Update model
Browse files- README.md +856 -0
- exp/asr_train_asr_raw_en_word/35epoch.pth +3 -0
- exp/asr_train_asr_raw_en_word/RESULTS.md +30 -0
- exp/asr_train_asr_raw_en_word/config.yaml +757 -0
- exp/asr_train_asr_raw_en_word/images/acc.png +0 -0
- exp/asr_train_asr_raw_en_word/images/backward_time.png +0 -0
- exp/asr_train_asr_raw_en_word/images/cer.png +0 -0
- exp/asr_train_asr_raw_en_word/images/cer_ctc.png +0 -0
- exp/asr_train_asr_raw_en_word/images/forward_time.png +0 -0
- exp/asr_train_asr_raw_en_word/images/gpu_max_cached_mem_GB.png +0 -0
- exp/asr_train_asr_raw_en_word/images/iter_time.png +0 -0
- exp/asr_train_asr_raw_en_word/images/loss.png +0 -0
- exp/asr_train_asr_raw_en_word/images/loss_att.png +0 -0
- exp/asr_train_asr_raw_en_word/images/loss_ctc.png +0 -0
- exp/asr_train_asr_raw_en_word/images/optim0_lr0.png +0 -0
- exp/asr_train_asr_raw_en_word/images/optim_step_time.png +0 -0
- exp/asr_train_asr_raw_en_word/images/train_time.png +0 -0
- exp/asr_train_asr_raw_en_word/images/wer.png +0 -0
- meta.yaml +8 -0
README.md
ADDED
@@ -0,0 +1,856 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- espnet
|
4 |
+
- audio
|
5 |
+
- automatic-speech-recognition
|
6 |
+
language: en
|
7 |
+
datasets:
|
8 |
+
- slurp_mixture
|
9 |
+
license: cc-by-4.0
|
10 |
+
---
|
11 |
+
|
12 |
+
## ESPnet2 ASR model
|
13 |
+
|
14 |
+
### `espnet/Yen-Ju_Lu_spatilaizedslurp_asr_train_asr_conformer_transformer_valid.acc.best`
|
15 |
+
|
16 |
+
This model was trained by neillu23 using slurp_mixture recipe in [espnet](https://github.com/espnet/espnet/).
|
17 |
+
|
18 |
+
### Demo: How to use in ESPnet2
|
19 |
+
|
20 |
+
```bash
|
21 |
+
cd espnet
|
22 |
+
git checkout 0fae8113d99d092e7cbe4bcc48f9361e7012cff2
|
23 |
+
pip install -e .
|
24 |
+
cd egs2/slurp_mixture/asr1
|
25 |
+
./run.sh --skip_data_prep false --skip_train true --download_model espnet/Yen-Ju_Lu_spatilaizedslurp_asr_train_asr_conformer_transformer_valid.acc.best
|
26 |
+
```
|
27 |
+
|
28 |
+
<!-- Generated by scripts/utils/show_asr_result.sh -->
|
29 |
+
# RESULTS
|
30 |
+
## Environments
|
31 |
+
- date: `Tue Mar 29 04:17:37 UTC 2022`
|
32 |
+
- python version: `3.8.12 (default, Oct 12 2021, 13:49:34) [GCC 7.5.0]`
|
33 |
+
- espnet version: `espnet 0.10.7a1`
|
34 |
+
- pytorch version: `pytorch 1.9.0`
|
35 |
+
- Git hash: `0fae8113d99d092e7cbe4bcc48f9361e7012cff2`
|
36 |
+
- Commit date: `Thu Mar 24 07:54:19 2022 +0000`
|
37 |
+
|
38 |
+
## asr_train_asr_raw_en_word
|
39 |
+
### WER
|
40 |
+
|
41 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
42 |
+
|---|---|---|---|---|---|---|---|---|
|
43 |
+
|inference_asr_model_valid.acc.best/devel|8690|109017|63.9|21.1|15.0|2.0|38.1|75.4|
|
44 |
+
|inference_asr_model_valid.acc.best/test|6099|77315|69.0|17.4|13.5|1.8|32.8|68.9|
|
45 |
+
|inference_asr_model_valid.acc.best/test_ineube|6099|77315|77.8|12.0|10.2|1.5|23.6|59.4|
|
46 |
+
|inference_asr_model_valid.acc.best/test_qut|6099|77315|68.4|17.9|13.6|1.8|33.3|69.5|
|
47 |
+
|inference_asr_model_valid.acc.best/test_qut_ineube|6099|77315|78.0|11.9|10.2|1.4|23.4|59.3|
|
48 |
+
|
49 |
+
### CER
|
