Skip to content
Permalink
Branch: master
Find file Copy path
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
530 lines (485 sloc) 21 KB
"""Runs deep learning experiments on speech dataset.
Usage:
run_exp.py [--dataset-path=path] [--dataset-name=timit]
[--iterator-type=sentences] [--batch-size=100] [--nframes=13]
[--features=fbank] [--init-lr=0.001] [--epochs=500]
[--network-type=dropout_net] [--trainer-type=adadelta]
[--prefix-output-fname=my_prefix_42] [--debug-test] [--debug-print=0]
[--debug-time] [--debug-plot=0]
Options:
-h --help Show this screen
--version Show version
--dataset-path=str A valid path to the dataset
default is timit
--dataset-name=str Name of the dataset (for outputs/saves)
default is "timit"
--iterator-type=str "sentences" | "batch" | "dtw"
default is "sentences"
--batch-size=int Batch size, used only by the batch iterator
default is 100 (unused for "sentences" iterator type)
--nframes=int Number of frames to base the first layer on
default is 13
--features=str "fbank" | "MFCC" (some others are not tested)
default is "fbank"
--init-lr=float Initial learning rate for SGD
default is 0.001 (that is very low intentionally)
--epochs=int Max number of epochs (always early stopping)
default is 500
--network-type=str "dropout*" | "*" | "dropout_ab_net*"
default is "dropout_net"
--trainer-type=str "SGD" | "adagrad" | "adadelta"
default is "adadelta"
--prefix-output-fname=str An additional prefix to the output file name
default is "" (empty string)
--debug-test Flag that activates training on the test set
default is False, using it makes it True
--debug-print=int Level of debug printing. 0: nothing, 1: network
default is 0 2: epochs/iters related
--debug-time Flag that activates timing epoch duration
default is False, using it makes it True
--debug-plot=int Level of debug plotting, 1: costs
default is 0 >= 2: gradients & updates
"""
import socket, docopt, cPickle, time, sys, os
import numpy
import matplotlib
matplotlib.use('Agg')
try:
import prettyplotlib as ppl
except:
print >> sys.stderr, "you should install prettyplotlib"
import matplotlib.pyplot as plt
import joblib
import random
from random import shuffle
from prep_timit import load_data
from test_dataset_iterators import DatasetSentencesIterator
from test_dataset_iterators import DatasetDTWIterator, DatasetBatchIteratorPhn
from test_dataset_iterators import DatasetDTWWrdSpkrIterator, DatasetDTReWIterator
from layers import Linear, ReLU, SigmoidLayer, SoftPlus
from classifiers import LogisticRegression
from nnet_archs import ABNeuralNet2Outputs
from nnet_archs import DropoutABNeuralNet # TODO
DEFAULT_DATASET = '/fhgfs/bootphon/scratch/gsynnaeve/TIMIT/train_dev_test_split'
if socket.gethostname() == "syhws-MacBook-Pro.local":
DEFAULT_DATASET = '/Users/gabrielsynnaeve/postdoc/datasets/TIMIT_train_dev_test'
elif socket.gethostname() == "TODO": # TODO
DEFAULT_DATASET = '/media/bigdata/TIMIT_train_dev_test'
DEBUG = False
REDTW = False
DIM_EMBEDDING = 100
def print_mean_weights_biases(params):
for layer_ind, param in enumerate(params):
filler = "weight"
if layer_ind % 2:
filler = "bias"
print("layer %i mean %s values %f and std devs %f" % (layer_ind/2,
filler, numpy.mean(param.eval()), numpy.std(param.eval())))
def plot_costs(cost):
# TODO
pass
def rolling_avg_pgu(iteration, pgu, l):
# (iteration * pgu + l) / (iteration + 1)
assert len(l) == len(pgu)
ll = len(l)/3
params, gparams, updates = l[:ll], l[ll:-ll], l[-ll:]
mpars, mgpars, mupds = pgu[:ll], pgu[ll:-ll], pgu[-ll:]
ii = iteration + 1
return [(iteration * mpars[k] + p) / ii for k, p in enumerate(params)] +\
[(iteration * mgpars[k] + g) / ii for k, g in enumerate(gparams)] +\
[(iteration * mupds[k] + u) / ii for k, u in enumerate(updates)]
def plot_params_gradients_updates(n, l):
# TODO currently works only with THEANO_FLAGS="device=cpu" (not working on
#CudaNDArrays)
def plot_helper(li, ti, p):
if ppl == None:
print >> sys.stderr, "cannot plot this without prettyplotlib"
return
fig, ax = plt.subplots(1)
if li % 2:
title = "biases" + ti
ppl.bar(ax, numpy.arange(p.shape[0]), p) # TODO with plt
else:
title = "weights" + ti
ppl.pcolormesh(fig, ax, p) # TODO with plt
plt.title(title)
plt.savefig(title + ".png")
#ppl.show()
plt.close()
ll = len(l)/3
params, gparams, updates = l[:ll], l[ll:-ll], l[-ll:]
if DEBUG:
print "params"
print params
print "===================="
print "gparams" # TODO find out why not CudaNDArray here
print gparams
print "===================="
print "updates" # TODO find out why not CudaNDArray here
print updates
title_iter = "_%04i" % n
for layer_ind, param in enumerate(params):
title = "_for_layer_" + str(layer_ind/3) + title_iter
plot_helper(layer_ind, title, param)
for layer_ind, gparam in enumerate(gparams):
title = "_gradients_for_layer_" + str(layer_ind/3) + title_iter
plot_helper(layer_ind, title, gparam)
for layer_ind, update in enumerate(updates):
title = "_updates_for_layer_" + str(layer_ind/3) + title_iter
plot_helper(layer_ind, title, update)
def run(dataset_path=DEFAULT_DATASET, dataset_name='timit',
iterator_type=DatasetDTWIterator, batch_size=100,
nframes=13, features="fbank",
init_lr=0.01, max_epochs=500,
network_type="dropout_net", trainer_type="adadelta",
layers_types=[ReLU, ReLU, ReLU, ReLU, LogisticRegression],
layers_sizes=[2400, 2400, 2400, 2400],
dropout_rates=[0.2, 0.5, 0.5, 0.5, 0.5],
recurrent_connections=[],
prefix_fname='',
debug_on_test_only=False,
debug_print=0,
debug_time=False,
debug_plot=0):
"""
FIXME TODO
"""
output_file_name = dataset_name
if prefix_fname != "":
output_file_name = prefix_fname + "_" + dataset_name
output_file_name += "_" + features + str(nframes)
output_file_name += "_" + network_type + "_" + trainer_type
output_file_name += "_emb_" + str(DIM_EMBEDDING)
print "output file name:", output_file_name
n_ins = None
n_outs = None
print "loading dataset from", dataset_path
# TODO DO A FUNCTION
if dataset_path[-7:] != '.joblib':
print >> sys.stderr, "prepare your dataset with align_words.py"
sys.exit(-1)
### LOADING DATA
data_same = joblib.load(dataset_path)
shuffle(data_same)
has_dev_and_test_set = True
dev_dataset_path = dataset_path[:-7].replace("train", "") + 'dev.joblib'
test_dataset_path = dataset_path[:-7].replace("train", "") + 'test.joblib'
dev_split_at = len(data_same)
test_split_at = len(data_same)
if not os.path.exists(dev_dataset_path) or not os.path.exists(test_dataset_path):
has_dev_and_test_set = False
dev_split_at = int(0.8 * dev_split_at)
test_split_at = int(0.9 * test_split_at)
n_ins = data_same[0][3].shape[1] * nframes
n_outs = DIM_EMBEDDING
normalize = True # TODO without
min_max_scale = False
marginf = (nframes-1)/2 # TODO
### TRAIN SET
if has_dev_and_test_set:
train_set_iterator = DatasetDTWWrdSpkrIterator(data_same,
normalize=normalize, min_max_scale=min_max_scale,
scale_f1=None, scale_f2=None, nframes=nframes,
batch_size=batch_size, marginf=marginf)
else:
train_set_iterator = DatasetDTWWrdSpkrIterator(
data_same[:dev_split_at], normalize=normalize,
