Merge pull request #67 from aierlma/master

Added the large-v3 option to some files, making it easier to run the repository directly

app.py

@@ -30,7 +30,7 @@ class App:
 
     @staticmethod
     def on_change_models(model_size: str):
-        translatable_model = ["large", "large-v1", "large-v2"]
+        translatable_model = ["large", "large-v1", "large-v2", "large-v3"]
         if model_size not in translatable_model:
             return gr.Checkbox.update(visible=False, value=False, interactive=False)
         else:
@@ -46,7 +46,7 @@ class App:
                     with gr.Row():
                         input_file = gr.Files(type="file", label="Upload File here")
                     with gr.Row():
-                        dd_model = gr.Dropdown(choices=self.whisper_inf.available_models, value="large-v2",
+                        dd_model = gr.Dropdown(choices=self.whisper_inf.available_models, value="large-v3",
                                                label="Model")
                         dd_lang = gr.Dropdown(choices=["Automatic Detection"] + self.whisper_inf.available_langs,
                                               value="Automatic Detection", label="Language")
@@ -84,7 +84,7 @@ class App:
                             tb_title = gr.Label(label="Youtube Title")
                             tb_description = gr.Textbox(label="Youtube Description", max_lines=15)
                     with gr.Row():
-                        dd_model = gr.Dropdown(choices=self.whisper_inf.available_models, value="large-v2",
+                        dd_model = gr.Dropdown(choices=self.whisper_inf.available_models, value="large-v3",
                                                label="Model")
                         dd_lang = gr.Dropdown(choices=["Automatic Detection"] + self.whisper_inf.available_langs,
                                               value="Automatic Detection", label="Language")
@@ -119,7 +119,7 @@ class App:
                     with gr.Row():
                         mic_input = gr.Microphone(label="Record with Mic", type="filepath", interactive=True)
                     with gr.Row():
-                        dd_model = gr.Dropdown(choices=self.whisper_inf.available_models, value="large-v2",
+                        dd_model = gr.Dropdown(choices=self.whisper_inf.available_models, value="large-v3",
                                                label="Model")
                         dd_lang = gr.Dropdown(choices=["Automatic Detection"] + self.whisper_inf.available_langs,
                                               value="Automatic Detection", label="Language")

modules/faster_whisper_inference.py

@@ -24,7 +24,7 @@ class FasterWhisperInference(BaseInterface):
         self.model = None
         self.available_models = whisper.available_models()
         self.available_langs = sorted(list(whisper.tokenizer.LANGUAGES.values()))
-        self.translatable_models = ["large", "large-v1", "large-v2"]
+        self.translatable_models = ["large", "large-v1", "large-v2", "large-v3"]
         self.device = "cuda" if torch.cuda.is_available() else "cpu"
         self.available_compute_types = ctranslate2.get_supported_compute_types("cuda") if self.device == "cuda" else ctranslate2.get_supported_compute_types("cpu")
         self.current_compute_type = "float16" if self.device == "cuda" else "float32"

modules/whisper_Inference.py

@@ -11,7 +11,7 @@ from .base_interface import BaseInterface
 from modules.subtitle_manager import get_srt, get_vtt, get_txt, write_file, safe_filename
 from modules.youtube_manager import get_ytdata, get_ytaudio
 
-DEFAULT_MODEL_SIZE = "large-v2"
+DEFAULT_MODEL_SIZE = "large-v3"
 
 
 class WhisperInference(BaseInterface):
@@ -327,7 +327,7 @@ class WhisperInference(BaseInterface):
         if lang == "Automatic Detection":
             lang = None
 
-        translatable_model = ["large", "large-v1", "large-v2"]
+        translatable_model = ["large", "large-v1", "large-v2", "large-v3"]
         segments_result = self.model.transcribe(audio=audio,
                                                 language=lang,
                                                 verbose=False,