|
|
|
@ -96,7 +96,7 @@ def start_server(config, args):
|
|
|
|
|
|
|
|
|
|
# prepare ASR inference handler
|
|
|
|
|
def file_to_transcript(filename):
|
|
|
|
|
feature = collate_fn.process_utterance(filename, "")
|
|
|
|
|
feature = test_loader.collate_fn.process_utterance(filename, "")
|
|
|
|
|
audio = np.array([feature[0]]).astype('float32') #[1, T, D]
|
|
|
|
|
audio_len = feature[0].shape[0]
|
|
|
|
|
audio_len = np.array([audio_len]).astype('int64') # [1]
|
|
|
|
|