Update app.py

app.py

@@ -24,6 +24,11 @@ from huggingface_hub import InferenceClient
 
 
 class VideoClassifier:
+    global audio_time , setup_time , caption_time , classification_time
+    audio_time = 0
+    setup_time = 0
+    caption_time = 0
+    classification_time = 0
     def __init__(self, no_of_frames, mode='interface',model='gemini'):
         
         self.no_of_frames = no_of_frames
@@ -105,6 +110,7 @@ class VideoClassifier:
         self.llm = HuggingFacePipeline(pipeline=self.generate_text)
 
     def audio_extraction(self,video_input):
+        global audio_time
         start_time_audio = time.time()
         print(f"Processing video: {video_input} with {self.no_of_frames} frames.")
         mp4_file = video_input
@@ -120,7 +126,8 @@ class VideoClassifier:
         transcript = result["text"]
         print("TRANSCRIPT",transcript)
         end_time_audio = time.time()
-        # print("TIME TAKEN FOR AUDIO CONVERSION (WHISPER)",end_time_audio-start_time_audio)
+        audio_time=end_time_audio-start_time_audio
+        # print("TIME TAKEN FOR AUDIO CONVERSION (WHISPER)",audio_time)
         
         return transcript
        
@@ -136,7 +143,7 @@ class VideoClassifier:
         return self.client(inputs, parameters)
     
     def classify_video(self,video_input):
-    
+        global classification_time , caption_time
         transcript=self.audio_extraction(video_input)
         start_time_caption = time.time()
         video = cv2.VideoCapture(video_input)
@@ -164,6 +171,7 @@ class VideoClassifier:
         video.release()
         cv2.destroyAllWindows()
         end_time_caption = time.time()
+        caption_time=end_time_caption-start_time_caption
         # print("TIME TAKEN FOR IMAGE CAPTIONING", end_time_caption-start_time_caption)
 
         start_time_generation = time.time()
@@ -312,14 +320,17 @@ class VideoClassifier:
 
         # return final_answer, first_video, second_video
         end_time_generation = time.time()
-        print("MODEL SETUP TIME",end_time_setup-start_time_setup)
-        print("TIME TAKEN FOR AUDIO CONVERSION (WHISPER)",end_time_audio-start_time_audio)
-        print("TIME TAKEN FOR IMAGE CAPTIONING", end_time_caption-start_time_caption)
-        print("TIME TAKEN FOR CLASS GENERATION",end_time_generation - start_time_generation)
+        classification_time = end_time_generation-start_time_generation
+        
+        print("MODEL SETUP TIME",setup_time)
+        print("TIME TAKEN FOR AUDIO CONVERSION (WHISPER)",audio_time)
+        print("TIME TAKEN FOR IMAGE CAPTIONING", caption_time)
+        print("TIME TAKEN FOR CLASS GENERATION",classification_time)
         return final_answer
     
     
     def save_model_choice(self,model_name):
+        global setup_time
         start_time_setup = time.time()
     
         self.model_name = model_name
@@ -330,7 +341,8 @@ class VideoClassifier:
             print("Setting up Gemini model for Class Selection")
             self.setup_gemini_model()
         end_time_setup = time.time()
-        # print("MODEL SETUP TIME",end_time_setup-start_time_setup)
+        setup_time=end_time_setup-start_time_setup
+        # print("MODEL SETUP TIME",setup_time)
         
         return "Model selected: " + model_name