| | |
|---|
| | | ark_reader = ReadHelper('ark:{}'.format(data_file)) |
|---|
| | | reader_list.append(ark_reader) |
|---|
| | | elif data_type == "text" or data_type == "sound" or data_type == 'text_hotword': |
|---|
| | | text_reader = open(data_file, "r") |
|---|
| | | text_reader = open(data_file, "r", encoding="utf-8") |
|---|
| | | reader_list.append(text_reader) |
|---|
| | | elif data_type == "none": |
|---|
| | | continue |
|---|
| | |
|---|
| | | # pre_prob = conf.get("pre_prob", 0) # unused yet |
|---|
| | | if pre_hwfile is not None: |
|---|
| | | pre_hwlist = [] |
|---|
| | | with open(pre_hwfile, 'r') as fin: |
|---|
| | | with open(pre_hwfile, 'r', encoding="utf-8") as fin: |
|---|
| | | for line in fin.readlines(): |
|---|
| | | pre_hwlist.append(line.strip()) |
|---|
| | | else: |
|---|
| | |
|---|
| | | mode=mode, |
|---|
| | | ) |
|---|
| | | |
|---|
| | | filter_conf = conf.get('filter_conf', {}) |
|---|
| | | filter_fn = partial(filter, **filter_conf) |
|---|
| | | dataset = FilterIterDataPipe(dataset, fn=filter_fn) |
|---|
| | | |
|---|
| | | if "text" in data_names: |
|---|
| | | vocab = {'vocab': dict, 'seg_dict': seg_dict, 'punc_dict': punc_dict, 'bpe_tokenizer': bpe_tokenizer, 'hw_config': hw_config} |
|---|
| | | tokenize_fn = partial(tokenize, **vocab) |
|---|
| | | dataset = MapperIterDataPipe(dataset, fn=tokenize_fn) |
|---|
| | | |
|---|
| | | filter_conf = conf.get('filter_conf', {}) |
|---|
| | | filter_fn = partial(filter, **filter_conf) |
|---|
| | | dataset = FilterIterDataPipe(dataset, fn=filter_fn) |
|---|
| | | |
|---|
| | | if shuffle: |
|---|
| | | buffer_conf = conf.get('shuffle_conf', {}) |
|---|
| | | buffer_size = buffer_conf['shuffle_size'] |
|---|
