Update app.py

app.py

@@ -12,13 +12,13 @@ from transformers import SpeechT5Processor, SpeechT5ForTextToSpeech, SpeechT5Hif
 import requests
 from requests.exceptions import Timeout
 
-checkpoint = "microsoft/speecht5_tts"
+checkpoint = "Edmon02/TTS_NB_2"
 processor = SpeechT5Processor.from_pretrained(checkpoint)
-model = SpeechT5ForTextToSpeech.from_pretrained("Edmon02/speecht5_finetuned_hy")
+model = SpeechT5ForTextToSpeech.from_pretrained(checkpoint)
 vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
 
 speaker_embeddings = {
-    "BDL": "cmu_us_bdl_arctic-wav-arctic_a0009.npy",
+    "BDL": "cmu_us_bdl_arctic-wav-arctic_a0004.npy",
 }
 
 def translate_text(text):
@@ -69,95 +69,11 @@ def process_text(text: str) -> str:
     processed_text = ' '.join(words)
     return processed_text
 
-replacements = [
-    ("՚", "?"),
-    ('՛', ""),
-    ('՝', ""),
-    ("«", "\""),
-    ("»", "\""),
-    ("՞", "?"),
-    ("ա", "a"),
-    ("բ", "b"),
-    ("գ", "g"),
-    ("դ", "d"),
-    ("զ", "z"),
-    ("է", "e"),
-    ("ը", "e'"),
-    ("թ",   "t'"),
-    ("ժ",	"jh"),
-    ("ի",	"i"),
-    ("լ",	"l"),
-    ("խ",	"kh"),
-    ("ծ",	"ts"),
-    ("կ",	"k"),
-    ("հ",	"h"),
-    ("ձ",	"dz"),
-    ("ղ",	"gh"),
-    ("ճ",	"ch"),
-    ("մ",	"m"),
-    ("յ",	"y"),
-    ("ն",	"n"),
-    ("շ",	"sh"),
-    ("չ",	"ch'"),
-    ("պ",	"p"),
-    ("ջ",	"j"),
-    ("ռ",	"r"),
-    ("ս",	"s"),
-    ("վ",	"v"),
-    ("տ",	"t"),
-    ("ր",	"r"),
-    ("ց",	"ts'"),
-    ("ւ",	""),
-    ("փ",	"p'"),
-    ("ք",	"k'"),
-    ("և",	"yev"),
-    ("օ",	"o"),
-    ("ֆ",	"f"),
-    ('։', "."),
-    ('–', "-"),
-    ('†', "e'"),
-]
-
-
-def cleanup_text(text):
-    
-    translator = str.maketrans("", "", string.punctuation)
-
-    text = text.translate(translator).lower()
-    text = text.lower()
-    
-    normalized_text = text
-
-    normalized_text = normalized_text.replace("ու", "u")
-    normalized_text = normalized_text.replace("եւ", "yev")
-    normalized_text = normalized_text.replace("եվ", "yev")
-
-    # Handle 'ո' at the beginning of a word
-    normalized_text = normalized_text.replace(" ո", " vo")
-
-    # Handle 'ո' in the middle of a word
-    normalized_text = normalized_text.replace("ո", "o")
-
-    # Handle 'ե' at the beginning of a word
-    normalized_text = normalized_text.replace(" ե", " ye")
-
-    # Handle 'ե' in the middle of a word
-    normalized_text = normalized_text.replace("ե", "e")
-
-    # Apply other replacements
-    for src, dst in replacements:
-        normalized_text = normalized_text.replace(src, dst)
-
-    inputs = normalized_text
-    return inputs
-
 def predict(text, speaker):
     if len(text.strip()) == 0:
         return (16000, np.zeros(0).astype(np.int16))
 
     text = process_text(text)
-    
-    text = cleanup_text(text)
 
     inputs = processor(text=text, return_tensors="pt")