50 |
+
|
51 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
52 |
+
|---|---|---|---|---|---|---|---|---|
|
53 |
+
|inference_asr_model_valid.acc.best/devel|8690|513265|79.6|9.3|11.0|3.6|23.9|75.4|
|
54 |
+
|inference_asr_model_valid.acc.best/test|6099|362039|82.6|7.6|9.8|3.0|20.4|68.9|
|
55 |
+
|inference_asr_model_valid.acc.best/test_ineube|6099|362039|87.5|4.9|7.6|2.1|14.6|59.4|
|
56 |
+
|inference_asr_model_valid.acc.best/test_qut|6099|362039|82.3|7.8|9.9|3.1|20.8|69.5|
|
57 |
+
|inference_asr_model_valid.acc.best/test_qut_ineube|6099|362039|87.5|4.9|7.6|2.1|14.6|59.3|
|
58 |
+
|
59 |
+
## ASR config
|
60 |
+
|
61 |
+
<details><summary>expand</summary>
|
62 |
+
|
63 |
+
```
|
64 |
+
config: conf/train_asr.yaml
|
65 |
+
print_config: false
|
66 |
+
log_level: INFO
|
67 |
+
dry_run: false
|
68 |
+
iterator_type: sequence
|
69 |
+
output_dir: exp/asr_train_asr_raw_en_word
|
70 |
+
ngpu: 1
|
71 |
+
seed: 0
|
72 |
+
num_workers: 1
|
73 |
+
num_att_plot: 3
|
74 |
+
dist_backend: nccl
|
75 |
+
dist_init_method: env://
|
76 |
+
dist_world_size: 3
|
77 |
+
dist_rank: 0
|
78 |
+
local_rank: 0
|
79 |
+
dist_master_addr: localhost
|
80 |
+
dist_master_port: 35953
|
81 |
+
dist_launcher: null
|
82 |
+
multiprocessing_distributed: true
|
83 |
+
unused_parameters: false
|
84 |
+
sharded_ddp: false
|
85 |
+
cudnn_enabled: true
|
86 |
+
cudnn_benchmark: false
|
87 |
+
cudnn_deterministic: true
|
88 |
+
collect_stats: false
|
89 |
+
write_collected_feats: false
|
90 |
+
max_epoch: 50
|
91 |
+
patience: null
|
92 |
+
val_scheduler_criterion:
|
93 |
+
- valid
|
94 |
+
- loss
|
95 |
+
early_stopping_criterion:
|
96 |
+
- valid
|
97 |
+
- loss
|
98 |
+
- min
|
99 |
+
best_model_criterion:
|
100 |
+
- - valid
|
101 |
+
- acc
|
102 |
+
- max
|
103 |
+
keep_nbest_models: 10
|
104 |
+
nbest_averaging_interval: 0
|
105 |
+
grad_clip: 5.0
|
106 |
+
grad_clip_type: 2.0
|
107 |
+
grad_noise: false
|
108 |
+
accum_grad: 1
|
109 |
+
no_forward_run: false
|
110 |
+
resume: true
|
111 |
+
train_dtype: float32
|
112 |
+
use_amp: false
|
113 |
+
log_interval: null
|
114 |
+
use_matplotlib: true
|
115 |
+
use_tensorboard: true
|
116 |
+
use_wandb: false
|
117 |
+
wandb_project: null
|
118 |
+
wandb_id: null
|
119 |
+
wandb_entity: null
|
120 |
+
wandb_name: null
|
121 |
+
wandb_model_log_interval: -1
|
122 |
+
detect_anomaly: false
|
123 |
+
pretrain_path: null
|
124 |
+
init_param: []
|
125 |
+
ignore_init_mismatch: false
|
126 |
+
freeze_param: []
|
127 |
+
num_iters_per_epoch: null
|
128 |
+
batch_size: 48
|
129 |
+
valid_batch_size: null
|
130 |
+
batch_bins: 1000000
|
131 |
+
valid_batch_bins: null
|
132 |
+
train_shape_file:
|
133 |
+
- exp/asr_stats_raw_en_word/train/speech_shape
|
134 |
+
- exp/asr_stats_raw_en_word/train/text_shape.word
|
135 |
+
valid_shape_file:
|
136 |
+
- exp/asr_stats_raw_en_word/valid/speech_shape
|
137 |
+
- exp/asr_stats_raw_en_word/valid/text_shape.word
|
138 |
+
batch_type: folded
|
139 |
+
valid_batch_type: null
|
140 |
+
fold_length:
|
141 |
+
- 80000
|
142 |
+
- 150
|
143 |
+
sort_in_batch: descending
|
144 |
+
sort_batch: descending
|
145 |
+
multiple_iterator: false
|
146 |
+
chunk_length: 500
|
147 |
+
chunk_shift_ratio: 0.5
|
148 |
+
num_cache_chunks: 1024
|
149 |
+
train_data_path_and_name_and_type:
|
150 |
+
- - dump/raw/train/wav.scp
|
151 |
+
- speech
|
152 |
+
- sound
|
153 |
+
- - dump/raw/train/text
|
154 |
+
- text
|
155 |
+
- text
|
156 |
+
valid_data_path_and_name_and_type:
|
157 |
+
- - dump/raw/devel/wav.scp
|
158 |
+
- speech
|
159 |
+
- sound
|
160 |
+
- - dump/raw/devel/text
|
161 |
+
- text
|
162 |
+
- text
|
163 |
+
allow_variable_data_keys: false
|
164 |
+
max_cache_size: 0.0
|
165 |
+
max_cache_fd: 32
|
166 |
+
valid_max_cache_size: null
|
167 |
+
optim: adam
|
168 |
+
optim_conf:
|
169 |
+
lr: 0.0002
|
170 |
+
scheduler: warmuplr
|
171 |
+
scheduler_conf:
|
172 |
+
warmup_steps: 25000
|
173 |
+
token_list:
|
174 |
+
- <blank>
|
175 |
+
- <unk>
|
176 |
+
- ▁the
|
177 |
+
- s
|
178 |
+
- ▁to
|
179 |
+
- ▁i
|
180 |
+
- ▁me
|
181 |
+
- ▁you
|
182 |
+
- ▁what
|
183 |
+
- ▁a
|
184 |
+
- ▁is
|
185 |
+
- a
|
186 |
+
- ▁my
|
187 |
+
- ▁please
|
188 |
+
- y
|
189 |
+
- ''''
|
190 |
+
- ▁in
|
191 |
+
- ing
|
192 |
+
- ▁s
|
193 |
+
- e
|
194 |
+
- o
|
195 |
+
- ▁for
|
196 |
+
- i
|
197 |
+
- ▁on
|
198 |
+
- d
|
199 |
+
- t
|
200 |
+
- u
|
201 |
+
- er
|
202 |
+
- p
|
203 |
+
- ▁of
|
204 |
+
- es
|
205 |
+
- re
|
206 |
+
- l
|
207 |
+
- ▁it
|
208 |
+
- ▁p
|
209 |
+
- le
|
210 |
+
- ▁f
|
211 |
+
- ▁m
|
212 |
+
- ▁email
|
213 |
+
- ▁d
|
214 |
+
- m
|
215 |
+
- ▁c
|
216 |
+
- ▁b
|
217 |
+
- st
|
218 |
+
- r
|
219 |
+
- n
|
220 |
+
- ar
|
221 |
+
- ▁t
|
222 |
+
- ▁h
|
223 |
+
- b
|
224 |
+
- ▁that
|
225 |
+
- c
|
226 |
+
- ▁this
|
227 |
+
- h
|
228 |
+
- an
|
229 |
+
- email_query
|
230 |
+
- ▁play
|
231 |
+
- ▁re
|
232 |
+
- ▁do
|
233 |
+
- ▁can
|
234 |
+
- at
|
235 |
+
- ▁have
|
236 |
+
- g
|
237 |
+
- ▁from
|
238 |
+
- ▁and
|
239 |
+
- en
|
240 |
+
- email_sendemail
|
241 |
+
- ▁olly
|
242 |
+
- 'on'
|
243 |
+
- ▁new
|
244 |
+
- it
|
245 |
+
- qa_factoid
|
246 |
+
- calendar_set
|
247 |
+
- ▁any
|
248 |
+
- or
|
249 |
+
- ▁g
|
250 |
+
- ent
|
251 |
+
- ▁how
|
252 |
+
- ▁tell
|
253 |
+
- ch
|
254 |
+
- ▁not
|
255 |
+
- ▁about
|
256 |
+
- ▁at
|
257 |
+
- ate
|
258 |
+
- general_negate
|
259 |
+
- f
|
260 |
+
- ▁today
|
261 |
+
- ▁e
|
262 |
+
- ed
|
263 |
+
- ▁list
|
264 |
+
- ▁r
|
265 |
+
- in
|
266 |
+
- k
|
267 |
+
- ic
|
268 |
+
- social_post
|
269 |
+
- ▁are
|
270 |
+
- play_music
|
271 |
+
- general_quirky
|
272 |
+
- ▁l
|
273 |
+
- al
|
274 |
+
- v
|
275 |
+
- ▁n
|
276 |
+
- ▁be
|
277 |
+
- ▁an
|
278 |
+
- ▁st
|
279 |
+
- et
|
280 |
+
- ▁am
|
281 |
+
- general_praise
|
282 |
+
- ▁time
|
283 |
+
- weather_query
|
284 |
+
- ▁up
|
285 |
+
- ▁check
|
286 |
+
- calendar_query
|
287 |
+
- ▁w
|
288 |
+
- om
|
289 |
+
- ur
|
290 |
+
- ▁send
|
291 |
+
- ▁with
|
292 |
+
- ly
|
293 |
+
- w
|
294 |
+
- general_explain
|
295 |
+
- ad
|
296 |
+
- ▁th
|
297 |
+
- news_query
|
298 |
+
- ▁one
|
299 |
+
- ▁emails
|
300 |
+
- day
|
301 |
+
- ▁sh
|
302 |
+
- ce
|
303 |
+
- ▁
|
304 |
+
- ▁last
|
305 |
+
- ve
|
306 |
+
- ▁he
|
307 |
+
- z
|
308 |
+
- ▁ch
|
309 |
+
- ▁will
|
310 |
+
- ▁set
|
311 |
+
- ▁would
|
312 |
+
- ▁was
|
313 |
+
- x
|
314 |
+
- general_repeat
|
315 |
+
- ▁add
|
316 |
+
- ▁again
|
317 |
+
- ou
|
318 |
+
- ▁ex
|
319 |
+
- is
|
320 |
+
- ct
|
321 |
+
- general_affirm
|
322 |
+
- general_confirm
|
323 |
+
- ▁song
|
324 |
+
- ▁next
|
325 |
+
- ▁j
|
326 |
+
- ▁meeting
|
327 |
+
- um
|
328 |
+
- ation
|
329 |
+
- ▁turn
|
330 |
+
- ▁did
|
331 |
+
- if
|
332 |
+
- ▁alarm
|
333 |
+
- am
|
334 |
+
- ▁like
|
335 |
+
- datetime_query
|
336 |
+
- ter
|
337 |
+
- ▁remind
|
338 |
+
- ▁o
|
339 |
+
- qa_definition
|
340 |
+
- ▁said
|
341 |
+
- ▁calendar
|
342 |
+
- ll
|
343 |
+
- se
|
344 |
+
- ers
|
345 |
+
- ▁pr
|
346 |
+
- th
|
347 |
+
- ▁get
|
348 |
+
- our
|
349 |
+
- ▁need
|
350 |
+
- ▁all
|
351 |
+
- ot
|
352 |
+
- ▁want
|
353 |
+
- ▁off
|
354 |
+
- and
|
355 |
+
- ▁right
|
356 |
+
- ▁de
|
357 |
+
- ▁tr
|
358 |
+
- ut
|
359 |
+
- general_dontcare
|
360 |
+
- as
|
361 |
+
- ▁week
|
362 |
+
- ▁tweet
|
363 |
+
- ight
|
364 |
+
- ir
|
365 |
+
- ▁your
|
366 |
+
- ▁event
|
367 |
+
- ▁news
|
368 |
+
- ▁se
|
369 |
+
- ay
|
370 |
+
- ion
|
371 |
+
- ▁com
|
372 |
+
- ▁there
|
373 |
+
- ▁ye
|
374 |
+
- ▁weather
|
375 |
+
- un
|
376 |
+
- ▁confirm
|
377 |
+
- ld
|
378 |
+
- calendar_remove
|
379 |
+
- ▁y
|
380 |
+
- ▁lights
|
381 |
+
- ▁more
|
382 |
+
- ▁v
|
383 |
+
- play_radio
|
384 |
+
- ▁does
|
385 |
+
- ▁po
|
386 |
+
- ▁now
|
387 |
+
- id
|
388 |
+
- email_querycontact
|
389 |
+
- ▁show
|
390 |
+
- ▁could
|
391 |
+
- ery
|
392 |
+
- op
|
393 |
+
- ▁day
|
394 |
+
- ▁pm
|
395 |
+
- ▁music
|
396 |
+
- ▁tomorrow
|
397 |
+
- ▁train
|
398 |
+
- ▁u
|
399 |
+
- ine
|
400 |
+
- ▁or
|
401 |
+
- ange
|
402 |
+
- qa_currency
|
403 |
+
- ice
|
404 |
+
- ▁contact
|
405 |
+
- ▁just
|
406 |
+
- ▁jo
|
407 |
+
- ▁think
|
408 |
+
- qa_stock
|
409 |
+
- end
|
410 |
+
- ss
|
411 |
+
- ber
|
412 |
+
- ▁tw
|
413 |
+
- ▁command
|
414 |
+
- ▁make
|
415 |
+
- ▁no
|
416 |
+
- ▁mo
|
417 |
+
- pe
|
418 |
+
- ▁find
|
419 |
+
- general_commandstop
|
420 |
+
- ▁when
|
421 |
+
- social_query
|
422 |
+
- ▁so
|
423 |
+
- ong
|
424 |
+
- ▁co
|
425 |
+
- ant
|
426 |
+
- ow
|
427 |
+
- q
|
428 |
+
- ▁much
|
429 |
+
- ▁where
|
430 |
+
- ue
|
431 |
+
- ul
|
432 |
+
- ri
|
433 |
+
- ake
|
434 |
+
- ap
|
435 |
+
- ▁start
|
436 |
+
- ▁mar
|
437 |
+
- ▁by
|
438 |
+
- one
|
439 |
+
- ▁know
|
440 |
+
- ▁wor
|
441 |
+
- oo
|
442 |
+
- ▁give
|
443 |
+
- ▁let
|
444 |
+
- ▁events
|
445 |
+
- der
|
446 |
+
- ▁ro
|
447 |
+
- ▁pl
|
448 |
+
- play_podcasts
|
449 |
+
- art
|
450 |
+
- us
|
451 |
+
- ▁work
|
452 |
+
- ▁current
|
453 |
+
- ol
|
454 |
+
- cooking_recipe
|
455 |
+
- nt
|
456 |
+
- ▁correct
|
457 |
+
- transport_query
|
458 |
+
- ia
|
459 |
+
- ▁stock
|
460 |
+
- ▁br
|
461 |
+
- ive
|
462 |
+
- ▁app
|
463 |
+
- ▁two
|
464 |
+
- ▁latest
|
465 |
+
- lists_query
|
466 |
+
- recommendation_events
|
467 |
+
- ab
|
468 |
+
- ▁go
|
469 |
+
- ▁but
|
470 |
+
- ook
|
471 |
+
- ▁some
|
472 |
+
- ke
|
473 |
+
- alarm_set
|
474 |
+
- play_audiobook
|
475 |
+
- ▁k
|
476 |
+
- ▁response
|
477 |
+
- ▁wr
|
478 |
+
- cast
|
479 |
+
- ▁open
|
480 |
+
- ▁cle
|
481 |
+
- ▁done
|
482 |
+
- ▁got
|
483 |
+
- ▁ca
|
484 |
+
- ite
|
485 |
+
- ase
|
486 |
+
- ▁thank
|
487 |
+
- iv
|
488 |
+
- ag
|
489 |
+
- ah
|
490 |
+
- ▁answer
|
491 |
+
- ie
|
492 |
+
- ▁five
|
493 |
+
- ▁book
|
494 |
+
- ▁rec
|
495 |
+
- ore
|
496 |
+
- ▁john
|
497 |
+
- ist
|
498 |
+
- ment
|
499 |
+
- ▁appreci
|
500 |
+
- ▁fri
|
501 |
+
- ack
|
502 |
+
- ▁remove
|
503 |
+
- ated
|
504 |
+
- ock
|
505 |
+
- ree
|
506 |
+
- j
|
507 |
+
- ▁good
|
508 |
+
- ▁many
|
509 |
+
- orn
|
510 |
+
- fe
|
511 |
+
- ▁radio
|
512 |
+
- ▁we
|
513 |
+
- int
|
514 |
+
- ▁facebook
|
515 |
+
- ▁cl
|
516 |
+
- ▁sev
|
517 |
+
- ▁schedule
|
518 |
+
- ard
|
519 |
+
- ▁per
|
520 |
+
- ▁li
|
521 |
+
- ▁going
|
522 |
+
- nd
|
523 |
+
- ain
|
524 |
+
- recommendation_locations
|
525 |
+
- ▁post
|
526 |
+
- lists_createoradd
|
527 |
+
- ff
|
528 |
+
- ▁su
|
529 |
+
- red
|
530 |
+
- iot_hue_lightoff
|
531 |
+
- lists_remove
|
532 |
+
- ▁ar
|
533 |
+
- een
|
534 |
+
- ▁say
|
535 |
+
- ro
|
536 |
+
- ▁volume
|
537 |
+
- ▁le
|
538 |
+
- ▁reply
|
539 |
+
- ▁complaint
|
540 |
+
- ▁delete
|
541 |
+
- ▁out
|
542 |
+
- lly
|
543 |
+
- ame
|
544 |
+
- ▁ne
|
545 |
+
- ▁detail
|
546 |
+
- ▁if
|
547 |
+
- im
|
548 |
+
- ▁happ
|
549 |
+
- orr
|
550 |
+
- ich
|
551 |
+
- em
|
552 |
+
- ▁ev
|
553 |
+
- ction
|
554 |
+
- ▁dollar
|
555 |
+
- ▁as
|
556 |
+
- alarm_query
|
557 |
+
- audio_volume_mute
|
558 |
+
- ac
|
559 |
+
- music_query
|
560 |
+
- ▁mon
|
561 |
+
- ther
|
562 |
+
- ▁thanks
|
563 |
+
- cel
|
564 |
+
- ▁who
|
565 |
+
- ave
|
566 |
+
- ▁service
|
567 |
+
- ▁mail
|
568 |
+
- ▁hear
|
569 |
+
- ty
|
570 |
+
- de
|
571 |
+
- ▁si
|
572 |
+
- ▁wh
|
573 |
+
- ood
|
574 |
+
- ell
|
575 |
+
- ▁con
|
576 |
+
- icket
|
577 |
+
- ▁once
|
578 |
+
- ound
|
579 |
+
- ▁don
|
580 |
+
- ▁loc
|
581 |
+
- ▁light
|
582 |
+
- ▁birthday
|
583 |
+
- ▁inf
|
584 |
+
- ffe
|
585 |
+
- ▁has
|
586 |
+
- ▁playlist
|
587 |
+
- ort
|
588 |
+
- el
|
589 |
+
- ening
|
590 |
+
- ▁us
|
591 |
+
- ▁un
|
592 |
+
- own
|
593 |
+
- ▁inc
|
594 |
+
- ai
|
595 |
+
- ▁speak
|
596 |
+
- age
|
597 |
+
- ▁mess
|
598 |
+
- ast
|
599 |
+
- ci
|
600 |
+
- ver
|
601 |
+
- ▁ten
|
602 |
+
- ▁underst
|
603 |
+
- gh
|
604 |
+
- audio_volume_up
|
605 |
+
- ome
|
606 |
+
- transport_ticket
|
607 |
+
- ind
|
608 |
+
- iot_hue_lightchange
|
609 |
+
- iot_coffee
|
610 |
+
- pp
|
611 |
+
- ▁res
|
612 |
+
- plain
|
613 |
+
- io
|
614 |
+
- lar
|
615 |
+
- takeaway_query
|
616 |
+
- ge
|
617 |
+
- takeaway_order
|
618 |
+
- email_addcontact
|
619 |
+
- play_game
|
620 |
+
- ak
|
621 |
+
- ▁fa
|
622 |
+
- transport_traffic
|
623 |
+
- music_likeness
|
624 |
+
- ▁rep
|
625 |
+
- act
|
626 |
+
- ust
|
627 |
+
- transport_taxi
|
628 |
+
- iot_hue_lightdim
|
629 |
+
- ▁mu
|
630 |
+
- ▁ti
|
631 |
+
- ick
|
632 |
+
- ▁ha
|
633 |
+
- ould
|
634 |
+
- general_joke
|
635 |
+
- '1'
|
636 |
+
- qa_maths
|
637 |
+
- ▁lo
|
638 |
+
- iot_cleaning
|
639 |
+
- ill
|
640 |
+
- her
|
641 |
+
- iot_hue_lightup
|
642 |
+
- pl
|
643 |
+
- '2'
|
644 |
+
- alarm_remove
|
645 |
+
- orrect
|
646 |
+
- ▁cont
|
647 |
+
- mail
|
648 |
+
- out
|
649 |
+
- audio_volume_down
|
650 |
+
- book
|
651 |
+
- ail
|
652 |
+
- recommendation_movies
|
653 |
+
- ck
|
654 |
+
- ▁man
|
655 |
+
- ▁mus
|
656 |
+
- ▁che
|
657 |
+
- me
|
658 |
+
- ume
|
659 |
+
- ▁answ
|
660 |
+
- datetime_convert
|
661 |
+
- ▁late
|
662 |
+
- iot_wemo_on
|
663 |
+
- ▁twe
|
664 |
+
- music_settings
|
665 |
+
- iot_wemo_off
|
666 |
+
- orre
|
667 |
+
- ith
|
668 |
+
- ▁tom
|
669 |
+
- ▁fr
|
670 |
+
- ere
|
671 |
+
- ▁ad
|
672 |
+
- xt
|
673 |
+
- ▁ab
|
674 |
+
- ank
|
675 |
+
- general_greet
|
676 |
+
- now
|
677 |
+
- ▁meet
|
678 |
+
- ▁curre
|
679 |
+
- ▁respon
|
680 |
+
- ▁ag
|
681 |
+
- audio_volume_other
|
682 |
+
- ink
|
683 |
+
- ▁spe
|
684 |
+
- iot_hue_lighton
|
685 |
+
- ght
|
686 |
+
- ▁rem
|
687 |
+
- '?'
|
688 |
+
- urn
|
689 |
+
- ▁op
|
690 |
+
- ▁complain
|
691 |
+
- ▁comm
|
692 |
+
- let
|
693 |
+
- music_dislikeness
|
694 |
+
- ove
|
695 |
+
- ▁sch
|
696 |
+
- ather
|
697 |
+
- ▁rad
|
698 |
+
- edule
|
699 |
+
- ▁under
|
700 |
+
- lease
|
701 |
+
- ▁bir
|
702 |
+
- erv
|
703 |
+
- ▁birth
|
704 |
+
- ▁face
|
705 |
+
- ▁cur
|
706 |
+
- sw
|
707 |
+
- ▁serv
|
708 |
+
- ek
|
709 |
+
- aid
|
710 |
+
- '9'
|
711 |
+
- ▁vol
|
712 |
+
- edu
|
713 |
+
- '5'
|
714 |
+
- cooking_query
|
715 |
+
- lete
|
716 |
+
- ▁joh
|
717 |
+
- ▁det
|
718 |
+
- firm
|
719 |
+
- nder
|
720 |
+
- '0'
|
721 |
+
- _
|
722 |
+
- irm
|
723 |
+
- '8'
|
724 |
+
- '&'
|
725 |
+
- list
|
726 |
+
- pon
|
727 |
+
- qa_query
|
728 |
+
- '7'
|
729 |
+
- '3'
|
730 |
+
- '-'
|
731 |
+
- N
|
732 |
+
- A
|
733 |
+
- M
|
734 |
+
- E
|
735 |
+
- ']'
|
736 |
+
- '['
|
737 |
+
- ':'
|
738 |
+
- reci
|
739 |
+
- ▁doll
|
740 |
+
- <sos/eos>
|
741 |
+
init: null
|
742 |
+
input_size: null
|
743 |
+
ctc_conf:
|
744 |
+
dropout_rate: 0.0
|
745 |
+
ctc_type: builtin
|
746 |
+
reduce: true
|
747 |
+
ignore_nan_grad: true
|
748 |
+
joint_net_conf: null
|
749 |
+
model_conf:
|
750 |
+
ctc_weight: 0.3
|
751 |
+
lsm_weight: 0.1
|
752 |
+
length_normalized_loss: false
|
753 |
+
extract_feats_in_collect_stats: false
|
754 |
+
use_preprocessor: true
|
755 |
+
token_type: word
|
756 |
+
bpemodel: null
|
757 |
+
non_linguistic_symbols: null
|
758 |
+
cleaner: null
|
759 |
+
g2p: null
|
760 |
+
speech_volume_normalize: null
|
761 |
+
rir_scp: null
|
762 |
+
rir_apply_prob: 1.0
|
763 |
+
noise_scp: null
|
764 |
+
noise_apply_prob: 1.0
|
765 |
+
noise_db_range: '13_15'
|
766 |
+
frontend: default
|
767 |
+
frontend_conf:
|
768 |
+
fs: 16k
|
769 |
+
specaug: specaug
|
770 |
+
specaug_conf:
|
771 |
+
apply_time_warp: true
|
772 |
+
time_warp_window: 5
|
773 |
+
time_warp_mode: bicubic
|
774 |
+
apply_freq_mask: true
|
775 |
+
freq_mask_width_range:
|
776 |
+
- 0
|
777 |
+
- 30
|
778 |
+
num_freq_mask: 2
|
779 |
+
apply_time_mask: true
|
780 |
+
time_mask_width_range:
|
781 |
+
- 0
|
782 |
+
- 40
|
783 |
+
num_time_mask: 2
|
784 |
+
normalize: utterance_mvn
|
785 |
+
normalize_conf: {}
|
786 |
+
preencoder: null
|
787 |
+
preencoder_conf: {}
|
788 |
+
encoder: conformer
|
789 |
+
encoder_conf:
|
790 |
+
output_size: 512
|
791 |
+
attention_heads: 8
|
792 |
+
linear_units: 2048
|
793 |
+
num_blocks: 12
|
794 |
+
