add reference

paddlespeech/t2s/data/batch.py

@@ -52,9 +52,8 @@ def batch_text_id(minibatch, pad_id=0, dtype=np.int64):
     """
     peek_example = minibatch[0]
     assert len(peek_example.shape) == 1, "text example is an 1D tensor"
-
+    # assume (channel, n_samples) or (n_samples, )
-    lengths = [example.shape[0] for example in
+    lengths = [example.shape[0] for example in minibatch]
-               minibatch]  # assume (channel, n_samples) or (n_samples, )
     max_len = np.max(lengths)
 
     batch = []

paddlespeech/t2s/data/get_feats.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 import librosa
 import numpy as np
 import pyworld

paddlespeech/t2s/models/fastspeech2/fastspeech2.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Fastspeech2 related modules for paddle"""
 from typing import Dict
 from typing import Sequence

paddlespeech/t2s/models/melgan/melgan.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """MelGAN Modules."""
 from typing import Any
 from typing import Dict

paddlespeech/t2s/models/parallel_wavegan/parallel_wavegan.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 import math
 from typing import Any
 from typing import Dict

paddlespeech/t2s/models/transformer_tts/transformer_tts.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Fastspeech2 related modules for paddle"""
 from typing import Dict
 from typing import Sequence

paddlespeech/t2s/modules/adversarial_loss.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Adversarial loss modules."""
 import paddle
 import paddle.nn.functional as F

paddlespeech/t2s/modules/fastspeech2_predictor/duration_predictor.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Duration predictor related modules."""
 import paddle
 from paddle import nn

paddlespeech/t2s/modules/fastspeech2_predictor/length_regulator.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Length regulator related modules."""
 import paddle
 from paddle import nn

paddlespeech/t2s/modules/fastspeech2_predictor/variance_predictor.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Variance predictor related modules."""
 import paddle
 from paddle import nn

paddlespeech/t2s/modules/fastspeech2_transformer/attention.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Multi-Head Attention layer definition."""
 import math
 

paddlespeech/t2s/modules/fastspeech2_transformer/decoder.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 # 暂时删除了 dyminic conv
 """Decoder definition."""
 import logging

paddlespeech/t2s/modules/fastspeech2_transformer/decoder_layer.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Decoder self-attention layer definition."""
 import paddle
 from paddle import nn

paddlespeech/t2s/modules/fastspeech2_transformer/embedding.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Positional Encoding Module."""
 import math
 

paddlespeech/t2s/modules/fastspeech2_transformer/encoder.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 import logging
 
 from paddle import nn

paddlespeech/t2s/modules/fastspeech2_transformer/encoder_layer.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Encoder self-attention layer definition."""
 import paddle
 from paddle import nn

paddlespeech/t2s/modules/fastspeech2_transformer/lightconv.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Lightweight Convolution Module."""
 import numpy
 import paddle

paddlespeech/t2s/modules/fastspeech2_transformer/positionwise_feed_forward.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Positionwise feed forward layer definition."""
 import paddle
 

paddlespeech/t2s/modules/fastspeech2_transformer/repeat.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Repeat the same layer definition."""
 import paddle
 

paddlespeech/t2s/modules/nets_utils.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 import paddle
 from paddle import nn
 from typeguard import check_argument_types

paddlespeech/t2s/modules/pqmf.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Pseudo QMF modules."""
 import numpy as np
 import paddle

paddlespeech/t2s/modules/residual_stack.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Residual stack module in MelGAN."""
 from typing import Any
 from typing import Dict

paddlespeech/t2s/modules/stft_loss.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 import paddle
 from paddle import nn
 from paddle.nn import functional as F

paddlespeech/t2s/modules/style_encoder.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Style encoder of GST-Tacotron."""
 from typing import Sequence
 

paddlespeech/t2s/modules/tacotron2/decoder.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Tacotron2 decoder related modules."""
 import paddle.nn.functional as F
 import six

paddlespeech/t2s/modules/tacotron2/encoder.py

@@ -11,6 +11,7 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+# Modified from espnet(https://github.com/espnet/espnet)
 """Tacotron2 encoder related modules."""
 import paddle
 import six