一个用于查找语言类型的分类器，印地语或英语。
ValueError: Tensor("ExponentialDecay_4:0", shape=(), dtype=float32) must be from the same graph as Tensor(("dnn/hiddenlayer_0/kernel/part_0:0", shape=(), dtype=resource)).
from __future__ import absolute_import, division, print_function, unicode_literals
from absl import logging
import tensorflow as tf
import tensorflow_hub as hub
import matplotlib.pyplot as plt
import numpy as np
import os
import pandas as pd
import re
import seaborn as sns
df = pd.read_csv("intern_test/data/english_text.csv", encoding="latin-1")
df2 = pd.read_csv("intern_test/data/hinglish_text.csv", encoding="latin-1")
df['label'] = 0 # English
df2['label'] = 1 # Hindi
df3 = pd.concat([df,df2])
df3.head()
# Training input on the whole training set with no limit on training epochs.
train_input_fn = tf.compat.v1.estimator.inputs.pandas_input_fn(
df3, df3["label"], num_epochs=None, shuffle=True)
# Prediction on the whole training set.
predict_train_input_fn = tf.compat.v1.estimator.inputs.pandas_input_fn(
df3, df3["label"], shuffle=False)
# Prediction on the test set.
#predict_test_input_fn = tf.compat.v1.estimator.inputs.pandas_input_fn(
# X_test, y_test, shuffle=False)
embedded_text_feature_column = hub.text_embedding_column(
key="text",
module_spec="https://tfhub.dev/google/nnlm-en-dim128/1")
optimizer = tf.compat.v1.train.AdamOptimizer(
learning_rate=tf.compat.v1.train.exponential_decay(
global_step=0,
learning_rate=0.1,
decay_steps=10000,
decay_rate=0.96, staircase=True))
estimator = tf.estimator.DNNClassifier(
hidden_units=[500, 100],
feature_columns=[embedded_text_feature_column],
n_classes=2,
optimizer=optimizer
)
estimator.train(input_fn=train_input_fn, steps=5000);
完全回溯。
ValueError Traceback (most recent call last)
<ipython-input-47-1c4563a14246> in <module>
2 # batch size. This is roughly equivalent to 25 epochs since the training dataset
3 # contains 25,000 examples.
----> 4 estimator.train(input_fn=train_input_fn, steps=5000);
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/estimator.py in train(self, input_fn, hooks, steps, max_steps, saving_listeners)
356
357 saving_listeners = _check_listeners_type(saving_listeners)
--> 358 loss = self._train_model(input_fn, hooks, saving_listeners)
359 logging.info('Loss for final step: %s.', loss)
360 return self
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/estimator.py in _train_model(self, input_fn, hooks, saving_listeners)
1122 return self._train_model_distributed(input_fn, hooks, saving_listeners)
1123 else:
-> 1124 return self._train_model_default(input_fn, hooks, saving_listeners)
1125
1126 def _train_model_default(self, input_fn, hooks, saving_listeners):
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/estimator.py in _train_model_default(self, input_fn, hooks, saving_listeners)
1152 worker_hooks.extend(input_hooks)
1153 estimator_spec = self._call_model_fn(
-> 1154 features, labels, model_fn_lib.ModeKeys.TRAIN, self.config)
1155 global_step_tensor = training_util.get_global_step(g)
1156 return self._train_with_estimator_spec(estimator_spec, worker_hooks,
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/estimator.py in _call_model_fn(self, features, labels, mode, config)
1110
1111 logging.info('Calling model_fn.')
-> 1112 model_fn_results = self._model_fn(features=features, **kwargs)
1113 logging.info('Done calling model_fn.')
