@ -44,7 +44,7 @@ class DeepSpeech2Model(object):
for GRU , weight sharing is not supported .
: type share_rnn_weights : bool
: param place : Program running place .
: type place : CPU or GPU
: type place : CPU Place or CUDAPlace
: param init_from_pretrained_model : Pretrained model path . If None , will train
from stratch .
: type init_from_pretrained_model : string | None
@ -60,7 +60,7 @@ class DeepSpeech2Model(object):
use_gru = False ,
share_rnn_weights = True ,
place = fluid . CPUPlace ( ) ,
init_from_pretrain _model= None ,
init_from_pretrain ed _model= None ,
output_model_dir = None ) :
self . _vocab_size = vocab_size
self . _num_conv_layers = num_conv_layers
@ -69,7 +69,7 @@ class DeepSpeech2Model(object):
self . _use_gru = use_gru
self . _share_rnn_weights = share_rnn_weights
self . _place = place
self . _init_from_pretrain _model = init_from_pretrain _model
self . _init_from_pretrain ed _model = init_from_pretrain ed _model
self . _output_model_dir = output_model_dir
self . _ext_scorer = None
self . logger = logging . getLogger ( " " )
@ -90,13 +90,14 @@ class DeepSpeech2Model(object):
if not is_infer :
input_fields = {
' names ' : [ ' audio_data ' , ' text_data ' , ' seq_len_data ' , ' masks ' ] ,
' shapes ' : [ [ - 1 , 161 , 161 ] , [ - 1 , 1 ] , [ - 1 , 1 ] , [ - 1 , 32 , 81 , 1 ] ] ,
' shapes ' :
[ [ None , 161 , None ] , [ None , 1 ] , [ None , 1 ] , [ None , 32 , 81 , None ] ] ,
' dtypes ' : [ ' float32 ' , ' int32 ' , ' int64 ' , ' float32 ' ] ,
' lod_levels ' : [ 0 , 1 , 0 , 0 ]
}
inputs = [
fluid . layers. data(
fluid . (
name = input_fields [ ' names ' ] [ i ] ,
shape = input_fields [ ' shapes ' ] [ i ] ,
dtype = input_fields [ ' dtypes ' ] [ i ] ,
@ -104,7 +105,7 @@ class DeepSpeech2Model(object):
for i in range ( len ( input_fields [ ' names ' ] ) )
]
reader = fluid . io . PyReade r(
reader = fluid . io . DataLoader. from_generato r(
feed_list = inputs ,
capacity = 64 ,
iterable = False ,
@ -112,16 +113,19 @@ class DeepSpeech2Model(object):
( audio_data , text_data , seq_len_data , masks ) = inputs
else :
audio_data = fluid . layers. data(
audio_data = fluid . (
name = ' audio_data ' ,
shape = [ - 1 , 161 , 161 ] ,
shape = [ None , 161 , None ] ,
dtype = ' float32 ' ,
lod_level = 0 )
seq_len_data = fluid . layers . data (
name = ' seq_len_data ' , shape = [ - 1 , 1 ] , dtype = ' int64 ' , lod_level = 0 )
masks = fluid . layers . data (
seq_len_data = fluid . data (
name = ' seq_len_data ' ,
shape = [ None , 1 ] ,
dtype = ' int64 ' ,
lod_level = 0 )
masks = fluid . data (
name = ' masks ' ,
shape = [ - 1 , 32 , 81 , 1 ] ,
shape = [ None , 32 , 81 , None ] ,
dtype = ' float32 ' ,
lod_level = 0 )
text_data = None
@ -141,26 +145,26 @@ class DeepSpeech2Model(object):
share_rnn_weights = self . _share_rnn_weights )
return reader , log_probs , loss
def init_from_pretrain _model( self , exe , program ) :
def init_from_pretrain ed _model( self , exe , program ) :
''' Init params from pretrain model. '''
assert isinstance ( self . _init_from_pretrain _model, str )
assert isinstance ( self . _init_from_pretrain ed _model, str )
if not os . path . exists ( self . _init_from_pretrain _model) :
print ( self . _init_from_pretrain _model)
if not os . path . exists ( self . _init_from_pretrain ed _model) :
print ( self . _init_from_pretrain ed _model)
raise Warning ( " The pretrained params do not exist. " )
return False
fluid . io . load_params (
exe ,
self . _init_from_pretrain _model,
self . _init_from_pretrain ed _model,
main_program = program ,
filename = " params.pdparams " )
print ( " finish initing model from pretrained params from %s " %
( self . _init_from_pretrain _model) )
( self . _init_from_pretrain ed _model) )
pre_epoch = 0
dir_name = self . _init_from_pretrain _model. split ( ' _ ' )
dir_name = self . _init_from_pretrain ed _model. split ( ' _ ' )
if len ( dir_name ) > = 2 and dir_name [ - 2 ] . endswith ( ' epoch ' ) and dir_name [
- 1 ] . isdigit ( ) :
pre_epoch = int ( dir_name [ - 1 ] )
@ -186,7 +190,7 @@ class DeepSpeech2Model(object):
return True
def test ( self , exe , dev_batch_reader , test_program , test_ py reader,
def test ( self , exe , dev_batch_reader , test_program , test_ reader,
fetch_list ) :
''' Test the model.
@ -196,14 +200,14 @@ class DeepSpeech2Model(object):
: type dev_batch_reader : read generator
: param test_program : The program of test .
: type test_program : Program
: param test_ pyreader: Pyr eader of test .
: type test_ pyreader: Pyr eader
: param test_ reader: R eader of test .
: type test_ reader: R eader
: param fetch_list : Fetch list .
: type fetch_list : list
: return : An output unnormalized log probability .
: rtype : array
'''
test_ py reader. start ( )
test_ reader. start ( )
epoch_loss = [ ]
while True :
try :
@ -214,7 +218,7 @@ class DeepSpeech2Model(object):
epoch_loss . extend ( np . array ( each_loss [ 0 ] ) )
except fluid . core . EOFException :
test_ py reader. reset ( )
test_ reader. reset ( )
break
return np . mean ( np . array ( epoch_loss ) )
@ -274,7 +278,7 @@ class DeepSpeech2Model(object):
startup_prog = fluid . Program ( )
with fluid . program_guard ( train_program , startup_prog ) :
with fluid . unique_name . guard ( ) :
train_ py reader, log_probs , ctc_loss = self . create_network ( )
train_ reader, log_probs , ctc_loss = self . create_network ( )
# prepare optimizer
optimizer = fluid . optimizer . AdamOptimizer (
learning_rate = fluid . layers . exponential_decay (
@ -290,7 +294,7 @@ class DeepSpeech2Model(object):
test_prog = fluid . Program ( )
with fluid . program_guard ( test_prog , startup_prog ) :
with fluid . unique_name . guard ( ) :
test_ py reader, _ , ctc_loss = self . create_network ( )
test_ reader, _ , ctc_loss = self . create_network ( )
test_prog = test_prog . clone ( for_test = True )
@ -299,8 +303,8 @@ class DeepSpeech2Model(object):
# init from some pretrain models, to better solve the current task
pre_epoch = 0
if self . _init_from_pretrain _model:
pre_epoch = self . init_from_pretrain _model( exe , train_program )
if self . _init_from_pretrain ed _model:
pre_epoch = self . init_from_pretrain ed _model( exe , train_program )
build_strategy = compiler . BuildStrategy ( )
exec_strategy = fluid . ExecutionStrategy ( )
@ -312,12 +316,12 @@ class DeepSpeech2Model(object):
build_strategy = build_strategy ,
exec_strategy = exec_strategy )
train_ pyreader. decorate _batch_generator( train_batch_reader )
test_ pyreader. decorate _batch_generator( dev_batch_reader )
train_ reader. set _batch_generator( train_batch_reader )
test_ reader. set _batch_generator( dev_batch_reader )
# run train
for epoch_id in range ( num_epoch ) :
train_ py reader. start ( )
train_ reader. start ( )
epoch_loss = [ ]
time_begin = time . time ( )
batch_id = 0
@ -346,7 +350,7 @@ class DeepSpeech2Model(object):
batch_id = batch_id + 1
except fluid . core . EOFException :
train_ py reader. reset ( )
train_ reader. reset ( )
break
time_end = time . time ( )
used_time = time_end - time_begin
@ -359,7 +363,7 @@ class DeepSpeech2Model(object):
exe ,
dev_batch_reader = dev_batch_reader ,
test_program = test_prog ,
test_ py reader= test_ py reader,
test_ reader= test_ reader,
fetch_list = [ ctc_loss ] )
print (
" --------Time: %f sec, epoch: %d , train loss: %f , test loss: %f "
@ -402,10 +406,10 @@ class DeepSpeech2Model(object):
exe = fluid . Executor ( self . _place )
exe . run ( startup_prog )
# init param from pretrain _model
if not self . _init_from_pretrain _model:
# init param from pretrain ed _model
if not self . _init_from_pretrain ed _model:
exit ( " No pretrain model file path! " )
self . init_from_pretrain _model( exe , infer_program )
self . init_from_pretrain ed _model( exe , infer_program )
infer_results = [ ]
time_begin = time . time ( )
lfchener
5 years ago