Update app.py

app.py

@@ -29,8 +29,8 @@ class_options = {
 }
 
 pipe = pipeline("automatic-speech-recognition", model="openai/whisper-large")
-
-
+model = whisper.load_model("large")
+tokenizer = get_tokenizer("large")
 
 def slider_logic(slider):
     threshold = 0
@@ -54,7 +54,7 @@ def classify_toxicity(audio_file, selected_sounds, slider):
     # transcribed_text = pipe(audio_file)["text"]
 
     threshold = slider_logic(slider)
-    model = whisper.load_model("large")
+    # MODEL LINE model = whisper.load_model("large")
     # model = model_cache[model_name]
     # class_names = classify_anxiety.split(",")
     classify_anxiety = "misophonia"
@@ -66,7 +66,7 @@ def classify_toxicity(audio_file, selected_sounds, slider):
     class_names = class_str.split(",")
     print("class names ", class_names, "classify_anxiety ", classify_anxiety)
     
-    tokenizer = get_tokenizer("large")
+    # TOKENIZER LINE tokenizer = get_tokenizer("large")
     # tokenizer= WhisperTokenizer.from_pretrained("openai/whisper-large")
 
     internal_lm_average_logprobs = classify.calculate_internal_lm_average_logprobs(
@@ -99,7 +99,7 @@ def classify_toxicity(audio_file, selected_sounds, slider):
     highest_float = float(highest_score)
 
     if highest_score is not None and highest_float > threshold:
-        affirm = "Threshold Exceeded"
+        affirm = "Threshold Exceeded, initiate intervention"
     else:
         affirm = " "