removed timeout and return the beam size to 5 and set up an enviroment vaiarable

app.py

@@ -7,6 +7,11 @@ import csv
 import json
 import tempfile
 import torch
+import os
+
+# Set environment variables for gradio
+os.environ["COMMANDLINE_ARGS"] = "--no-gradio-queue"
+
 
 
 # Initiate checkpoints for model loading
@@ -28,7 +33,7 @@ llama_tokenizer = AutoTokenizer.from_pretrained(llama_checkpoint)
 # Function to transcribe audio
 def transcribe_audio(audio_file_path):
     try:
-        segments, info = whisper_model.transcribe(audio_file_path, beam_size=1)
+        segments, info = whisper_model.transcribe(audio_file_path, beam_size=5)
         text = "".join([segment.text for segment in segments])
         return text
     except Exception as e:
@@ -193,8 +198,7 @@ demo = gr.Interface(
         gr.Textbox(label="Plan of Action"),
     ],
     title="Medical Diagnostic Form Assistant",
-    description="Upload an audio file or record audio to generate a medical diagnostic form.",
-    max_timeout=300
+    description="Upload an audio file or record audio to generate a medical diagnostic form."
 )
 
 with demo: