# training loop
for x, y_, epoch in txt.rnn_minibatch_sequencer(codetext, BATCHSIZE, SEQLEN, nb_epochs=20):
# train on one minibatch
feed_dict = {X: x, Y_: y_, Hin: istate, lr: learning_rate, pkeep: dropout_pkeep, batchsize: BATCHSIZE}
_, y, ostate, smm = sess.run([train_step, Y, H, summaries], feed_dict=feed_dict)
# save training data for Tensorboard
summary_writer.add_summary(smm, step)
# display a visual validation of progress (every 50 batches)
if step % _50_BATCHES == 0:
feed_dict = {X: x, Y_: y_, Hin: istate, pkeep: 1.0, batchsize: BATCHSIZE} # no dropout for validation
y, l, bl, acc = sess.run([Y, seqloss, batchloss, accuracy], feed_dict=feed_dict)
txt.print_learning_learned_comparison(x, y, l, bookranges, bl, acc, epoch_size, step, epoch)
# run a validation step every 50 batches
# The validation text should be a single sequence but that's too slow (1s per 1024 chars!),
# so we cut it up and batch the pieces (slightly inaccurate)
# tested: validating with 5K sequences instead of 1K is only slightly more accurate, but a lot slower.
if step % _50_BATCHES == 0 and len(valitext) > 0:
VALI_SEQLEN = 1*1024 # Sequence length for validation. State will be wrong at the start of each sequence.
bsize = len(valitext) // VALI_SEQLEN
txt.print_validation_header(len(codetext), bookranges)
vali_x, vali_y, _ = next(txt.rnn_minibatch_sequencer(valitext, bsize, VALI_SEQLEN, 1)) # all data in 1 batch
vali_nullstate = np.zeros([bsize, INTERNALSIZE*NLAYERS])
feed_dict = {X: vali_x, Y_: vali_y, Hin: vali_nullstate, pkeep: 1.0, # no dropout for validation
batchsize: bsize}
ls, acc, smm = sess.run([batchloss, accuracy, summaries], feed_dict=feed_dict)
txt.print_validation_stats(ls, acc)
def main(_):
# load data, either shakespeare, or the Python source of Tensorflow itself
shakedir = FLAGS.text_dir
# shakedir = "../tensorflow/**/*.py"
codetext, valitext, bookranges = txt.read_data_files(shakedir,
validation=True)
# display some stats on the data
epoch_size = len(codetext) // (FLAGS.train_batch_size * FLAGS.seqlen)
txt.print_data_stats(len(codetext), len(valitext), epoch_size)
#
# the model (see FAQ in README.md)
#
lr = tf.placeholder(tf.float32, name='lr') # learning rate
pkeep = tf.placeholder(tf.float32, name='pkeep') # dropout parameter
batchsize = tf.placeholder(tf.int32, name='batchsize')
# inputs
X = tf.placeholder(tf.uint8, [None, None],
name='X') # [ BATCHSIZE, FLAGS.seqlen ]
Xo = tf.one_hot(X, ALPHASIZE, 1.0,
0.0) # [ BATCHSIZE, FLAGS.seqlen, ALPHASIZE ]
# expected outputs = same sequence shifted by 1 since we are trying to predict the next character
Y_ = tf.placeholder(tf.uint8, [None, None],
name='Y_') # [ BATCHSIZE, FLAGS.seqlen ]
Yo_ = tf.one_hot(Y_, ALPHASIZE, 1.0,
0.0) # [ BATCHSIZE, FLAGS.seqlen, ALPHASIZE ]
# input state
Hin = tf.placeholder(tf.float32, [None, INTERNALSIZE * NLAYERS],
name='Hin') # [ BATCHSIZE, INTERNALSIZE * NLAYERS]
# using a NLAYERS=3 layers of GRU cells, unrolled FLAGS.seqlen=30 times
# dynamic_rnn infers FLAGS.seqlen from the size of the inputs Xo
onecell = rnn.GRUCell(INTERNALSIZE)
dropcell = rnn.DropoutWrapper(onecell, input_keep_prob=pkeep)
multicell = rnn.MultiRNNCell([dropcell] * NLAYERS, state_is_tuple=False)
multicell = rnn.DropoutWrapper(multicell, output_keep_prob=pkeep)
Yr, H = tf.nn.dynamic_rnn(multicell,
Xo,
dtype=tf.float32,
initial_state=Hin)
# Yr: [ BATCHSIZE, FLAGS.seqlen, INTERNALSIZE ]
# H: [ BATCHSIZE, INTERNALSIZE*NLAYERS ] # this is the last state in the sequence
H = tf.identity(H, name='H') # just to give it a name
# Softmax layer implementation:
# Flatten the first two dimension of the output [ BATCHSIZE, FLAGS.seqlen, ALPHASIZE ] => [ BATCHSIZE x FLAGS.seqlen, ALPHASIZE ]
# then apply softmax readout layer. This way, the weights and biases are shared across unrolled time steps.
# From the readout point of view, a value coming from a cell or a minibatch is the same thing
Yflat = tf.reshape(
Yr, [-1, INTERNALSIZE]) # [ BATCHSIZE x FLAGS.seqlen, INTERNALSIZE ]
Ylogits = layers.linear(
Yflat, ALPHASIZE) # [ BATCHSIZE x FLAGS.seqlen, ALPHASIZE ]
Yflat_ = tf.reshape(
Yo_, [-1, ALPHASIZE]) # [ BATCHSIZE x FLAGS.seqlen, ALPHASIZE ]
loss = tf.nn.softmax_cross_entropy_with_logits(
logits=Ylogits, labels=Yflat_) # [ BATCHSIZE x FLAGS.seqlen ]
loss = tf.reshape(loss, [batchsize, -1]) # [ BATCHSIZE, FLAGS.seqlen ]
Yo = tf.nn.softmax(Ylogits,
name='Yo') # [ BATCHSIZE x FLAGS.seqlen, ALPHASIZE ]
Y = tf.argmax(Yo, 1) # [ BATCHSIZE x FLAGS.seqlen ]
Y = tf.reshape(Y, [batchsize, -1], name="Y") # [ BATCHSIZE, FLAGS.seqlen ]
train_step = tf.train.AdamOptimizer(lr).minimize(loss)
# stats for display
seqloss = tf.reduce_mean(loss, 1)
batchloss = tf.reduce_mean(seqloss)
accuracy = tf.reduce_mean(
tf.cast(tf.equal(Y_, tf.cast(Y, tf.uint8)), tf.float32))
loss_summary = tf.summary.scalar("batch_loss", batchloss)
acc_summary = tf.summary.scalar("batch_accuracy", accuracy)
summaries = tf.summary.merge([loss_summary, acc_summary])
# Init Tensorboard stuff. This will save Tensorboard information into a different
# folder at each run named 'log/<timestamp>/'. Two sets of data are saved so that
# you can compare training and validation curves visually in Tensorboard.
timestamp = str(math.trunc(time.time()))
summary_writer = tf.summary.FileWriter(
os.path.join(FLAGS.summaries_dir, timestamp + "-training"))
validation_writer = tf.summary.FileWriter(
os.path.join(FLAGS.summaries_dir, timestamp + "-validation"))
# Init for saving models. They will be saved into a directory named 'checkpoints'.
# Only the last checkpoint is kept.
if not os.path.exists(FLAGS.checkpoint_dir):
os.mkdir(FLAGS.checkpoint_dir)
saver = tf.train.Saver(max_to_keep=1)
# for display: init the progress bar
DISPLAY_FREQ = 50
_50_BATCHES = DISPLAY_FREQ * FLAGS.train_batch_size * FLAGS.seqlen
progress = txt.Progress(DISPLAY_FREQ,
size=111 + 2,
msg="Training on next " + str(DISPLAY_FREQ) +
" batches")
# init
istate = np.zeros([FLAGS.train_batch_size,
INTERNALSIZE * NLAYERS]) # initial zero input state
init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init)
step = 0
# training loop
for x, y_, epoch in txt.rnn_minibatch_sequencer(codetext,
FLAGS.train_batch_size,
FLAGS.seqlen,
nb_epochs=1000):
# train on one minibatch
feed_dict = {
X: x,
Y_: y_,
Hin: istate,
lr: FLAGS.learning_rate,
pkeep: FLAGS.dropout_pkeep,
batchsize: FLAGS.train_batch_size
}
_, y, ostate, smm = sess.run([train_step, Y, H, summaries],
feed_dict=feed_dict)
# save training data for Tensorboard
summary_writer.add_summary(smm, step)
# display a visual validation of progress (every 50 batches)
if step % _50_BATCHES == 0:
feed_dict = {
X: x,
Y_: y_,
Hin: istate,
pkeep: 1.0,
batchsize: FLAGS.train_batch_size
} # no dropout for validation
y, l, bl, acc = sess.run([Y, seqloss, batchloss, accuracy],
feed_dict=feed_dict)
txt.print_learning_learned_comparison(x, y, l, bookranges, bl, acc,
epoch_size, step, epoch)
# run a validation step every 50 batches
# The validation text should be a single sequence but that's too slow (1s per 1024 chars!),
# so we cut it up and batch the pieces (slightly inaccurate)
# tested: validating with 5K sequences instead of 1K is only slightly more accurate, but a lot slower.
if step % _50_BATCHES == 0 and len(valitext) > 0:
VALI_SEQLEN = 1 * 1024 # Sequence length for validation. State will be wrong at the start of each sequence.
bsize = len(valitext) // VALI_SEQLEN
txt.print_validation_header(len(codetext), bookranges)
vali_x, vali_y, _ = next(
txt.rnn_minibatch_sequencer(valitext, bsize, VALI_SEQLEN,
1)) # all data in 1 batch
vali_nullstate = np.zeros([bsize, INTERNALSIZE * NLAYERS])
feed_dict = {
X: vali_x,
Y_: vali_y,
Hin: vali_nullstate,
pkeep: 1.0, # no dropout for validation
batchsize: bsize
}
ls, acc, smm = sess.run([batchloss, accuracy, summaries],
feed_dict=feed_dict)
txt.print_validation_stats(ls, acc)
# save validation data for Tensorboard
validation_writer.add_summary(smm, step)
# display a short text generated with the current weights and biases (every 150 batches)
if step // 3 % _50_BATCHES == 0:
txt.print_text_generation_header()
ry = np.array([[txt.convert_from_alphabet(ord("K"))]])
rh = np.zeros([1, INTERNALSIZE * NLAYERS])
for k in range(1000):
ryo, rh = sess.run([Yo, H],
feed_dict={
X: ry,
pkeep: 1.0,
Hin: rh,
batchsize: 1
})
rc = txt.sample_from_probabilities(
ryo, topn=10 if epoch <= 1 else 2)
print(chr(txt.convert_to_alphabet(rc)), end="")
ry = np.array([[rc]])
txt.print_text_generation_footer()
# save a checkpoint (every 500 batches)
if step // 10 % _50_BATCHES == 0:
saver.save(sess,
FLAGS.checkpoint_dir + '/rnn_train_' + timestamp,
global_step=step)
# display progress bar
progress.step(reset=step % _50_BATCHES == 0)
# loop state around
istate = ostate
step += FLAGS.train_batch_size * FLAGS.seqlen
batchsize: BATCH_SIZE
}
_, y, ostate = sess.run([train_step, Y, H], feed_dict=feed_dict)
# Store training TensorBoard data
if step % _50_BATCHES == 0:
feed_dict = {
X: x,
Y_: y_,
Hin: istate,
pkeep: 1.0,
batchsize: BATCH_SIZE
}
y, l, bl, acc, smm = sess.run(
[Y, seqloss, batchloss, accuracy, summaries], feed_dict=feed_dict)
txt.print_learning_learned_comparison(x, y, l, scriptranges, bl, acc,
size_of_epoch, step, epoch)
summary_writer.add_summary(smm, step)
# Validation is batched in order to run quicker
if step % _50_BATCHES == 0 and len(validtext) > 0:
VALI_SEQ_LEN = 1 * 1024
bsize = len(validtext) // VALI_SEQ_LEN
txt.print_validation_header(len(traintext), scriptranges)
vali_x, vali_y, _ = next(
txt.rnn_minibatch_sequencer(validtext, bsize, VALI_SEQ_LEN, 1))
vali_nullstate = np.zeros([bsize, NUM_OF_GRUS * NUM_LAYERS])
feed_dict = {
X: vali_x,
Y_: vali_y,
Hin: vali_nullstate,
pkeep: 1.0,
def fit(self, data, epochs=1000, displayFreq=50, genFreq=150, saveFreq=5000, verbosity=2):
progress = txt.Progress(displayFreq, size=111+2, msg="Training on next "+str(displayFreq)+" batches")
