update branch (#4074)

Signed-off-by: ericharper <complex451@gmail.com> Co-authored-by: Subhankar Ghosh <subhankar2321@gmail.com> Co-authored-by: Somshubra Majumdar <titu1994@gmail.com>
NVIDIA · Apr 28, 2022 · 655ff80 · 655ff80
1 parent f776442
commit 655ff80
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 13 deletions.
diff --git a/nemo/collections/tts/modules/fastpitch.py b/nemo/collections/tts/modules/fastpitch.py
@@ -158,8 +158,8 @@ def __init__(
         else:
             self.speaker_emb = None
 
-        self.register_buffer('max_token_duration', torch.tensor(max_token_duration), persistent=False)
-        self.register_buffer('min_token_duration', torch.tensor(0.0), persistent=False)
+        self.max_token_duration = max_token_duration
+        self.min_token_duration = 0
 
         self.pitch_emb = torch.nn.Conv1d(
             1,
@@ -171,7 +171,6 @@ def __init__(
         # Store values precomputed from training data for convenience
         self.register_buffer('pitch_mean', torch.zeros(1))
         self.register_buffer('pitch_std', torch.zeros(1))
-        self.register_buffer('zero_emb', torch.zeros(1), persistent=False)
 
         self.proj = torch.nn.Linear(self.decoder.d_model, n_mel_channels, bias=True)
 
@@ -225,7 +224,7 @@ def forward(
 
         # Calculate speaker embedding
         if self.speaker_emb is None or speaker is None:
-            spk_emb = self.zero_emb
+            spk_emb = 0
         else:
             spk_emb = self.speaker_emb(speaker).unsqueeze(1)
 

diff --git a/tutorials/AudioTranslationSample.ipynb b/tutorials/AudioTranslationSample.ipynb
@@ -116,7 +116,7 @@
                 "nmt_model = nemo_nlp.models.MTEncDecModel.from_pretrained(model_name='nmt_zh_en_transformer6x6').cuda()\n",
                 "\n",
                 "# Spectrogram generator which takes text as an input and produces spectrogram\n",
-                "spectrogram_generator = nemo_tts.models.Tacotron2Model.from_pretrained(model_name=\"tts_en_tacotron2\").cuda()\n",
+                "spectrogram_generator = nemo_tts.models.FastPitchModel.from_pretrained(model_name=\"tts_en_fastpitch\").cuda()\n",
                 "\n",
                 "# Vocoder model which takes spectrogram and produces actual audio\n",
                 "vocoder = nemo_tts.models.HifiGanModel.from_pretrained(model_name=\"tts_hifigan\").cuda()"
@@ -261,7 +261,7 @@
             "toc_visible": true
         },
         "kernelspec": {
-            "display_name": "Python 3",
+            "display_name": "Python 3 (ipykernel)",
             "language": "python",
             "name": "python3"
         },
@@ -275,9 +275,9 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.7.7"
+            "version": "3.8.12"
         }
     },
     "nbformat": 4,
-    "nbformat_minor": 1
-}
+    "nbformat_minor": 4
+}
diff --git a/tutorials/VoiceSwapSample.ipynb b/tutorials/VoiceSwapSample.ipynb
@@ -116,7 +116,7 @@
                 "punctuation = nemo_nlp.models.PunctuationCapitalizationModel.from_pretrained(model_name='punctuation_en_distilbert').cuda()\n",
                 "\n",
                 "# Spectrogram generator which takes text as an input and produces spectrogram\n",
-                "spectrogram_generator = nemo_tts.models.Tacotron2Model.from_pretrained(model_name=\"tts_en_tacotron2\").cuda()\n",
+                "spectrogram_generator = nemo_tts.models.FastPitchModel.from_pretrained(model_name=\"tts_en_fastpitch\").cuda()\n",
                 "\n",
                 "# Vocoder model which takes spectrogram and produces actual audio\n",
                 "vocoder = nemo_tts.models.HifiGanModel.from_pretrained(model_name=\"tts_hifigan\").cuda()"
@@ -306,7 +306,7 @@
             "toc_visible": true
         },
         "kernelspec": {
-            "display_name": "Python 3",
+            "display_name": "Python 3 (ipykernel)",
             "language": "python",
             "name": "python3"
         },
@@ -320,9 +320,9 @@
             "name": "python",
             "nbconvert_exporter": "python",
             "pygments_lexer": "ipython3",
-            "version": "3.8.5"
+            "version": "3.8.12"
         }
     },
     "nbformat": 4,
-    "nbformat_minor": 1
+    "nbformat_minor": 4
 }