Interface improvement

accent_gradio.py

@@ -30,17 +30,12 @@ from models import *
 from utils import *
 from text_utils import TextCleaner
 textclenaer = TextCleaner()
-
 import phonemizer
-from cached_path import cached_path
-#from dp.phonemizer import Phonemizer
-
-
 
 
 # Global values
 sample_rate_value=24000
-original_voice_path = "original_voice.wav"
+original_voice_path = "ref_voice.wav"
 
 to_mel = torchaudio.transforms.MelSpectrogram(
     n_mels=80, n_fft=2048, win_length=1200, hop_length=300)
@@ -225,7 +220,9 @@ def record_speaker(audio):
 with gr.Blocks() as demo:
     gr.Markdown(""" # AccentCraft
     ### Transform your non-native accent into a native North American accent.
-    *For newcomers, kindly navigate to the "First-Time Users" section by clicking on the designated link.*
+    **This is an educational app designed to transform the speech of a non-native English speaker into a native American accent.**
+    **The tool aims to assist learners in accent reduction and pronunciation improvement.**
+    **For newcomers, kindly navigate to the *First-Time Users* section by clicking on the designated link.**
     """)
     with gr.Accordion("First-Time Users (Click Here):", open=False):
         gr.Markdown("""
@@ -264,16 +261,16 @@ with gr.Blocks() as demo:
     gr.Markdown(
         """
         ## Remark: 
-        - Longer sentences yield a more naturally flowing result.
-        Brief expressions like "Hi" or "How are you" may yield suboptimal outcomes.
-        - The model might occasionally produce noise or generate random speech. 
-        Consider re-recording or re-running for enhanced clarity and accuracy.
-        - By utilizing this application, you provide consent for your voice to
-         be synthesized by pre-trained models.
-        - If encountering an error, please try re-running or reloading the page.
-        - This app primarily functions as an educational tool for English learners. 
-        The author does not endorse or support any malicious or misuse of this application.
-        - The user acknowledges and agrees that the use of the software is at the user's sole risk.
+        - **Longer sentences yield a more naturally flowing result.
+        Brief expressions like "Hi" or "How are you" may yield suboptimal outcomes.**
+        - **The model might occasionally produce noise or generate random speech. 
+        Consider re-recording or re-running for enhanced clarity and accuracy.**
+        - **By utilizing this application, you provide consent for your voice to
+         be synthesized by pre-trained models.**
+        - **If encountering an error, please try re-running or reloading the page.**
+        - **This app primarily functions as an educational tool for English learners. 
+        The author does not endorse or support any malicious or misuse of this application.**
+        - **The user acknowledges and agrees that the use of the software is at the user's sole risk.**
         """)
 
 

requirements.txt

@@ -18,6 +18,4 @@ git+https://github.com/resemble-ai/monotonic_align.git
 gradio
 git+https://github.com/openai/whisper.git
 scipy
-phonemizer
-deep-phonemizer
-cached-path
+phonemizer