optimization on preprocessing and other classif model

models/audio_classification__knn.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a54b7efa8866a57c8b8f45ca1b6b5d04caf83d0e79fb42cfa77ca6e375e1e5d7
+size 4811702

tasks/audio.py

@@ -60,12 +60,12 @@ async def evaluate_audio(request: AudioEvaluationRequest):
         features = []
         for row in dataset:
             # Load the audio file and resample it
-            target_sr = 25000
+            target_sr = 6000
             audio = row['audio']['array']
             audio = librosa.resample(audio, orig_sr=12000, target_sr=target_sr)
 
             # Extract MFCC features
-            mfccs = librosa.feature.mfcc(y=audio, sr=target_sr, n_mfcc=40)
+            mfccs = librosa.feature.mfcc(y=audio, sr=target_sr, n_mfcc=7)
             mfccs_scaled = np.mean(mfccs.T, axis=0)
 
             # Append features and labels
@@ -75,7 +75,7 @@ async def evaluate_audio(request: AudioEvaluationRequest):
 
     X_test = preprocess(test_dataset)
 
-    classification_model = joblib.load("./models/audio_classification_baseline.pkl")
+    classification_model = joblib.load("./models/audio_classification_knn.pkl")
 
     predictions = classification_model.predict(X_test)
     true_labels = test_dataset["label"]