min_max_scale=min_max_scale, scale_f1=None, scale_f2=None,
nframes=nframes, batch_size=batch_size, marginf=marginf)
f1 = train_set_iterator._scale_f1
f2 = train_set_iterator._scale_f2
### DEV SET
if has_dev_and_test_set:
data_same = joblib.load(dev_dataset_path)
valid_set_iterator = DatasetDTWWrdSpkrIterator(data_same,
normalize=normalize, min_max_scale=min_max_scale,
scale_f1=f1, scale_f2=f2,
nframes=nframes, batch_size=batch_size, marginf=marginf)
else:
valid_set_iterator = DatasetDTWWrdSpkrIterator(
data_same[dev_split_at:test_split_at], normalize=normalize,
min_max_scale=min_max_scale, scale_f1=f1, scale_f2=f2,
nframes=nframes, batch_size=batch_size, marginf=marginf)
### TEST SET
if has_dev_and_test_set:
data_same = joblib.load(test_dataset_path)
test_set_iterator = DatasetDTWWrdSpkrIterator(data_same,
normalize=normalize, min_max_scale=min_max_scale,
scale_f1=f1, scale_f2=f2, nframes=nframes,
batch_size=batch_size, marginf=marginf)
else:
test_set_iterator = DatasetDTWWrdSpkrIterator(
data_same[test_split_at:], normalize=normalize,
min_max_scale=min_max_scale, scale_f1=f1, scale_f2=f2,
nframes=nframes, batch_size=batch_size, marginf=marginf)
assert n_ins != None
assert n_outs != None
# numpy random generator
numpy_rng = numpy.random.RandomState(123)
print '... building the model'
# TODO the proper network type other than just dropout or not
nnet = None
fast_dropout = False
if "fast_dropout" in network_type:
fast_dropout = True
if "dropout" in network_type:
nnet = DropoutABNeuralNet(numpy_rng=numpy_rng,
n_ins=n_ins,
layers_types=layers_types,
layers_sizes=layers_sizes,
n_outs=n_outs,
loss='cos_cos2',
rho=0.95,
eps=1.E-6,
max_norm=4.,
fast_drop=fast_dropout,
debugprint=debug_print)
else:
# nnet = ABNeuralNet2Outputs(numpy_rng=numpy_rng,
# n_ins=n_ins,
# layers_types=layers_types,
# layers_sizes=layers_sizes,
# n_outs=n_outs,
# loss='cos_cos2',
# #loss='euclidean',
# rho=0.90,
# eps=1.E-6,
# max_norm=0.,
# debugprint=debug_print)
from nnet_archs import ABNeuralNet
nnet = ABNeuralNet(numpy_rng=numpy_rng,
n_ins=n_ins,
layers_types=layers_types,
layers_sizes=layers_sizes,
n_outs=n_outs,
loss='cos_cos2',
rho=0.9,
eps=1.E-6,
max_norm=0.,
debugprint=debug_print)
print "Created a neural net as:",
print str(nnet)
# get the training, validation and testing function for the model
print '... getting the training functions'
print trainer_type
train_fn = None
if debug_plot or debug_print:
if trainer_type == "adadelta":
train_fn = nnet.get_adadelta_trainer(debug=True)
elif trainer_type == "adagrad":
train_fn = nnet.get_adagrad_trainer(debug=True)
else:
train_fn = nnet.get_SGD_trainer(debug=True)
else:
if trainer_type == "adadelta":
train_fn = nnet.get_adadelta_trainer()
elif trainer_type == "adagrad":
train_fn = nnet.get_adagrad_trainer()
else:
train_fn = nnet.get_SGD_trainer()
train_scoref = nnet.score_classif_same_diff_separated(train_set_iterator)
valid_scoref = nnet.score_classif_same_diff_separated(valid_set_iterator)
test_scoref = nnet.score_classif(test_set_iterator)
data_iterator = train_set_iterator
if debug_on_test_only:
print >> sys.stderr, "NOT IMPLEMENTED"
sys.exit(-1)
data_iterator = test_set_iterator
print '... training the model'
# early-stopping parameters
patience = 1000 # look as this many examples regardless TODO
patience_increase = 2. # wait this much longer when a new best is
# found
improvement_threshold = 0.995 # a relative improvement of this much is
# considered significant
best_validation_loss = numpy.inf
test_score = 0.
start_time = time.clock()
done_looping = False
epoch = 0
lr = init_lr
timer = None
if debug_plot:
print_mean_weights_biases(nnet.params)
#with open(output_file_name + 'epoch_0.pickle', 'wb') as f:
# cPickle.dump(nnet, f, protocol=-1)
while (epoch < max_epochs) and (not done_looping):
if REDTW and ("ab_net" in network_type or "abnet" in network_type) and ((epoch + 1) % 20) == 0:
print "recomputing DTW:"
data_iterator.recompute_DTW(nnet.transform_x1())
epoch = epoch + 1
avg_costs = []
avg_params_gradients_updates = []
if debug_time:
timer = time.time()
for iteration, (x, y) in enumerate(data_iterator):
#print "x[0][0]", x[0][0]
#print "x[1][0]", x[1][0]
#print "y[0][0]", y[0][0]
#print "y[1][0]", y[1][0]
avg_cost = 0.
if "delta" in trainer_type: # TODO remove need for this if
avg_cost = train_fn(x[0], x[1], y)
else:
avg_cost = train_fn(x[0], x[1], y, lr)
if debug_print >= 3:
print "cost:", avg_cost[0]
if debug_plot >= 2:
plot_costs(avg_cost[0])
if not len(avg_params_gradients_updates):
avg_params_gradients_updates = map(numpy.asarray, avg_cost[1:])
else:
avg_params_gradients_updates = rolling_avg_pgu(
iteration, avg_params_gradients_updates,
map(numpy.asarray, avg_cost[1:]))
if debug_plot >= 3:
plot_params_gradients_updates(iteration, avg_cost[1:])
if type(avg_cost) == list:
avg_costs.append(avg_cost[0])
else:
avg_costs.append(avg_cost)
if debug_print >= 2:
print_mean_weights_biases(nnet.params)
if debug_plot >= 2:
plot_params_gradients_updates(epoch, avg_params_gradients_updates)
if debug_time:
print(' epoch %i took %f seconds' % (epoch, time.time() - timer))
avg_cost = numpy.mean(avg_costs)
if numpy.isnan(avg_cost):
print("avg costs is NaN so we're stopping here!")
break
print(' epoch %i, avg costs %f' % \
(epoch, avg_cost))
tmp_train = zip(*train_scoref())
print(' epoch %i, training sim same %f, diff %f' % \
(epoch, numpy.mean(tmp_train[0]), numpy.mean(tmp_train[1])))
# TODO update lr(t) = lr(0) / (1 + lr(0) * lambda * t)
lr = numpy.float32(init_lr / (numpy.sqrt(iteration) + 1.)) ### TODO
#lr = numpy.float32(init_lr / (iteration + 1.)) ### TODO
# or another scheme for learning rate decay
#with open(output_file_name + 'epoch_' +str(epoch) + '.pickle', 'wb') as f:
# cPickle.dump(nnet, f, protocol=-1)
if debug_on_test_only:
continue
# we check the validation loss on every epoch
validation_losses = zip(*valid_scoref())
this_validation_loss = 0.5*(1.-numpy.mean(validation_losses[0])) +\
0.5*numpy.mean(validation_losses[1])
print(' epoch %i, valid sim same %f, diff %f' % \
(epoch, numpy.mean(validation_losses[0]), numpy.mean(validation_losses[1])))
# if we got the best validation score until now
if this_validation_loss < best_validation_loss:
with open(output_file_name + '.pickle', 'wb') as f:
cPickle.dump(nnet, f, protocol=-1)
# improve patience if loss improvement is good enough
if (this_validation_loss < best_validation_loss *
improvement_threshold):
patience = max(patience, iteration * patience_increase)
# save best validation score and iteration number
best_validation_loss = this_validation_loss
# test it on the test set
test_losses = test_scoref()
test_score_same = numpy.mean(test_losses[0]) # TODO this is a mean of means (with different lengths)
test_score_diff = numpy.mean(test_losses[1]) # TODO this is a mean of means (with different lengths)
print((' epoch %i, test sim of best model same %f diff %f') %
(epoch, test_score_same, test_score_diff))
if patience <= iteration: # TODO correct that
done_looping = True
break
end_time = time.clock()
print(('Optimization complete with best validation score of %f, '
'with test performance %f') %
(best_validation_loss, test_score))
print >> sys.stderr, ('The fine tuning code for file ' +
os.path.split(__file__)[1] +
' ran for %.2fm' % ((end_time - start_time)
/ 60.))
with open(output_file_name + '_final.pickle', 'wb') as f:
cPickle.dump(nnet, f, protocol=-1)
if __name__=='__main__':
arguments = docopt.docopt(__doc__, version='run_exp version 0.1')
dataset_path=DEFAULT_DATASET
if arguments['--dataset-path'] != None:
dataset_path = arguments['--dataset-path']
dataset_name = 'timit'
if arguments['--dataset-name'] != None:
dataset_name = arguments['--dataset-name']
iterator_type = DatasetSentencesIterator
if arguments['--iterator-type'] != None:
if "sentences" in arguments['--iterator-type']:
iterator_type = DatasetSentencesIterator
elif "dtw" in arguments['--iterator-type']:
if "spkr" in arguments['--iterator-type']:
iterator_type = DatasetDTWWrdSpkrIterator
else:
if REDTW:
iterator_type = DatasetDTReWIterator
else:
iterator_type = DatasetDTWIterator
else:
iterator_type = DatasetBatchIteratorPhn # TODO
batch_size = 50 # TODO 10 || 200
if arguments['--batch-size'] != None:
batch_size = int(arguments['--batch-size'])
nframes = 13
if arguments['--nframes'] != None:
nframes = int(arguments['--nframes'])
features = 'fbank'
if arguments['--features'] != None:
features = arguments['--features']
init_lr = 0.01
if arguments['--init-lr'] != None:
init_lr = float(arguments['--init-lr'])
max_epochs = 500
if arguments['--epochs'] != None:
max_epochs = int(arguments['--epochs'])
network_type = 'dropout_net'
if arguments['--network-type'] != None:
network_type = arguments['--network-type']
trainer_type = 'adadelta'
if arguments['--trainer-type'] != None:
trainer_type = arguments['--trainer-type']
prefix_fname = ''
if arguments['--prefix-output-fname'] != None:
prefix_fname = arguments['--prefix-output-fname']
debug_on_test_only = False
if arguments['--debug-test']:
debug_on_test_only = True
debug_print = 0
if arguments['--debug-print']:
debug_print = int(arguments['--debug-print'])
debug_time = False
if arguments['--debug-time']:
debug_time = True
debug_plot = 0
if arguments['--debug-plot']:
debug_plot = int(arguments['--debug-plot'])
run(dataset_path=dataset_path, dataset_name=dataset_name,
iterator_type=iterator_type, batch_size=batch_size,
nframes=nframes, features=features,
init_lr=init_lr, max_epochs=max_epochs,
network_type=network_type, trainer_type=trainer_type,
#layers_types=[ReLU, ReLU, ReLU, SigmoidLayer],
#layers_types=[ReLU, ReLU, ReLU, ReLU],
#layers_types=[SoftPlus, SoftPlus, SoftPlus, SoftPlus],
#layers_types=[SoftPlus, SoftPlus, SoftPlus, SigmoidLayer],
#layers_sizes=[1000, 1000, 1000],
#dropout_rates=[0.2, 0.5, 0.5, 0.5],
#layers_types=[ReLU, ReLU],
layers_types=[ReLU, SigmoidLayer],
#layers_types=[SoftPlus, SigmoidLayer],
#layers_types=[SigmoidLayer, SigmoidLayer],
layers_sizes=[200],
#layers_sizes=[1000],
#layers_types=[ReLU],
#layers_types=[SoftPlus],
#layers_types=[SigmoidLayer],
#layers_sizes=[],
recurrent_connections=[], # TODO in opts
prefix_fname=prefix_fname,
debug_on_test_only=debug_on_test_only,
debug_print=debug_print,
debug_time=debug_time,
debug_plot=debug_plot)
# TODO I-vector features that are averaged at least on a whole word (UBM like)
#THEANO_FLAGS='device=gpu0' python run_exp_AB_phn_spkr.py --dataset-path=LUCID_9chars.joblib --dataset-name=LUCID_9chars --nframes=7 --network-type=abnet --debug-print=1 --debug-time
You can’t perform that action at this time.