Bug fix for res combine

shixian.shi

2024-01-25 292d34f2366d05a249088db844fc177e61708281

Bug fix for res combine

 examples/industrial_data_pretraining/seaco_paraformer/demo.py

@@ -15,6 +15,8 @@
                  # spk_model_revision="v2.0.2",
                  )

res = model.generate(input="https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav",
                     hotword='达摩院 魔搭')
res = model.generate(input="/Users/shixian/Downloads/output_16000.wav",
                     hotword='达摩院 魔搭',
                     # sentence_timestamp=True,
                    )
print(res)

 funasr/auto/auto_model.py

@@ -123,7 +123,6 @@
            self.preset_spk_num = kwargs.get("preset_spk_num", None)
            if self.preset_spk_num:
                logging.warning("Using preset speaker number: {}".format(self.preset_spk_num))
            logging.warning("Many to print when using speaker model...")
            
        self.kwargs = kwargs
        self.model = model
@@ -329,8 +328,6 @@
                speech_j, speech_lengths_j = slice_padding_audio_samples(speech, speech_lengths, sorted_data[beg_idx:end_idx])       
                results = self.inference(speech_j, input_len=None, model=model, kwargs=kwargs, disable_pbar=True, **cfg)
                if self.spk_model is not None:

                  
                    # compose vad segments: [[start_time_sec, end_time_sec, speech], [...]]
                    for _b in range(len(speech_j)):
                        vad_segments = [[sorted_data[beg_idx:end_idx][_b][0][0]/1000.0,
@@ -345,8 +342,6 @@
                if len(results) < 1:
                    continue
                results_sorted.extend(results)


            
            end_asr_total = time.time()
            time_escape_total_per_sample = end_asr_total - beg_asr_total
@@ -355,7 +350,6 @@
                                 f"time_speech_total_per_sample: {time_speech_total_per_sample: 0.3f}, "
                                 f"time_escape_total_per_sample: {time_escape_total_per_sample:0.3f}")
            

            restored_data = [0] * n
            for j in range(n):
                index = sorted_data[j][1]
@@ -378,7 +372,7 @@
                            result[k] = restored_data[j][k]
                        else:
                            result[k] = torch.cat([result[k], restored_data[j][k]], dim=0)
                    elif k == 'raw_text':
                    elif 'text' in k:
                        if k not in result:
                            result[k] = restored_data[j][k]
                        else:
@@ -393,8 +387,9 @@
            if self.punc_model is not None:
                self.punc_kwargs.update(cfg)
                punc_res = self.inference(result["text"], model=self.punc_model, kwargs=self.punc_kwargs, **cfg)
                import copy; raw_text = copy.copy(result["text"])
                result["text"] = punc_res[0]["text"]
                     
                
            # speaker embedding cluster after resorted
            if self.spk_model is not None:
                all_segments = sorted(all_segments, key=lambda x: x[0])
@@ -402,19 +397,24 @@
                labels = self.cb_model(spk_embedding.cpu(), oracle_num=self.preset_spk_num)
                del result['spk_embedding']
                sv_output = postprocess(all_segments, None, labels, spk_embedding.cpu())
                if self.spk_mode == 'vad_segment':
                if self.spk_mode == 'vad_segment':  # recover sentence_list
                    sentence_list = []
                    for res, vadsegment in zip(restored_data, vadsegments):
                        sentence_list.append({"start": vadsegment[0],\
                                                "end": vadsegment[1],
                                                "sentence": res['raw_text'],
                                                "timestamp": res['timestamp']})
                else: # punc_segment
                elif self.spk_mode == 'punc_segment':
                    sentence_list = timestamp_sentence(punc_res[0]['punc_array'], \
                                                        result['timestamp'], \
                                                        result['raw_text'])
                distribute_spk(sentence_list, sv_output)
                result['sentence_info'] = sentence_list
            elif kwargs.get("sentence_timestamp", False):
                sentence_list = timestamp_sentence(punc_res[0]['punc_array'], \
                                                        result['timestamp'], \
                                                        result['raw_text'])
                result['sentence_info'] = sentence_list
                    
            result["key"] = key
            results_ret_list.append(result)

 funasr/models/contextual_paraformer/model.py

@@ -65,11 +65,9 @@


        if bias_encoder_type == 'lstm':
            logging.warning("enable bias encoder sampling and contextual training")
            self.bias_encoder = torch.nn.LSTM(inner_dim, inner_dim, 1, batch_first=True, dropout=bias_encoder_dropout_rate)
            self.bias_embed = torch.nn.Embedding(self.vocab_size, inner_dim)
        elif bias_encoder_type == 'mean':
            logging.warning("enable bias encoder sampling and contextual training")
            self.bias_embed = torch.nn.Embedding(self.vocab_size, inner_dim)
        else:
            logging.error("Unsupport bias encoder type: {}".format(bias_encoder_type))

 funasr/models/seaco_paraformer/model.py

@@ -66,7 +66,6 @@
  
        # bias encoder
        if self.bias_encoder_type == 'lstm':
            logging.warning("enable bias encoder sampling and contextual training")
            self.bias_encoder = torch.nn.LSTM(self.inner_dim, 
                                              self.inner_dim, 
                                              2, 
@@ -79,7 +78,6 @@
                self.lstm_proj = None
            self.bias_embed = torch.nn.Embedding(self.vocab_size, self.inner_dim)
        elif self.bias_encoder_type == 'mean':
            logging.warning("enable bias encoder sampling and contextual training")
            self.bias_embed = torch.nn.Embedding(self.vocab_size, self.inner_dim)
        else:
            logging.error("Unsupport bias encoder type: {}".format(self.bias_encoder_type))