Update main.py

main.py

@@ -11,6 +11,7 @@ from uuid import uuid4
 # Load the necessary models and tokenizers
 model_path = "Vikhrmodels/salt-116k"
 tokenizer = AutoTokenizer.from_pretrained(model_path)
+print(tokenizer)
 # Специальные токены
 start_audio_token = "<soa>"
 end_audio_token = "<eoa>"
@@ -98,9 +99,14 @@ def decode_audio(tokens, quantizer, pad_tokens, n_original_tokens):
 
 
 # Inference functions
-def infer_text_to_audio(text, model, tokenizer, quantizer, max_seq_length=1024, top_k=20):
-    print(text)
+def infer_text_to_audio(text):
+    
+    max_seq_length=1024
+    top_k=20
+    
     print(type(tokenizer))
+    print(text)
+    
     text_tokenized = tokenizer(str(text), return_tensors="pt")
     text_input_tokens = text_tokenized["input_ids"].to(device)
 
@@ -117,7 +123,9 @@ def infer_text_to_audio(text, model, tokenizer, quantizer, max_seq_length=1024,
 
     return audio_signal
 
-def infer_audio_to_text(audio_path, model, tokenizer, quantizer, max_seq_length=1024, top_k=20):
+def infer_audio_to_text(audio_path):
+    max_seq_length=1024
+    top_k=20
     audio_data, sample_rate = torchaudio.load(audio_path)
 
     audio = audio_data.view(1, 1, -1).float().to(device)