Merge branch 'main' of github.com:alibaba-damo-academy/FunASR

add

egs_modelscope/asr/paraformer/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/infer.py

@@ -34,6 +34,6 @@ for sample_offset in range(0, speech_length, min(stride_size, speech_length - sa
     rec_result = inference_pipeline(audio_in=speech[sample_offset: sample_offset + stride_size],
                                     param_dict=param_dict)
     if len(rec_result) != 0:
-        final_result += " ".join(rec_result['text']) + " "
+        final_result += rec_result['text'] + " "
         print(rec_result)
 print(final_result)

egs_modelscope/asr/paraformer/speech_paraformer_asr_nat-zh-cn-16k-common-vocab8404-online/infer.py

@@ -34,6 +34,6 @@ for sample_offset in range(0, speech_length, min(stride_size, speech_length - sa
     rec_result = inference_pipeline(audio_in=speech[sample_offset: sample_offset + stride_size],
                                     param_dict=param_dict)
     if len(rec_result) != 0:
-        final_result += " ".join(rec_result['text']) + " "
+        final_result += rec_result['text'] + " "
         print(rec_result)
 print(final_result.strip())

funasr/bin/asr_inference_paraformer_streaming.py

@@ -553,12 +553,12 @@ def inference_modelscope(
                 asr_result = speech2text(cache, raw_inputs[:, sample_offset: sample_offset + stride_size], input_lens)
                 if len(asr_result) != 0: 
                     final_result += " ".join(asr_result) + " "
-            item = {'key': "utt", 'value': [final_result.strip()]}
+            item = {'key': "utt", 'value': final_result.strip()}
         else:
             input_lens = torch.tensor([raw_inputs.shape[1]])
             cache["encoder"]["is_final"] = is_final
             asr_result = speech2text(cache, raw_inputs, input_lens)
-            item = {'key': "utt", 'value': asr_result}
+            item = {'key': "utt", 'value': " ".join(asr_result)}
 
         asr_result_list.append(item)
         if is_final: