Update lid.py

lid.py

@@ -16,15 +16,13 @@ with open(f"data/lid/all_langs.tsv") as f:
         iso, name = line.split(" ", 1)
         LID_LANGUAGES[iso] = name.strip()
 
-def identify_language(audio_source=None, microphone=None, file_upload=None):
-    audio_fp = file_upload if "upload" in str(audio_source or "").lower() else microphone
-    if audio_fp is None:
+def identify_language(audio=None):
+    if audio is None:
         return "ERROR: You have to either use the microphone or upload an audio file"
     
-    audio_samples = librosa.load(audio_fp, sr=LID_SAMPLING_RATE, mono=True)[0]
+    audio_samples = librosa.load(audio, sr=LID_SAMPLING_RATE, mono=True)[0]
     inputs = processor(audio_samples, sampling_rate=LID_SAMPLING_RATE, return_tensors="pt")
 
-    # Set device
     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
     model.to(device)
     inputs = inputs.to(device)
@@ -36,6 +34,8 @@ def identify_language(audio_source=None, microphone=None, file_upload=None):
     scores, indices = torch.topk(logit_lsm, 5, dim=-1)
     scores, indices = torch.exp(scores).to("cpu").tolist(), indices.to("cpu").tolist()
     iso2score = {model.config.id2label[int(i)]: s for s, i in zip(scores, indices)}
+    
     if max(iso2score.values()) < LID_THRESHOLD:
         return "Low confidence in the language identification predictions. Output is not shown!"
+    
     return {LID_LANGUAGES[iso]: score for iso, score in iso2score.items()}