eden-emu/opus
15
0
Fork 0
mirror of https://github.com/xiph/opus.git synced 2025-06-01 16:17:42 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix non-128 batch sizes

Browse source 
Avoid hardcoding the batch size in the model
This commit is contained in:
Jean-Marc Valin 2021-10-09 03:20:22 -04:00
parent 37c9bd8d28
commit 8cdc8081d8
2 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

8
dnn/training_tf2/lpcnet.py
View file

@ -230,10 +230,10 @@ class WeightClip(Constraint):
constraint = WeightClip(0.992)
def new_lpcnet_model(rnn_units1=384, rnn_units2=16, nb_used_features = 20, training=False, adaptation=False, quantize=False, flag_e2e = False):
pcm = Input(shape=(None, 3), batch_size=128)
feat = Input(shape=(None, nb_used_features), batch_size=128)
pitch = Input(shape=(None, 1), batch_size=128)
def new_lpcnet_model(rnn_units1=384, rnn_units2=16, nb_used_features=20, batch_size=128, training=False, adaptation=False, quantize=False, flag_e2e = False):
pcm = Input(shape=(None, 3), batch_size=batch_size)
feat = Input(shape=(None, nb_used_features), batch_size=batch_size)
pitch = Input(shape=(None, 1), batch_size=batch_size)
dec_feat = Input(shape=(None, 128))
dec_state1 = Input(shape=(rnn_units1,))
dec_state2 = Input(shape=(rnn_units2,))

2
dnn/training_tf2/train_lpcnet.py
View file

@ -121,7 +121,7 @@ opt = Adam(lr, decay=decay, beta_2=0.99)
strategy = tf.distribute.experimental.MultiWorkerMirroredStrategy()
with strategy.scope():
model, _, _ = lpcnet.new_lpcnet_model(rnn_units1=args.grua_size, rnn_units2=args.grub_size, training=True, quantize=quantize, flag_e2e = flag_e2e)
model, _, _ = lpcnet.new_lpcnet_model(rnn_units1=args.grua_size, rnn_units2=args.grub_size, batch_size=batch_size, training=True, quantize=quantize, flag_e2e = flag_e2e)
if not flag_e2e:
model.compile(optimizer=opt, loss='sparse_categorical_crossentropy', metrics='sparse_categorical_crossentropy')
else: