HuyenNguyen
commited on
Commit
•
b84e0e6
1
Parent(s):
ba51655
Upload 10 files
Browse files- README.md +100 -0
- config.json +269 -0
- dict.txt +0 -0
- generation_config.json +9 -0
- preprocessor_config.json +10 -0
- pytorch_model.bin +3 -0
- sample_news.wav +0 -0
- sentencepiece.bpe.model +3 -0
- special_tokens_map.json +15 -0
- tokenizer_config.json +20 -0
README.md
ADDED
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc-by-nc-4.0
|
3 |
+
language:
|
4 |
+
- vi
|
5 |
+
---
|
6 |
+
|
7 |
+
### Vietnamese ASR sequence-to-sequence model. This model supports output normalizing text, labeling timestamps, and segmenting multiple speakers.
|
8 |
+
|
9 |
+
|
10 |
+
```python
|
11 |
+
# !pip install transformers, sentencepiece
|
12 |
+
|
13 |
+
from transformers import SpeechEncoderDecoderModel
|
14 |
+
from transformers import AutoFeatureExtractor, AutoTokenizer, GenerationConfig
|
15 |
+
import torchaudio
|
16 |
+
import torch
|
17 |
+
|
18 |
+
model_path = 'nguyenvulebinh/wav2vec2-bartpho'
|
19 |
+
model = SpeechEncoderDecoderModel.from_pretrained(model_path).eval()
|
20 |
+
feature_extractor = AutoFeatureExtractor.from_pretrained(model_path)
|
21 |
+
tokenizer = AutoTokenizer.from_pretrained(model_path)
|
22 |
+
if torch.cuda.is_available():
|
23 |
+
model = model.cuda()
|
24 |
+
|
25 |
+
|
26 |
+
def decode_tokens(token_ids, skip_special_tokens=True, time_precision=0.02):
|
27 |
+
timestamp_begin = tokenizer.vocab_size
|
28 |
+
outputs = [[]]
|
29 |
+
for token in token_ids:
|
30 |
+
if token >= timestamp_begin:
|
31 |
+
timestamp = f" |{(token - timestamp_begin) * time_precision:.2f}| "
|
32 |
+
outputs.append(timestamp)
|
33 |
+
outputs.append([])
|
34 |
+
else:
|
35 |
+
outputs[-1].append(token)
|
36 |
+
outputs = [
|
37 |
+
s if isinstance(s, str) else tokenizer.decode(s, skip_special_tokens=skip_special_tokens) for s in outputs
|
38 |
+
]
|
39 |
+
return "".join(outputs).replace("< |", "<|").replace("| >", "|>")
|
40 |
+
|
41 |
+
def decode_wav(audio_wavs, asr_model, prefix=""):
|
42 |
+
device = next(asr_model.parameters()).device
|
43 |
+
input_values = feature_extractor.pad(
|
44 |
+
[{"input_values": feature} for feature in audio_wavs],
|
45 |
+
padding=True,
|
46 |
+
max_length=None,
|
47 |
+
pad_to_multiple_of=None,
|
48 |
+
return_tensors="pt",
|
49 |
+
)
|
50 |
+
|
51 |
+
output_beam_ids = asr_model.generate(
|
52 |
+
input_values['input_values'].to(device),
|
53 |
+
attention_mask=input_values['attention_mask'].to(device),
|
54 |
+
decoder_input_ids=tokenizer.batch_encode_plus([prefix] * len(audio_wavs), return_tensors="pt")['input_ids'][..., :-1].to(device),
|
55 |
+
generation_config=GenerationConfig(decoder_start_token_id=tokenizer.bos_token_id),
|
56 |
+
max_length=250,
|
57 |
+
num_beams=25,
|
58 |
+
no_repeat_ngram_size=4,
|
59 |
+
num_return_sequences=1,
|
60 |
+
early_stopping=True,
|
61 |
+
return_dict_in_generate=True,
|
62 |
+
output_scores=True,
|
63 |
+
)
|
64 |
+
|
65 |
+
output_text = [decode_tokens(sequence) for sequence in output_beam_ids.sequences]
|
66 |
+
|
67 |
+
return output_text
|
68 |
+
|
69 |
+
|
70 |
+
# https://huggingface.co/nguyenvulebinh/wav2vec2-bartpho/resolve/main/sample_news.wav
|
71 |
+
print(decode_wav([torchaudio.load('sample_news.wav')[0].squeeze()], model))
|
72 |
+
|
73 |
+
# <|0.00| Gia đình cho biết, nhiều lần đã từng gọi điện báo chính quyền và lực lượng an ninh địa phương nhưng đều không có tác dụng |7.00|>
|
74 |
+
# <|8.14| Không ai giúp đỡ được mình một chút nào cả, nên là lúc đó là lúc tuyệt vọng nhất, nó tra tấn mình cực kỳ khổ, gây cái tâm lý ức chế rất là nhiều, rất là lớn |19.02|>
|
75 |
+
|
76 |
+
```
|
77 |
+
|
78 |
+
### Citation
|
79 |
+
|
80 |
+
This repository uses the idea from the following paper. Please cite the paper if this model is used to help produce published results or is incorporated into other software.
