Update app.py

app.py

@@ -1,63 +1,50 @@
+from jax._src.custom_derivatives import linear_call
 from transformers import pipeline
 import gradio as gr
 from pytube import YouTube
 
 pipe = pipeline(model="Mei000/whisper-small-sv-SE")
 
-class GradioInference():
-  def __init__(self):
-    self.yt = None
-    self.loaded_model = pipe
       
-  def __call__(self, link):
-    if self.yt is None:
-       self.yt = YouTube(link)
-    path = self.yt.streams.filter(only_audio=True)[0].download(filename="tmp.mp4")
-    results_text = self.loaded_model(path)["text"]
-    return results_text
+def link_transcribe(link):
+  path = YouTube(link).streams.filter(only_audio=True)[0].download(filename="tmp.mp4")
+  results_text = pipe(path)["text"]
+  return results_text
       
-  def transcribe(audio):
-    text = pipe(audio)["text"]
-    return text
+def transcribe(audio):
+  text = pipe(audio)["text"]
+  return text
 
-  def populate_metadata(self, link):
-    self.yt = YouTube(link)
-    return self.yt.thumbnail_url, self.yt.title
+def populate_metadata(link):
+  lin = YouTube(link)
+  return lin.thumbnail_url, lin.title
 
-gio = GradioInference()
-fn=transcribe,
-title="Whisper Small Swedishr",
-description="Demo for Swedish speech recognition using a fine-tuned Whisper small model.",
-
-block = gr.Blocks()
-with block:
-    gr.HTML(
-        """
-            <div style="text-align: center; max-width: 500px; margin: 0 auto;">
-              <div>
-                <h1>Youtube Whisperer</h1>
-              </div>
-              <p style="margin-bottom: 10px; font-size: 94%">
-                Speech to text transcription of Youtube videos using OpenAI's Whisper
-              </p>
-            </div>
-        """
-    )
-
-    
-    with gr.Group():
-        with gr.Box():
-          with gr.Row().style(equal_height=True):
-            inputs=gr.Audio(source="microphone", type="filepath")
-            out = gr.Textbox(label="Transcription", placeholder="Transcription Output", lines=10)
-          with gr.Row().style(equal_height=True):
-            link = gr.Textbox(label="YouTube Link")
-            title = gr.Label(label="Video Title")
-          with gr.Row().style(equal_height=True):
-            img = gr.Image(label="Thumbnail")
-            text = gr.Textbox(label="Transcription", placeholder="Transcription Output", lines=10)
-          with gr.Row().style(equal_height=True): 
-              btn = gr.Button("Transcribe")       
-          btn.click(gio, inputs=[link], outputs=[text])
-          link.change(gio.populate_metadata, inputs=[link], outputs=[img, title])
-block.launch()
+with gr.Blocks() as demo:
+    gr.Markdown("Whisper-Small Cantonese Recognition")
+    with gr.Row():
+        with gr.TabItem("Record from Microphone"):
+            record_file = gr.Audio(source="microphone", type="filepath",label="Record from microphone")
+            record_button = gr.Button("Submit")
+            record_outputs = [gr.Textbox(label="Recognized result from Microphone"),] 
+    """with gr.Row():
+        with gr.TabItem("Transcribe from youtube URL"):
+            url = gr.Text(max_lines=1, label="Transcribe from youtube URL")
+            youtube_button = gr.Button("Submit")
+            youtube_outputs = [
+                gr.Textbox(label="Recognized speech from youtube URL")
+            ]   """
+    with gr.Row().style(equal_height=True):
+        link = gr.Textbox(label="YouTube Link")
+        title = gr.Label(label="Video Title")
+    with gr.Row().style(equal_height=True):
+        img = gr.Image(label="Thumbnail")
+        youtube_outputs = [
+                gr.Textbox(label="Transcription", placeholder="Transcription Output", lines=10)
+            ]         
+    with gr.Row().style(equal_height=True):
+        youtube_button = gr.Button("Submit")
+  
+    record_button.click( fn=transcribe, inputs=record_file, outputs=record_outputs,)   
+    youtube_button.click( fn=link_transcribe, inputs=link, outputs=youtube_outputs,)
+    link.change(gio.populate_metadata, inputs=[link], outputs=[img, title])
+demo.launch()