Update audio_processing/config.py

audio_processing/config.py

@@ -1,8 +1,6 @@
 import torch
 
 from transformers import pipeline
-from parler_tts import ParlerTTSForConditionalGeneration
-from transformers import AutoTokenizer, AutoFeatureExtractor
 
 # ArticMonkey:19.03.24:1700 example of version name in plaintext  will be convert into hex using this site -> https://magictool.ai/tool/text-to-hex-converter/
 # Here ArticMonkey is name of version and rest of all is data and time
@@ -18,11 +16,21 @@ pipe = pipeline(
     chunk_length_s=30, 
 )
 
-checkpoint_parler = "parler-tts/parler_tts_mini_v0.1"
+# from parler_tts import ParlerTTSForConditionalGeneration
+# from transformers import AutoTokenizer, AutoFeatureExtractor
 
-model_parler = ParlerTTSForConditionalGeneration.from_pretrained(checkpoint_parler).to(device)
-tokenizer = AutoTokenizer.from_pretrained(checkpoint_parler)
-feature_extractor = AutoFeatureExtractor.from_pretrained(checkpoint_parler)
+# checkpoint_parler = "parler-tts/parler_tts_mini_v0.1"
 
-SAMPLE_RATE = feature_extractor.sampling_rate
-SEED = 42
+# model_parler = ParlerTTSForConditionalGeneration.from_pretrained(checkpoint_parler).to(device)
+# tokenizer = AutoTokenizer.from_pretrained(checkpoint_parler)
+# feature_extractor = AutoFeatureExtractor.from_pretrained(checkpoint_parler)
+
+# SAMPLE_RATE = feature_extractor.sampling_rate
+# SEED = 42
+
+from transformers import VitsModel, AutoTokenizer
+
+checkpoint_mms_tts_eng = "facebook/mms-tts-eng"
+
+model_mms_tts_eng = VitsModel.from_pretrained(checkpoint_mms_tts_eng)
+tokenizer_mms_tts_eng = AutoTokenizer.from_pretrained(checkpoint_mms_tts_eng)