tfStuff = self.tfStuff
valitext = data.valitext
# todo: check if batchSize != data.batchSize or if seqLen != data.seqLen (if so, I think we need to raise an exception?)
firstEpoch = self.curEpoch
lastEpoch = firstEpoch + epochs
isFirstStepInThisFitCall = True
try:
with tfStuff.graph.as_default():
#with tf.name_scope(self.scopeName):
with tf.variable_scope(self.fullName, reuse=tf.AUTO_REUSE):
sess = tfStuff.sess
# training loop
for x, y_, epoch, batch in txt.rnn_minibatch_sequencer(data.codetext, self.batchSize, self.seqLen, nb_epochs=epochs, startBatch=self.curBatch, startEpoch=self.curEpoch):
nSteps = self.step // (self.batchSize*self.seqLen)
# train on one minibatch
feed_dict = {tfStuff.X: x, tfStuff.Y_: y_, tfStuff.Hin: tfStuff.istate, tfStuff.lr: self.learningRate, tfStuff.pkeep: self.dropoutPkeep, tfStuff.batchsize: self.batchSize}
_, y, ostate = sess.run([tfStuff.train_step, tfStuff.Y, tfStuff.H], feed_dict=feed_dict)
# log training data for Tensorboard display a mini-batch of sequences (every 50 batches)
if nSteps % displayFreq == 0 or isFirstStepInThisFitCall:
feed_dict = {tfStuff.X: x, tfStuff.Y_: y_, tfStuff.Hin: tfStuff.istate, tfStuff.pkeep: 1.0, tfStuff.batchsize: self.batchSize} # no dropout for validation
y, l, bl, acc, smm = sess.run([tfStuff.Y, tfStuff.seqloss, tfStuff.batchloss, tfStuff.accuracy, tfStuff.summaries], feed_dict=feed_dict)
txt.print_learning_learned_comparison(x, y, l, data.bookranges, bl, acc, data.epoch_size, self.step, epoch, lastEpoch, verbosity=verbosity)
self.tbStuff.summary_writer.add_summary(smm, self.step)
# run a validation step every 50 batches
# The validation text should be a single sequence but that's too slow (1s per 1024 chars!),
# so we cut it up and batch the pieces (slightly inaccurate)
# tested: validating with 5K sequences instead of 1K is only slightly more accurate, but a lot slower.
if (nSteps % displayFreq == 0 or isFirstStepInThisFitCall) and len(data.valitext) > 0:
VALI_seqLen = 1*1024 # Sequence length for validation. State will be wrong at the start of each sequence.
bsize = len(data.valitext) // VALI_seqLen
if verbosity >= 1: txt.print_validation_header(len(data.codetext), data.bookranges)
vali_x, vali_y, _, _ = next(txt.rnn_minibatch_sequencer(data.valitext, bsize, VALI_seqLen, 1)) # all data in 1 batch
vali_nullstate = np.zeros([bsize, self.internalSize * self.nLayers])
feed_dict = {tfStuff.X: vali_x, tfStuff.Y_: vali_y, tfStuff.Hin: vali_nullstate, tfStuff.pkeep: 1.0, # no dropout for validation
tfStuff.batchsize: bsize}
ls, acc, smm = sess.run([tfStuff.batchloss, tfStuff.accuracy, tfStuff.summaries], feed_dict=feed_dict)
if verbosity >= 1: txt.print_validation_stats(ls, acc)
# save validation data for Tensorboard
self.tbStuff.validation_writer.add_summary(smm, self.step)
# display a short text generated with the current weights and biases (every 150 batches)
if nSteps % genFreq == 0 or isFirstStepInThisFitCall:
txt.print_text_generation_header()
ry = np.array([[txt.convert_from_alphabet(ord("K"))]])
rh = np.zeros([1, self.internalSize * self.nLayers])
for k in range(1000):
ryo, rh = sess.run([tfStuff.Yo, tfStuff.H], feed_dict={tfStuff.X: ry, tfStuff.pkeep: 1.0, tfStuff.Hin: rh, tfStuff.batchsize: 1})
rc = txt.sample_from_probabilities(ryo, topn=10 if epoch <= 1 else 2)
print(chr(txt.convert_to_alphabet(rc)), end="")
ry = np.array([[rc]])
txt.print_text_generation_footer()
if isFirstStepInThisFitCall:
for i in range(nSteps % displayFreq):
progress.step()
isFirstStepInThisFitCall = False
# save a checkpoint (every 500 batches)
if nSteps % saveFreq == 0:
self.save(alreadyInGraph=True)
# display progress bar
progress.step(reset=nSteps % displayFreq == 0)
# loop state around
tfStuff.istate = ostate
self.step += self.batchSize * self.seqLen
self.curEpoch = epoch
self.curBatch = batch
except KeyboardInterrupt as e:
print("\npressed ctrl-c, saving")
self.save()