope_patterns.split(',') self.include_scope_patterns = None if (not include_scope_patterns) else include_scope_patterns.split(',') def begin(self): # You can add ops to the graph here. print('Before starting the session.') # 1. Create saver #exclusions = [] #if self.checkpoint_exclude_scopes: # exclusions = [scope.strip() # for scope in self.checkpoint_exclude_scopes.split(',')] # #variables_to_restore = [] #for var in slim.get_model_variables(): #tf.global_variables(): # excluded = False # for exclusion in exclusions: # if var.op.name.startswith(exclusion): # excluded = True # break # if not excluded: # variables_to_restore.append(var) #inclusions #[var for var in tf.trainable_variables() if var.op.name.startswith('InceptionResnetV1')] variables_to_restore = tf.contrib.framework.filter_variables( slim.get_model_variables(), include_patterns=self.include_scope_patterns, # ['Conv'], exclude_patterns=self.exclude_scope_patterns, # ['biases', 'Logits'], # If True (default), performs re.search to find matches # (i.e. pattern can match any substring of the variable name). # If False, performs re.match (i.e. regexp should match from the beginning of the variable name). reg_search = True ) self.saver = tf.train.Saver(variables_to_restore) def after_create_session(self, session, coord): # When this is called, the graph is finalized and # ops can no longer be added to the graph. print('Session created.') tf.logging.info('Fine-tuning from %s' % self.checkpoint_path) self.saver.restore(session, os.path.expanduser(self.checkpoint_path)) tf.logging.info('End fineturn from %s' % self.checkpoint_path) def before_run(self, run_context): #print('Before calling session.run().') return None #SessionRunArgs(self.your_tensor) def after_run(self, run_context, run_values): #print('Done running one step. The value of my tensor: %s', run_values.results) #if you-need-to-stop-loop: # run_context.request_stop() pass def end(self, session): #print('Done with the session.') pass |
以上这篇tensorflow estimator 使用hook实现finetune方式就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多支持服务器之家。
原文链接：https://blog.csdn.net/andylei777/article/details/79074757
