Update audio2text/a2t.py

audio2text/a2t.py

@@ -3,6 +3,7 @@ import numpy as np
 from .init import pipe
 
 TASK = "transcribe"
+BATCH_SIZE = 16
 
 class A2T:
     def __init__(self, mic):
@@ -12,7 +13,7 @@ class A2T:
         if inputs is  None:
             print("Inputs None")
 
-        transcribed_text = pipe(inputs, generate_kwargs={"task": task}, return_timestamps=True)["text"]
+        transcribed_text = pipe(inputs, batch_size=BATCH_SIZE, generate_kwargs={"task": task}, return_timestamps=True)["text"]
         print("transcribed_text : ", transcribed_text)
         return transcribed_text
         
@@ -20,7 +21,7 @@ class A2T:
         try:
             if self.mic is not None:
                 chunk = self.mic
-                audio = np.array(chunk)
+                audio = np.frombuffer(chunk, dtype=np.uint8)
                 print("audio : ", audio)
             else:
                 return "please provide audio"