TensorFlow PR Curve Plugin：pr_curve_streaming_op

Question

マルチクラス分類にpr_curve_streaming_opを使用する方法を知りましたか？

私はthis demoとthis demoに従おうとしています。しかし、これまでのところ、私はデータ収集段階にぶつかりつつあります。私は実行時にデータを収集することができるようにする必要があり、実行時などにboolテンソルを変更できないなどのいくつかの問題に直面しているようです。どんなヘルプも非常に便利です。おかげ

は、ここに私のコードスニペットであると私は取得していますエラー：

self.input_y = tf.placeholder(tf.float32, [None, num_classes], name='input_y') 

with tf.device('/cpu:0'), tf.name_scope('embedding'): 
    self.emb_var = tf.Variable(embedding_mat, name='emb_var') 
    # if not non_static: 
    # self.emb_var = tf.constant(embedding_mat, name='emb_var') 
    # else: 
    # self.emb_var = tf.Variable(embedding_mat, name='emb_var', trainable=True) 
    self.embedded_chars = tf.nn.embedding_lookup(self.emb_var, self.input_x) 
    self.emb = tf.expand_dims(self.embedded_chars, -1) 
pooled_concat = [] 
reduced = np.int32(np.ceil((sequence_length) * 1.0/max_pool_size)) 

### 
. 
. 
Removed some code for simplicity... 
. 
. 
### 
with tf.name_scope('probabilities'): 
    self.probabilities = tf.nn.softmax(self.scores) 

with tf.name_scope('confidence'): 
    self.conf = tf.reduce_max(self.probabilities, reduction_indices=[1]) 

with tf.name_scope('Low_confidence'): 
    self.conf_low = tf.reduce_min(self.conf, name='low_conf') 

with tf.name_scope('Avg_confidence'): 
    self.Avg_conf = tf.reduce_mean(self.conf, name='confidence') 

with tf.name_scope('high_confidence'): 
    self.conf_high = tf.reduce_max(self.conf, name='high_conf') 

with tf.name_scope('loss'): 
    losses = tf.nn.softmax_cross_entropy_with_logits(labels=self.input_y, logits=self.scores) # only named arguments accepted 
    self.loss = tf.reduce_mean(losses) + l2_reg_lambda * l2_loss 

with tf.name_scope('accuracy'): 
    correct_predictions = tf.equal(self.predictions, tf.argmax(self.input_y, 1)) 
    self.accuracy = tf.reduce_mean(tf.cast(correct_predictions, "float"), name='accuracy') 

with tf.name_scope('num_correct'): 
    correct = tf.equal(self.predictions, tf.argmax(self.input_y, 1)) 
    self.num_correct = tf.reduce_sum(tf.cast(correct, 'float')) 

''' start ''' 
with tf.name_scope('pr_curve'): 
    self.pr_curve, update_op = summary_lib.pr_curve_streaming_op('pr_curve', predictions=self.probabilities, labels=tf.cast(self.input_y,tf.bool), num_thresholds=11) 

''' end ''' 

をここでは、私が取得していますエラーです：

2017-12-25 18:37:30.433250: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433565: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433613: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433630: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433652: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433699: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433729: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433767: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
2017-12-25 18:37:30.433805: W tensorflow/core/framework/op_kernel.cc:1192] Failed precondition: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
Traceback (most recent call last): 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1323, in _do_call 
    return fn(*args) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1302, in _run_fn 
    status, run_metadata) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py", line 473, in __exit__ 
    c_api.TF_GetCode(self.status.status)) 
tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
     [[Node: rnn/LessEqual_4/_37 = _HostRecv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_105_rnn/LessEqual_4", tensor_type=DT_BOOL, _device="/job:localhost/replica:0/task:0/device:CPU:0"]()]] 

During handling of the above exception, another exception occurred: 

Traceback (most recent call last): 
    File "./code/train.py", line 357, in <module> 
    train_cnn_rnn() 
    File "./code/train.py", line 276, in train_cnn_rnn 
    train_step(x_train_batch, y_train_batch) 
    File "./code/train.py", line 223, in train_step 
    _, step, summaries = sess.run([train_op, global_step, train_summary_op], feed_dict) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/debug/wrappers/framework.py", line 534, in run 
    run_metadata=run_metadata) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 889, in run 
    run_metadata_ptr) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1120, in _run 
    feed_dict_tensor, options, run_metadata) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1317, in _do_run 
    options, run_metadata) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1336, in _do_call 
    raise type(e)(node_def, op, message) 
tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
     [[Node: rnn/LessEqual_4/_37 = _HostRecv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_105_rnn/LessEqual_4", tensor_type=DT_BOOL, _device="/job:localhost/replica:0/task:0/device:CPU:0"]()]] 

Caused by op 'pr_curve/pr_curve/true_negatives/true_negatives/read', defined at: 
    File "./code/train.py", line 357, in <module> 
    train_cnn_rnn() 
    File "./code/train.py", line 170, in train_cnn_rnn 
    l2_reg_lambda=params['l2_reg_lambda']) 
    File "/media/hemant/MVV/MyValueVest-local/main/code/text_cnn_rnn.py", line 127, in __init__ 
    self.pr_curve, update_op = summary_lib.pr_curve_streaming_op('pr_curve', predictions=self.probabilities, labels=tf.cast(self.input_y,tf.bool), num_thresholds=11) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorboard/plugins/pr_curve/summary.py", line 234, in streaming_op 
    weights=weights) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/metrics_impl.py", line 1482, in true_negatives_at_thresholds 
    labels, predictions, thresholds, weights=weights, includes=('tn',)) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/metrics_impl.py", line 543, in _confusion_matrix_at_thresholds 
    true_n = _create_local('true_negatives', shape=[num_thresholds]) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/metrics_impl.py", line 196, in _create_local 
    validate_shape=validate_shape) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/variable_scope.py", line 1927, in variable 
    caching_device=caching_device, name=name, dtype=dtype) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/variables.py", line 213, in __init__ 
    constraint=constraint) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/variables.py", line 356, in _init_from_args 
    self._snapshot = array_ops.identity(self._variable, name="read") 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/array_ops.py", line 125, in identity 
    return gen_array_ops.identity(input, name=name) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/ops/gen_array_ops.py", line 2071, in identity 
    "Identity", input=input, name=name) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper 
    op_def=op_def) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 2956, in create_op 
    op_def=op_def) 
    File "/home/hemant/anaconda3/envs/tf14/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 1470, in __init__ 
    self._traceback = self._graph._extract_stack() # pylint: disable=protected-access 

FailedPreconditionError (see above for traceback): Attempting to use uninitialized value pr_curve/pr_curve/true_negatives/true_negatives 
     [[Node: pr_curve/pr_curve/true_negatives/true_negatives/read = Identity[T=DT_FLOAT, _class=["loc:@pr_curve/pr_curve/true_negatives/true_negatives"], _device="/job:localhost/replica:0/task:0/device:GPU:0"](pr_curve/pr_curve/true_negatives/true_negatives)]] 
     [[Node: rnn/LessEqual_4/_37 = _HostRecv[client_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_105_rnn/LessEqual_4", tensor_type=DT_BOOL, _device="/job:localhost/replica:0/task:0/device:CPU:0"]()]] 

Answer 1

はそれを考え出した... - 別々に作成されました私はすべての予測値を必要としていたので - すべてのクラスをループしました。

pr_graph = tf.Graph() 
with pr_graph.as_default(): 
    session_conf = tf.ConfigProto(allow_soft_placement=True, log_device_placement=False) 
    pr_sess = tf.Session(config=session_conf) 
    with pr_sess.as_default(): 
     for cat in range(y_train.shape[1]): 
      with tf.name_scope('%s' % labels[cat]): 
       _, update_op = summary_lib.pr_curve_streaming_op('pr_curve', predictions=probs[:, cat], labels=tf.cast(y_test[:, cat], tf.bool), num_thresholds=500, metrics_collections='pr') 
     pr_summary_op = tf.summary.merge_all() 
     pr_summary_dir = os.path.join(checkpoint_dir, "s", "pr") 
     pr_summary_writer = tf.summary.FileWriter(pr_summary_dir, pr_sess.graph) 
     pr_sess.run(tf.local_variables_initializer()) 
     pr_sess.run([update_op]) 
     pr_summary_writer.add_summary(pr_sess.run(pr_summary_op)) 
     pr_summary_writer.close()