Update app.py

app.py

@@ -1,3 +1,4 @@
+import string
 import torch
 from transformers import pipeline
 from transformers.pipelines.audio_utils import ffmpeg_read
@@ -47,8 +48,9 @@ def transcribe(file, string, return_timestamps):
             for chunk in timestamps
         ]
         text = "\n".join(str(feature) for feature in timestamps)
+    text_nopunc = text.translate(str.maketrans('', '', string.punctuation))
     grade = ''
-    if text == string:
+    if text_nopunc.lower() == string.lower():
         grade = "good!"
     else:
         grade = "could use some work..."
@@ -64,14 +66,14 @@ mic_transcribe = gr.Interface(
         gr.inputs.Textbox(label="Word/Phrase"),
         gr.inputs.Checkbox(default=False, label="Return timestamps"),
     ],
-    outputs=["text", "text"],
+    outputs=[gr.Textbox(label="What I heard..."), gr.Textbox(label="Grade")],
     layout="vertical",
     theme="huggingface",
-    title="Whisper Demo: Transcribe Audio",
+    title="Vietnamese Pronounciation Checker",
     description=(
-        "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the"
-        f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
-        " of arbitrary length."
+        "This space transcribes Vietnamese words, phrases, and sentences via microphone or audio files then compares the user's text input to what the language model hears." 
+        "You will then be given a PASS/FAIL grade to tell you if your spoken audio matches the text you entered."
+        f"[{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) is a Vietnamese Speech-to-Text model and powers the analysis of the audio files."
     ),
     allow_flagging="never",
 )
@@ -83,14 +85,14 @@ file_transcribe = gr.Interface(
         gr.inputs.Textbox(label="Word/Phrase"),
         gr.inputs.Checkbox(default=False, label="Return timestamps"),
     ],
-    outputs=["text", "text"],
+    outputs=[gr.Textbox(label="What I heard..."), gr.Textbox(label="Grade")],
     layout="vertical",
     theme="huggingface",
-    title="Whisper Demo: Transcribe Audio",
+    title="Vietnamese Pronounciation Checker"",
     description=(
-        "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the"
-        f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
-        " of arbitrary length."
+        "This space transcribes Vietnamese words, phrases, and sentences via microphone or audio files then compares the user's text input to what the language model hears." 
+        "You will then be given a PASS/FAIL grade to tell you if your spoken audio matches the text you entered."
+        f"[{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) is a Vietnamese Speech-to-Text model and powers the analysis of the audio files."
     ),
     examples=[
         ["./example.flac", "transcribe", False],
@@ -101,6 +103,6 @@ file_transcribe = gr.Interface(
 )
 
 with demo:
-    gr.TabbedInterface([mic_transcribe, file_transcribe], ["Transcribe Microphone", "Transcribe Audio File"])
+    gr.TabbedInterface([mic_transcribe, file_transcribe], ["Pronounce via Microphone", "Pronounce via Audio File"])
 
 demo.launch(enable_queue=True)