1114
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/canned/dnn.py in _model_fn(features, labels, mode, config)
520 input_layer_partitioner=input_layer_partitioner,
521 config=config,
--> 522 batch_norm=batch_norm)
523
524 super(DNNClassifier, self).__init__(
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/canned/dnn.py in _dnn_model_fn(features, labels, mode, head, hidden_units, feature_columns, optimizer, activation_fn, dropout, input_layer_partitioner, config, use_tpu, batch_norm)
300 labels=labels,
301 optimizer=optimizer,
--> 302 logits=logits)
303
304
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/canned/head.py in create_estimator_spec(self, features, mode, logits, labels, optimizer, train_op_fn, regularization_losses)
238 self._create_tpu_estimator_spec(
239 features, mode, logits, labels, optimizer, train_op_fn,
--> 240 regularization_losses))
241 return tpu_estimator_spec.as_estimator_spec()
242 except NotImplementedError:
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow_estimator/python/estimator/canned/head.py in _create_tpu_estimator_spec(self, features, mode, logits, labels, optimizer, train_op_fn, regularization_losses)
1244 train_op = optimizer.minimize(
1245 regularized_training_loss,
-> 1246 global_step=training_util.get_global_step())
1247 elif train_op_fn is not None:
1248 train_op = train_op_fn(regularized_training_loss)
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/training/optimizer.py in minimize(self, loss, global_step, var_list, gate_gradients, aggregation_method, colocate_gradients_with_ops, name, grad_loss)
411
412 return self.apply_gradients(grads_and_vars, global_step=global_step,
--> 413 name=name)
414
415 def compute_gradients(self, loss, var_list=None,
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/training/optimizer.py in apply_gradients(self, grads_and_vars, global_step, name)
610 scope_name = var.op.name
611 with ops.name_scope("update_" + scope_name), ops.colocate_with(var):
--> 612 update_ops.append(processor.update_op(self, grad))
613 if global_step is None:
614 apply_updates = self._finish(update_ops, name)
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/training/optimizer.py in update_op(self, optimizer, g)
169 return optimizer._resource_apply_sparse_duplicate_indices(
170 g.values, self._v, g.indices)
--> 171 update_op = optimizer._resource_apply_dense(g, self._v)
172 if self._v.constraint is not None:
173 with ops.control_dependencies([update_op]):
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/training/adam.py in _resource_apply_dense(self, grad, var)
173 math_ops.cast(self._beta2_t, grad.dtype.base_dtype),
174 math_ops.cast(self._epsilon_t, grad.dtype.base_dtype),
--> 175 grad, use_locking=self._use_locking)
176
177 def _apply_sparse_shared(self, grad, var, indices, scatter_add):
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/training/gen_training_ops.py in resource_apply_adam(var, m, v, beta1_power, beta2_power, lr, beta1, beta2, epsilon, grad, use_locking, use_nesterov, name)
1300 beta2=beta2, epsilon=epsilon, grad=grad,
1301 use_locking=use_locking,
-> 1302 use_nesterov=use_nesterov, name=name)
1303 return _op
1304 _result = None
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py in _apply_op_helper(self, op_type_name, name, **keywords)
348 # Need to flatten all the arguments into a list.
349 # pylint: disable=protected-access
--> 350 g = ops._get_graph_from_inputs(_Flatten(keywords.values()))
351 # pylint: enable=protected-access
352 except AssertionError as e:
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/framework/ops.py in _get_graph_from_inputs(op_input_list, graph)
5711 graph = graph_element.graph
5712 elif original_graph_element is not None:
-> 5713 _assert_same_graph(original_graph_element, graph_element)
5714 elif graph_element.graph is not graph:
5715 raise ValueError("%s is not from the passed-in graph." % graph_element)
~/anaconda3/envs/tf/lib/python3.7/site-packages/tensorflow/python/framework/ops.py in _assert_same_graph(original_item, item)
5647 if original_item.graph is not item.graph:
5648 raise ValueError("%s must be from the same graph as %s." % (item,
-> 5649 original_item))
5650
5651
ValueError: Tensor("ExponentialDecay_4:0", shape=(), dtype=float32) must be from the same graph as Tensor("dnn/hiddenlayer_0/kernel/part_0:0", shape=(), dtype=resource).
你可以尝试使用函数给出指数衰减，该函数将学习率和步数作为输入，并将其作为 keras.callbacks.LearningRateScheduler 的输入。
这里是一个使用指数衰减的例子。
def exponential_decay_fn(epoch):
return 0.01 * 0.1**(epoch / 20)
def exponential_decay(lr0, s):
def exponential_decay_fn(epoch):
return lr0 * 0.1**(epoch / s)
return exponential_decay_fn
exponential_decay_fn = exponential_decay(lr0=0.01, s=20)
model.compile(loss="sparse_categorical_crossentropy", optimizer="nadam", metrics=["accuracy"])
n_epochs = 25
lr_scheduler = keras.callbacks.LearningRateScheduler(exponential_decay_fn)
history = model.fit(X_train_scaled, y_train, epochs=n_epochs,
validation_data=(X_valid_scaled, y_valid),
callbacks=[lr_scheduler])
我希望这能回答你的问题。