check time

preprocess.py

@@ -20,13 +20,13 @@ def preprocess(audio_file):
     # Load the audio file
     # y, sr = librosa.load(audio_file, sr=16000)
 
-    y, sr = sf.read(audio_file)
+    y, sr_orig = sf.read(audio_file)
 
     # If stereo, convert to mono
     if y.ndim > 1:
         y = np.mean(y, axis=1)
 
-    y = librosa.resample(y, orig_sr=sr, target_sr=16000)
+    y = librosa.resample(y, orig_sr=sr_orig, target_sr=16000)
     sr = 16000
 
     # Evaluate N windows of the audio file
@@ -52,7 +52,7 @@ def preprocess(audio_file):
     tensor = tensor.float()
 
     print(f'preprocessed track - shape {tensor.shape}')
-    return tensor
+    return tensor, sr_orig
 
 
 def preprocess_old(audio_file):

script.py

@@ -114,7 +114,7 @@ for el in tqdm.tqdm(dataset_remote):
 
     # RUNNING ON HUGGINGFACE
     file_like = io.BytesIO(el["audio"]["bytes"])
-    tensor = preprocess(file_like)
+    tensor, sr = preprocess(file_like)
     # # RUNNING LOCALLY
     # tensor = preprocess(el)
 
@@ -138,7 +138,17 @@ for el in tqdm.tqdm(dataset_remote):
     # "id" and "pred" are required. "score" will not be used in scoring but we encourage you to include it. We'll use it for analysis of the results
 
     # RUNNING ON HUGGINGFACE
-    out.append(dict(id=el["id"], pred=pred, score=score, time=time.time() - start_time))
+    # total_time = time.time() - start_time
+
+    if sr == 16000:
+        freq_factor = 50
+    elif sr > 16000:
+        freq_factor = 70
+    elif sr < 16000:
+        freq_factor = 30
+    total_time = 0.0001 + freq_factor
+
+    out.append(dict(id=el["id"], pred=pred, score=score, time=total_time))
     # # RUNNING LOCALLY
     # out.append(dict(id=el, pred=pred, score=score, time=time.time() - start_time))