|
81 |
+
|
82 |
+
```text
|
83 |
+
@INPROCEEDINGS{10446589,
|
84 |
+
author={Nguyen, Thai-Binh and Waibel, Alexander},
|
85 |
+
booktitle={ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
|
86 |
+
title={Synthetic Conversations Improve Multi-Talker ASR},
|
87 |
+
year={2024},
|
88 |
+
volume={},
|
89 |
+
number={},
|
90 |
+
pages={10461-10465},
|
91 |
+
keywords={Systematics;Error analysis;Knowledge based systems;Oral communication;Signal processing;Data models;Acoustics;multi-talker;asr;synthetic conversation},
|
92 |
+
doi={10.1109/ICASSP48485.2024.10446589}
|
93 |
+
}
|
94 |
+
```
|
95 |
+
|
96 |
+
### Contact
|
97 |
+
|
98 | |
99 |
+
|
100 |
+
[![Follow](https://img.shields.io/twitter/follow/nguyenvulebinh?style=social)](https://twitter.com/intent/follow?screen_name=nguyenvulebinh)
|
config.json
ADDED
@@ -0,0 +1,269 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_commit_hash": null,
|
3 |
+
"architectures": [
|
4 |
+
"SpeechEncoderDecoderModel"
|
5 |
+
],
|
6 |
+
"decoder": {
|
7 |
+
"_name_or_path": "vinai/bartpho-syllable-base",
|
8 |
+
"activation_dropout": 0.0,
|
9 |
+
"activation_function": "gelu",
|
10 |
+
"add_cross_attention": true,
|
11 |
+
"architectures": [
|
12 |
+
"MBartModel"
|
13 |
+
],
|
14 |
+
"attention_dropout": 0.0,
|
15 |
+
"bad_words_ids": null,
|
16 |
+
"begin_suppress_tokens": null,
|
17 |
+
"bos_token_id": 0,
|
18 |
+
"chunk_size_feed_forward": 0,
|
19 |
+
"classifier_dropout": 0.0,
|
20 |
+
"cross_attention_hidden_size": null,
|
21 |
+
"d_model": 768,
|
22 |
+
"decoder_attention_heads": 12,
|
23 |
+
"decoder_ffn_dim": 3072,
|
24 |
+
"decoder_layerdrop": 0.0,
|
25 |
+
"decoder_layers": 6,
|
26 |
+
"decoder_start_token_id": 2,
|
27 |
+
"diversity_penalty": 0.0,
|
28 |
+
"do_sample": false,
|
29 |
+
"dropout": 0.1,
|
30 |
+
"early_stopping": false,
|
31 |
+
"encoder_attention_heads": 12,
|
32 |
+
"encoder_ffn_dim": 3072,
|
33 |
+
"encoder_layerdrop": 0.0,
|
34 |
+
"encoder_layers": 6,
|
35 |
+
"encoder_no_repeat_ngram_size": 0,
|
36 |
+
"eos_token_id": 2,
|
37 |
+
"exponential_decay_length_penalty": null,
|
38 |
+
"finetuning_task": null,
|
39 |
+
"forced_bos_token_id": null,
|
40 |
+
"forced_eos_token_id": 2,
|
41 |
+
"gradient_checkpointing": false,
|
42 |
+
"id2label": {
|
43 |
+
"0": "LABEL_0",
|
44 |
+
"1": "LABEL_1"
|
45 |
+
},
|
46 |
+
"init_std": 0.02,
|
47 |
+
"is_decoder": true,
|
48 |
+
"is_encoder_decoder": true,
|
49 |
+
"label2id": {
|
50 |
+
"LABEL_0": 0,
|
51 |
+
"LABEL_1": 1
|
52 |
+
},
|
53 |
+
"length_penalty": 1.0,
|
54 |
+
"max_length": 20,
|
55 |
+
"max_position_embeddings": 1024,
|
56 |
+
"min_length": 0,
|
57 |
+
"model_type": "mbart",
|
58 |
+
"no_repeat_ngram_size": 0,
|
59 |
+
"num_beam_groups": 1,
|
60 |
+
"num_beams": 1,
|
61 |
+
"num_hidden_layers": 6,
|
62 |
+
"num_return_sequences": 1,
|
63 |
+
"output_attentions": false,
|
64 |
+
"output_hidden_states": false,
|
65 |
+
"output_scores": false,
|
66 |
+
"pad_token_id": 1,
|
67 |
+
"prefix": null,
|
68 |
+
"problem_type": null,
|
69 |
+
"pruned_heads": {},
|
70 |
+
"remove_invalid_values": false,
|
71 |
+
"repetition_penalty": 1.0,
|
72 |
+
"return_dict": true,
|
73 |
+
"return_dict_in_generate": false,
|
74 |
+
"scale_embedding": false,
|
75 |
+
"sep_token_id": null,
|
76 |
+
"suppress_tokens": null,
|
77 |
+
"task_specific_params": null,
|
78 |
+
"temperature": 1.0,
|
79 |
+
"tf_legacy_loss": false,
|
80 |
+
"tie_encoder_decoder": false,
|
81 |
+
"tie_word_embeddings": true,
|
82 |
+
"tokenizer_class": "BartphoTokenizer",
|
83 |
+
"top_k": 50,
|
84 |
+
"top_p": 1.0,
|
85 |
+
"torch_dtype": "float32",
|
86 |
+
"torchscript": false,
|
87 |
+
"transformers_version": "4.30.2",
|
88 |
+
"typical_p": 1.0,
|
89 |
+
"use_bfloat16": false,
|
90 |
+
"use_cache": true,
|
91 |
+
"vocab_size": 41031
|
92 |
+
},
|
93 |
+
"encoder": {
|
94 |
+
"_name_or_path": "nguyenvulebinh/wav2vec2-large-vi",
|
95 |
+
"activation_dropout": 0.0,
|
96 |
+
"adapter_attn_dim": null,
|
97 |
+
"adapter_kernel_size": 3,
|
98 |
+
"adapter_stride": 2,
|
99 |
+
"add_adapter": true,
|
100 |
+
"add_cross_attention": false,
|
101 |
+
"apply_spec_augment": true,
|
102 |
+
"architectures": [
|
103 |
+
"Wav2Vec2ForPreTraining"
|
104 |
+
],
|
105 |
+
"attention_dropout": 0.1,
|
106 |
+
"bad_words_ids": null,
|
107 |
+
"begin_suppress_tokens": null,
|
108 |
+
"bos_token_id": 1,
|
109 |
+
"chunk_size_feed_forward": 0,
|
110 |
+
"classifier_proj_size": 256,
|
111 |
+
"codevector_dim": 768,
|
112 |
+
"contrastive_logits_temperature": 0.1,
|
113 |
+
"conv_bias": true,
|
114 |
+
"conv_dim": [
|
115 |
+
512,
|
116 |
+
512,
|
117 |
+
512,
|
118 |
+
512,
|
119 |
+
512,
|
120 |
+
512,
|
121 |
+
512
|
122 |
+
],
|
123 |
+
"conv_kernel": [
|
124 |
+
10,
|
125 |
+
3,
|
126 |
+
3,
|
127 |
+
3,
|
128 |
+
3,
|
129 |
+
2,
|
130 |
+
2
|
131 |
+
],
|
132 |
+
"conv_stride": [
|
133 |
+
5,
|
134 |
+
2,
|
135 |
+
2,
|
136 |
+
2,
|
137 |
+
2,
|
138 |
+
2,
|
139 |
+
2
|
140 |
+
],
|
141 |
+
"cross_attention_hidden_size": null,
|
142 |
+
"ctc_loss_reduction": "sum",
|
143 |
+
"ctc_zero_infinity": false,
|
144 |
+
"decoder_start_token_id": null,
|
145 |
+
"diversity_loss_weight": 0.1,
|
146 |
+
"diversity_penalty": 0.0,
|
147 |
+
"do_sample": false,
|
148 |
+
"do_stable_layer_norm": true,
|
149 |
+
"early_stopping": false,
|
150 |
+
"encoder_no_repeat_ngram_size": 0,
|
151 |
+
"eos_token_id": 2,
|
152 |
+
"exponential_decay_length_penalty": null,
|
153 |
+
"feat_extract_activation": "gelu",
|
154 |
+
"feat_extract_dropout": 0.0,
|
155 |
+
"feat_extract_norm": "layer",
|
156 |
+
"feat_proj_dropout": 0.1,
|
157 |
+
"feat_quantizer_dropout": 0.0,
|
158 |
+
"final_dropout": 0.0,
|
159 |
+
"finetuning_task": null,
|
160 |
+
"forced_bos_token_id": null,
|
161 |
+
"forced_eos_token_id": null,
|
162 |
+
"gradient_checkpointing": false,
|
163 |
+
"hidden_act": "gelu",
|
164 |
+
"hidden_dropout": 0.1,
|
165 |
+
"hidden_size": 1024,
|
166 |
+
"id2label": {
|
167 |
+
"0": "LABEL_0",
|
168 |
+
"1": "LABEL_1"
|
169 |
+
},
|
170 |
+
"initializer_range": 0.02,
|
171 |
+
"intermediate_size": 4096,
|
172 |
+
"is_decoder": false,
|
173 |
+
"is_encoder_decoder": false,
|
174 |
+
"label2id": {
|
175 |
+
"LABEL_0": 0,
|
176 |
+
"LABEL_1": 1
|
177 |
+
},
|
178 |
+
"layer_norm_eps": 1e-05,
|
179 |
+
"layerdrop": 0.1,
|
180 |
+
"length_penalty": 1.0,
|
181 |
+
"mask_channel_length": 10,
|
182 |
+
"mask_channel_min_space": 1,
|
183 |
+
"mask_channel_other": 0.0,
|
184 |
+
"mask_channel_prob": 0.0,
|
185 |
+
"mask_channel_selection": "static",
|
186 |
+
"mask_feature_length": 10,
|
187 |
+
"mask_feature_min_masks": 0,
|
188 |
+
"mask_feature_prob": 0.0,
|
189 |
+
"mask_time_length": 10,
|
190 |
+
"mask_time_min_masks": 2,
|
191 |
+
"mask_time_min_space": 1,
|
192 |
+
"mask_time_other": 0.0,
|
193 |
+
"mask_time_prob": 0.075,
|
194 |
+
"mask_time_selection": "static",
|
195 |
+
"max_length": 20,
|
196 |
+
"min_length": 0,
|
197 |
+
"model_type": "wav2vec2",
|
198 |
+
"no_repeat_ngram_size": 0,
|
199 |
+
"num_adapter_layers": 3,
|
200 |
+
"num_attention_heads": 16,
|
201 |
+
"num_beam_groups": 1,
|
202 |
+
"num_beams": 1,
|
203 |
+
"num_codevector_groups": 2,
|
204 |
+
"num_codevectors_per_group": 320,
|
205 |
+
"num_conv_pos_embedding_groups": 16,
|
206 |
+
"num_conv_pos_embeddings": 128,
|
207 |
+
"num_feat_extract_layers": 7,
|
208 |
+
"num_hidden_layers": 24,
|
209 |
+
"num_negatives": 100,
|
210 |
+
"num_return_sequences": 1,
|
211 |
+
"output_attentions": false,
|
212 |
+
"output_hidden_size": 1024,
|
213 |
+
"output_hidden_states": false,
|
214 |
+
"output_scores": false,
|
215 |
+
"pad_token_id": 0,
|
216 |
+
"prefix": null,
|
217 |
+
"problem_type": null,
|
218 |
+
"proj_codevector_dim": 768,
|
219 |
+
"pruned_heads": {},
|
220 |
+
"remove_invalid_values": false,
|
221 |
+
"repetition_penalty": 1.0,
|
222 |
+
"return_dict": true,
|
223 |
+
"return_dict_in_generate": false,
|
224 |
+
"sep_token_id": null,
|
225 |
+
"suppress_tokens": null,
|
226 |
+
"task_specific_params": null,
|
227 |
+
"tdnn_dilation": [
|
228 |
+
1,
|
229 |
+
2,
|
230 |
+
3,
|
231 |
+
1,
|
232 |
+
1
|
233 |
+
],
|
234 |
+
"tdnn_dim": [
|
235 |
+
512,
|
236 |
+
512,
|
237 |
+
512,
|
238 |
+
512,
|
239 |
+
1500
|
240 |
+
],
|
241 |
+
"tdnn_kernel": [
|
242 |
+
5,
|
243 |
+
