# ==================================================
print("Entering into graph\n")
with tf.Graph().as_default():
session_conf = tf.ConfigProto(
allow_soft_placement=FLAGS.allow_soft_placement,
log_device_placement=FLAGS.log_device_placement)
sess = tf.Session(config=session_conf)
print("Entering into session\n")
with sess.as_default():
model = Model(
num_classes=y_train.shape[1],
vocab_size=len(vocab_processor.vocabulary_),
embedding_size = embeddings.shape[1],
max_length = FLAGS.max_doc_length,
vocab_proc = vocab_processor,
filter_sizes=list(map(int, FLAGS.filter_sizes.split(","))),
num_filters=FLAGS.num_filters,
l2_reg_lambda=FLAGS.l2_reg_lambda)
print("Defined Model\n")
# Define Training procedure
global_step = tf.Variable(0, name="global_step", trainable=False)
optimizer = tf.train.AdamOptimizer(FLAGS.learning_rate)
grads_and_vars = optimizer.compute_gradients(model.loss) # ,tf.trainable_variables()
def ClipIfNotNone(grad):
if grad is None:
print("NAN----------------------------------------------------")
return grad
# Training
# ==================================================
print("Entering into graph\n")
with tf.Graph().as_default():
session_conf = tf.ConfigProto(
allow_soft_placement=FLAGS.allow_soft_placement,
log_device_placement=FLAGS.log_device_placement)
sess = tf.Session(config=session_conf)
print("Entering into session\n")
with sess.as_default():
model = Model(
#sequence_length=q_train.shape[1],
num_classes=y_train.shape[1],
#vocab_size=len(vocab_processor.vocabulary_),
embedding_size=q_train.shape[1],
filter_sizes=list(map(int, FLAGS.filter_sizes.split(","))),
num_filters=FLAGS.num_filters,
l2_reg_lambda=FLAGS.l2_reg_lambda)
# Output directory for models and summaries
timestamp = str(int(time.time()))
out_dir = os.path.abspath(os.path.join("../../runs", timestamp)) #os.path.curdir,"runs", timestamp
print("Writing to {}\n".format(out_dir))
print("Defined Model\n")
# Define Training procedure
global_step = tf.Variable(0, name="global_step", trainable=False)
optimizer = tf.train.AdamOptimizer(FLAGS.learning_rate)
grads_and_vars = optimizer.compute_gradients(model.loss, tf.trainable_variables())
train_op = optimizer.apply_gradients(grads_and_vars, global_step=global_step)