Create app.py

app.py

@@ -0,0 +1,133 @@
+from pprint import pprint
+
+import gradio as gr
+import torch
+
+# from IPython.display import Audio, display
+from omegaconf import OmegaConf
+
+torch.hub.download_url_to_file(
+    "https://raw.githubusercontent.com/snakers4/silero-models/master/models.yml",
+    "latest_silero_models.yml",
+    progress=False,
+)
+
+all_models = OmegaConf.load("latest_silero_models.yml")
+
+model_id = "v3_1_ru"
+device = torch.device("cpu")
+
+model, example_text = torch.hub.load(
+    repo_or_dir="snakers4/silero-models",
+    model="silero_tts",
+    language=language,
+    speaker=model_id,
+)
+model.to(device)  # gpu or cpu
+
+sample_rate = 48000
+speaker = "aidar"
+put_accent = True
+put_yo = True
+example_text = "В недрах тундры выдры в г+етрах т+ырят в вёдра ядра к+едров."
+
+models = list(all_models.tts_models.get(language).keys())
+
+model, example_text = torch.hub.load(
+    repo_or_dir='snakers4/silero-models',
+    model='silero_tts',
+    language='ru',
+    speaker=model_id
+)
+
+def change_language(language):
+    models = list(all_models.tts_models.get(language).keys())
+    return model_input.update(choices=models)
+
+def change_model(language, model_name):
+    model, example_text = torch.hub.load(
+        repo_or_dir='snakers4/silero-models',
+        model='silero_tts',
+        language=language,
+        speaker=model_name
+    )
+    
+    return speaker_input.update(choices=model.speakers)
+
+
+def generate_audio_by_text(text, text_type, speaker):
+    if text_type == 'SSML':
+        return model.save_wav(
+            ssml_text=text,
+            speaker=speaker,
+            sample_rate=sample_rate,
+            put_accent=put_accent,
+            put_yo=put_yo,
+        )
+    else:
+        return model.save_wav(
+            text=text,
+            speaker=speaker,
+            sample_rate=sample_rate,
+            put_accent=put_accent,
+            put_yo=put_yo,
+        )
+
+with gr.Blocks() as demo:
+    with gr.Row():
+        with gr.Column():
+            language_input = gr.Dropdown(
+                label="Language",
+                choices=list(all_models.tts_models.keys()),
+                value="ru",
+                interactive=True,
+            )
+
+            model_input = gr.Dropdown(
+                label="Model (based on selected language)",
+                value="v3_1_ru",
+                choices=models,
+                interactive=True,
+            )
+
+            speaker_input = gr.Dropdown(
+                label="Speaker (based on selected model)",
+                value="kseniya",
+                choices=model.speakers,
+                interactive=True,
+            )
+
+            text_input = gr.Textbox(
+                label="Text for generating",
+                value="В недрах тундры выдры в г+етрах т+ырят в вёдра +ядра к+едров.",
+                lines=5,
+                interactive=True,
+            )
+
+            text_type_input = gr.Radio(
+                label="Text type",
+                choices=["Common", "SSML"],
+                value="Common",
+                interactive=True,
+            )
+
+            language_input.change(change_language, inputs=language_input, outputs=model_input)
+            model_input.change(change_model, inputs=[language_input, model_input], outputs=speaker_input)
+
+        with gr.Column():
+            audio_output = gr.Audio(label="Output audio")
+            generate_btn = gr.Button(value="Generate", variant="primary")
+            generate_btn.click(
+                generate_audio_by_text,
+                inputs=[text_input, text_type_input, speaker_input],
+                outputs=audio_output,
+            )
+
+    gr.Markdown(
+        "This is a simple frontend for [silero](https://github.com/snakers4/silero-models) project (Text-To-Speech part only)."
+    )
+    gr.Markdown(
+        "You can check [official docs](https://github.com/snakers4/silero-models/wiki/SSML) to find information about SSML syntax."
+    )
+
+demo.launch()