Add ASR with TTS Tutorial. Fix enhancer usage. (NVIDIA#6955)

* Add ASR with TTS Tutorial * Fix enhancer usage Signed-off-by: Vladimir Bataev <[email protected]> Signed-off-by: zhehuaichen <[email protected]>
zhehuaichen · Oct 4, 2023 · 2f89f46 · 2f89f46
1 parent cfd8357
commit 2f89f46
Show file tree

Hide file tree

Showing 3 changed files with 851 additions and 0 deletions.
diff --git a/docs/source/starthere/tutorials.rst b/docs/source/starthere/tutorials.rst
@@ -106,6 +106,9 @@ To run a tutorial:
    * - ASR
      - Multi-lingual ASR
      - `Multi-lingual ASR <https://colab.research.google.com/github/NVIDIA/NeMo/blob/stable/tutorials/asr/Multilang_ASR.ipynb>`_
+   * - ASR
+     - Hybrid ASR-TTS Models Tutorial
+     - `Multi-lingual ASR <https://colab.research.google.com/github/NVIDIA/NeMo/blob/stable/tutorials/asr/ASR_TTS_Tutorial.ipynb>`_
    * - NLP
      - Using Pretrained Language Models for Downstream Tasks
      - `Pretrained Language Models for Downstream Tasks <https://colab.research.google.com/github/NVIDIA/NeMo/blob/stable/tutorials/nlp/01_Pretrained_Language_Models_for_Downstream_Tasks.ipynb>`_

diff --git a/nemo/collections/asr/models/hybrid_asr_tts_models.py b/nemo/collections/asr/models/hybrid_asr_tts_models.py
@@ -311,8 +311,10 @@ def from_pretrained_models(
                 )
             )
         else:
+            cfg = copy.deepcopy(cfg)  # copy to avoid modifying original config
             cfg.tts_model_path = f"{tts_model_path}"
             cfg.asr_model_path = f"{asr_model_path}"
+            cfg.enhancer_model_path = f"{enhancer_model_path}" if enhancer_model_path is not None else None
         return ASRWithTTSModel(cfg, trainer=trainer)
 
     def __setattr__(self, name, value):