Upload ChunkFormer Classification Model

README.md

@@ -57,19 +57,27 @@ result = model.classify_audio(
     audio_path="path/to/your/audio.wav",
     chunk_size=-1,  # -1 for full attention
     left_context_size=-1,
-    right_context_size=-1,
-    return_probabilities=True
+    right_context_size=-1
 )
 
 print(result)
 # Output example:
 # {
-#   'gender': 0,
-#   'gender_probability': [0.95, 0.05],
-#   'dialect': 3,
-#   'dialect_probability': [0.1, 0.15, 0.05, 0.7],
-#   'emotion': 5,
-#   'emotion_probability': [0.05, 0.02, 0.03, 0.08, 0.02, 0.8]
+#   'gender': {
+#       'label': 'female',
+#       'label_id': 0,
+#       'prob': 0.95
+#   },
+#   'dialect': {
+#       'label': 'northern dialect',
+#       'label_id': 3,
+#       'prob': 0.70
+#   },
+#   'emotion': {
+#       'label': 'neutral',
+#       'label_id': 5,
+#       'prob': 0.80
+#   }
 # }
 ```
 
@@ -78,8 +86,7 @@ print(result)
 ```bash
 chunkformer-decode \
     --model_checkpoint khanhld/chunkformer-gender-emotion-dialect-age-classification \
-    --audio_file path/to/audio.wav \
-    --return_probabilities
+    --audio_file path/to/audio.wav
 ```
 
 ## Training