Update app.py

app.py

@@ -33,14 +33,8 @@ pipe_dict = {
 }
 
 title =      """
-# Explore MMS finetuning
-## Or how to access truely multilingual TTS
-Massively Multilingual Speech (MMS) models are light-weight, low-latency TTS models based on the [VITS architecture](https://huggingface.co/docs/transformers/model_doc/vits).
-Meta's [MMS](https://arxiv.org/abs/2305.13516) project, aiming to provide speech technology across a diverse range of languages. You can find more details about the supported languages and their ISO 639-3 codes in the [MMS Language Coverage Overview](https://dl.fbaipublicfiles.com/mms/misc/language_coverage_mms.html),
-and see all MMS-TTS checkpoints on the Hugging Face Hub: [facebook/mms-tts](https://huggingface.co/models?sort=trending&search=facebook%2Fmms-tts).
-    
-Coupled with the right data and the right training recipe, you can get an excellent finetuned version of every MMS checkpoints in **20 minutes** with as little as **80 to 150 samples**.            
-Training recipe available in this [github repository](https://github.com/ylacombe/finetune-hf-vits)!
+Marathi Parkinson Enabler: Speaking is a big challenge during Parakinsons. Patients show slurred speech and cannot communicate effectively. 
+This is marathi text to speech model for parkinson users who want to communicate in Marathi. 
             """
 
 max_speakers = 1
@@ -129,47 +123,12 @@ with gr.Blocks(css=css) as demo_blocks:
     with gr.Accordion("Datasets and models details", open=False):
         gr.Markdown("""
         
-For each language, we used 100 to 150 samples of a single speaker to finetune the model.
-### Spanish
-* **Model**: [Spanish MMS TTS](https://huggingface.co/facebook/mms-tts-spa).
-* **Datasets**:
-    - [Chilean Spanish TTS dataset](https://huggingface.co/datasets/ylacombe/google-chilean-spanish).
-### Tamil
-* **Model**: [Tamil MMS TTS](https://huggingface.co/facebook/mms-tts-tam).
-* **Datasets**:
-    - [Tamil TTS dataset](https://huggingface.co/datasets/ylacombe/google-tamil).
-### Gujarati
-* **Model**: [Gujarati MMS TTS](https://huggingface.co/facebook/mms-tts-guj).
-* **Datasets**:
-    - [Gujarati TTS dataset](https://huggingface.co/datasets/ylacombe/google-gujarati).
 ### Marathi
 * **Model**: [Marathi MMS TTS](https://huggingface.co/facebook/mms-tts-mar).
 * **Datasets**:
-    - [Marathi TTS dataset](https://huggingface.co/datasets/ylacombe/google-chilean-marathi).
-### English
-* **Model**: [VITS-ljs](https://huggingface.co/kakao-enterprise/vits-ljs)
-* **Dataset**: [British Isles Accent](https://huggingface.co/datasets/ylacombe/english_dialects). For each accent, we used 100 to 150 samples of a single speaker to finetune [VITS-ljs](https://huggingface.co/kakao-enterprise/vits-ljs).
-                    
+    - [Marathi TTS dataset](https://huggingface.co/datasets/ylacombe/google-chilean-marathi). 
                     """) 
 
-    with gr.Accordion("Run VITS and MMS with transformers", open=False):
-        gr.Markdown(
-            """
-        ```bash
-        pip install transformers
-        ```
-        ```py
-        from transformers import pipeline
-        import scipy
-        pipe = pipeline("text-to-speech", model="kakao-enterprise/vits-ljs", device=0)
-        
-        results = pipe("A cinematic shot of a baby racoon wearing an intricate italian priest robe")
-        # write to a wav file
-        scipy.io.wavfile.write("audio_vits.wav", rate=results["sampling_rate"], data=results["audio"].squeeze())
-        ```
-        """
-        )
-
 
     language.change(lambda language: gr.Dropdown(
                     models_per_language[language],