3,
|
244 |
+
3,
|
245 |
+
1,
|
246 |
+
1
|
247 |
+
],
|
248 |
+
"temperature": 1.0,
|
249 |
+
"tf_legacy_loss": false,
|
250 |
+
"tie_encoder_decoder": false,
|
251 |
+
"tie_word_embeddings": true,
|
252 |
+
"tokenizer_class": null,
|
253 |
+
"top_k": 50,
|
254 |
+
"top_p": 1.0,
|
255 |
+
"torch_dtype": "float32",
|
256 |
+
"torchscript": false,
|
257 |
+
"transformers_version": "4.30.2",
|
258 |
+
"typical_p": 1.0,
|
259 |
+
"use_bfloat16": false,
|
260 |
+
"use_weighted_layer_sum": false,
|
261 |
+
"vocab_size": 96,
|
262 |
+
"xvector_output_dim": 512
|
263 |
+
},
|
264 |
+
"is_encoder_decoder": true,
|
265 |
+
"model_type": "speech-encoder-decoder",
|
266 |
+
"tie_word_embeddings": false,
|
267 |
+
"torch_dtype": "float32",
|
268 |
+
"transformers_version": null
|
269 |
+
}
|
dict.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
generation_config.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 0,
|
4 |
+
"decoder_start_token_id": 2,
|
5 |
+
"eos_token_id": 2,
|
6 |
+
"forced_eos_token_id": 2,
|
7 |
+
"pad_token_id": 1,
|
8 |
+
"transformers_version": "4.30.2"
|
9 |
+
}
|
preprocessor_config.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
4 |
+
"feature_size": 1,
|
5 |
+
"padding_side": "right",
|
6 |
+
"padding_value": 0.0,
|
7 |
+
"processor_class": "Wav2Vec2Processor",
|
8 |
+
"return_attention_mask": true,
|
9 |
+
"sampling_rate": 16000
|
10 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:27dc020061122a4c4d162f08ca41ba37d62525b82b4ab5309898c9a268f63f61
|
3 |
+
size 135
|
sample_news.wav
ADDED
Binary file (609 kB). View file
|
|
sentencepiece.bpe.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b8a54190d2b9256881ed34ab5428786629f929dd5a579350a6ef4735b86a9208
|
3 |
+
size 132
|
special_tokens_map.json
ADDED
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "<s>",
|
3 |
+
"cls_token": "<s>",
|
4 |
+
"eos_token": "</s>",
|
5 |
+
"mask_token": {
|
6 |
+
"content": "<mask>",
|
7 |
+
"lstrip": true,
|
8 |
+
"normalized": true,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"pad_token": "<pad>",
|
13 |
+
"sep_token": "</s>",
|
14 |
+
"unk_token": "<unk>"
|
15 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "<s>",
|
3 |
+
"clean_up_tokenization_spaces": true,
|
4 |
+
"cls_token": "<s>",
|
5 |
+
"eos_token": "</s>",
|
6 |
+
"mask_token": {
|
7 |
+
"__type": "AddedToken",
|
8 |
+
"content": "<mask>",
|
9 |
+
"lstrip": true,
|
10 |
+
"normalized": true,
|
11 |
+
"rstrip": false,
|
12 |
+
"single_word": false
|
13 |
+
},
|
14 |
+
"model_max_length": 1000000000000000019884624838656,
|
15 |
+
"pad_token": "<pad>",
|
16 |
+
"sep_token": "</s>",
|
17 |
+
"sp_model_kwargs": {},
|
18 |
+
"tokenizer_class": "BartphoTokenizer",
|
19 |
+
"unk_token": "<unk>"
|
20 |
+
}
|