| | |
|---|
| | | for char, timestamp in zip(new_char_list, timestamp_list): |
|---|
| | | res_str += "{} {} {};".format(char, str(timestamp[0]+0.0005)[:5], str(timestamp[1]+0.0005)[:5]) |
|---|
| | | res = [] |
|---|
| | | for char, timestamp in zip(char_list, timestamp_list): |
|---|
| | | for char, timestamp in zip(new_char_list, timestamp_list): |
|---|
| | | if char != '<sil>': |
|---|
| | | res.append([int(timestamp[0] * 1000), int(timestamp[1] * 1000)]) |
|---|
| | | return res_str, res |
|---|
| | |
|---|
| | | ) |
|---|
| | | if 'cuda' in device: |
|---|
| | | tp_model = tp_model.cuda() |
|---|
| | | |
|---|
| | | |
|---|
| | | frontend = None |
|---|
| | | if tp_train_args.frontend is not None: |
|---|
| | | frontend = WavFrontend(cmvn_file=timestamp_cmvn_file, **tp_train_args.frontend_conf) |
|---|
| | |
|---|
| | | token = speechtext2timestamp.converter.ids2tokens(batch['text'][batch_id]) |
|---|
| | | ts_str, ts_list = time_stamp_lfr6_advance(us_alphas[batch_id], us_cif_peak[batch_id], token) |
|---|
| | | logging.warning(ts_str) |
|---|
| | | tp_result_list.append({'text':"".join([i for i in token if i != '<sil>']), 'timestamp': ts_list}) |
|---|
| | | item = {'key': key, 'value': ts_str, 'timestamp':ts_list} |
|---|
| | | # tp_result_list.append({'text':"".join([i for i in token if i != '<sil>']), 'timestamp': ts_list}) |
|---|
| | | tp_result_list.append(item) |
|---|
| | | return tp_result_list |
|---|
| | | |
|---|
| | | return _forward |
|---|