dropout_rate: 0.1
|
795 |
+
positional_dropout_rate: 0.1
|
796 |
+
attention_dropout_rate: 0.1
|
797 |
+
input_layer: conv2d
|
798 |
+
normalize_before: true
|
799 |
+
macaron_style: true
|
800 |
+
pos_enc_layer_type: rel_pos
|
801 |
+
selfattention_layer_type: rel_selfattn
|
802 |
+
activation_type: swish
|
803 |
+
use_cnn_module: true
|
804 |
+
cnn_module_kernel: 31
|
805 |
+
postencoder: null
|
806 |
+
postencoder_conf: {}
|
807 |
+
decoder: transformer
|
808 |
+
decoder_conf:
|
809 |
+
attention_heads: 8
|
810 |
+
linear_units: 2048
|
811 |
+
num_blocks: 6
|
812 |
+
dropout_rate: 0.1
|
813 |
+
positional_dropout_rate: 0.1
|
814 |
+
self_attention_dropout_rate: 0.1
|
815 |
+
src_attention_dropout_rate: 0.1
|
816 |
+
required:
|
817 |
+
- output_dir
|
818 |
+
- token_list
|
819 |
+
version: 0.10.7a1
|
820 |
+
distributed: true
|
821 |
+
```
|
822 |
+
|
823 |
+
</details>
|
824 |
+
|
825 |
+
|
826 |
+
|
827 |
+
### Citing ESPnet
|
828 |
+
|
829 |
+
```BibTex
|
830 |
+
@inproceedings{watanabe2018espnet,
|
831 |
+
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
|
832 |
+
title={{ESPnet}: End-to-End Speech Processing Toolkit},
|
833 |
+
year={2018},
|
834 |
+
booktitle={Proceedings of Interspeech},
|
835 |
+
pages={2207--2211},
|
836 |
+
doi={10.21437/Interspeech.2018-1456},
|
837 |
+
url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
|
838 |
+
}
|
839 |
+
|
840 |
+
|
841 |
+
|
842 |
+
|
843 |
+
```
|
844 |
+
|
845 |
+
or arXiv:
|
846 |
+
|
847 |
+
```bibtex
|
848 |
+
@misc{watanabe2018espnet,
|
849 |
+
title={ESPnet: End-to-End Speech Processing Toolkit},
|
850 |
+
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai},
|
851 |
+
year={2018},
|
852 |
+
eprint={1804.00015},
|
853 |
+
archivePrefix={arXiv},
|
854 |
+
primaryClass={cs.CL}
|
855 |
+
}
|
856 |
+
```
|
exp/asr_train_asr_raw_en_word/35epoch.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f1a301e560edb1d5351d8d39d3b19b78d5da8415629efa9fdc7f31b9a33ef797
|
3 |
+
size 437693801
|
exp/asr_train_asr_raw_en_word/RESULTS.md
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
<!-- Generated by scripts/utils/show_asr_result.sh -->
|
2 |
+
# RESULTS
|
3 |
+
## Environments
|
4 |
+
- date: `Tue Mar 29 04:17:37 UTC 2022`
|
5 |
+
- python version: `3.8.12 (default, Oct 12 2021, 13:49:34) [GCC 7.5.0]`
|
6 |
+
- espnet version: `espnet 0.10.7a1`
|
7 |
+
- pytorch version: `pytorch 1.9.0`
|
8 |
+
- Git hash: `0fae8113d99d092e7cbe4bcc48f9361e7012cff2`
|
9 |
+
- Commit date: `Thu Mar 24 07:54:19 2022 +0000`
|
10 |
+
|
11 |
+
## asr_train_asr_raw_en_word
|
12 |
+
### WER
|
13 |
+
|
14 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
15 |
+
|---|---|---|---|---|---|---|---|---|
|
16 |
+
|inference_asr_model_valid.acc.best/devel|8690|109017|63.9|21.1|15.0|2.0|38.1|75.4|
|
17 |
+
|inference_asr_model_valid.acc.best/test|6099|77315|69.0|17.4|13.5|1.8|32.8|68.9|
|
18 |
+
|inference_asr_model_valid.acc.best/test_ineube|6099|77315|77.8|12.0|10.2|1.5|23.6|59.4|
|
19 |
+
|inference_asr_model_valid.acc.best/test_qut|6099|77315|68.4|17.9|13.6|1.8|33.3|69.5|
|
20 |
+
|inference_asr_model_valid.acc.best/test_qut_ineube|6099|77315|78.0|11.9|10.2|1.4|23.4|59.3|
|
21 |
+
|
22 |
+
### CER
|
23 |
+
|
24 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
25 |
+
|---|---|---|---|---|---|---|---|---|
|
26 |
+
|inference_asr_model_valid.acc.best/devel|8690|513265|79.6|9.3|11.0|3.6|23.9|75.4|
|
27 |
+
|inference_asr_model_valid.acc.best/test|6099|362039|82.6|7.6|9.8|3.0|20.4|68.9|
|
28 |
+
|inference_asr_model_valid.acc.best/test_ineube|6099|362039|87.5|4.9|7.6|2.1|14.6|59.4|
|
29 |
+
|inference_asr_model_valid.acc.best/test_qut|6099|362039|82.3|7.8|9.9|3.1|20.8|69.5|
|
30 |
+
|inference_asr_model_valid.acc.best/test_qut_ineube|6099|362039|87.5|4.9|7.6|2.1|14.6|59.3|
|
exp/asr_train_asr_raw_en_word/config.yaml
ADDED
@@ -0,0 +1,757 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
config: conf/train_asr.yaml
|
2 |
+
print_config: false
|
3 |
+
log_level: INFO
|
4 |
+
dry_run: false
|
5 |
+
iterator_type: sequence
|
6 |
+
output_dir: exp/asr_train_asr_raw_en_word
|
7 |
+
ngpu: 1
|
8 |
+
seed: 0
|
9 |
+
num_workers: 1
|
10 |
+
num_att_plot: 3
|
11 |
+
dist_backend: nccl
|
12 |
+
dist_init_method: env://
|
13 |
+
dist_world_size: 3
|
14 |
+
dist_rank: 0
|
15 |
+
local_rank: 0
|
16 |
+
dist_master_addr: localhost
|
17 |
+
dist_master_port: 35953
|
18 |
+
dist_launcher: null
|
19 |
+
multiprocessing_distributed: true
|
20 |
+
unused_parameters: false
|
21 |
+
sharded_ddp: false
|
22 |
+
cudnn_enabled: true
|
23 |
+
cudnn_benchmark: false
|
24 |
+
cudnn_deterministic: true
|
25 |
+
collect_stats: false
|
26 |
+
write_collected_feats: false
|
27 |
+
max_epoch: 50
|
28 |
+
patience: null
|
29 |
+
val_scheduler_criterion:
|
30 |
+
- valid
|
31 |
+
- loss
|
32 |
+
early_stopping_criterion:
|
33 |
+
- valid
|
34 |
+
- loss
|
35 |
+
- min
|
36 |
+
best_model_criterion:
|
37 |
+
- - valid
|
38 |
+
- acc
|
39 |
+
- max
|
40 |
+
keep_nbest_models: 10
|
41 |
+
nbest_averaging_interval: 0
|
42 |
+
grad_clip: 5.0
|
43 |
+
grad_clip_type: 2.0
|
44 |
+
grad_noise: false
|
45 |
+
accum_grad: 1
|
46 |
+
no_forward_run: false
|
47 |
+
resume: true
|
48 |
+
train_dtype: float32
|
49 |
+
use_amp: false
|
50 |
+
log_interval: null
|
51 |
+
use_matplotlib: true
|
52 |
+
use_tensorboard: true
|
53 |
+
use_wandb: false
|
54 |
+
wandb_project: null
|
55 |
+
wandb_id: null
|
56 |
+
wandb_entity: null
|
57 |
+
wandb_name: null
|
58 |
+
wandb_model_log_interval: -1
|
59 |
+
detect_anomaly: false
|
60 |
+
pretrain_path: null
|
61 |
+
init_param: []
|
62 |
+
ignore_init_mismatch: false
|
63 |
+
freeze_param: []
|
64 |
+
num_iters_per_epoch: null
|
65 |
+
batch_size: 48
|
66 |
+
valid_batch_size: null
|
67 |
+
batch_bins: 1000000
|
68 |
+
valid_batch_bins: null
|
69 |
+
train_shape_file:
|
70 |
+
- exp/asr_stats_raw_en_word/train/speech_shape
|
71 |
+
- exp/asr_stats_raw_en_word/train/text_shape.word
|
72 |
+
valid_shape_file:
|
73 |
+
- exp/asr_stats_raw_en_word/valid/speech_shape
|
74 |
+
- exp/asr_stats_raw_en_word/valid/text_shape.word
|
75 |
+
batch_type: folded
|
76 |
+
valid_batch_type: null
|
77 |
+
fold_length:
|
78 |
+
- 80000
|
79 |
+
- 150
|
80 |
+
sort_in_batch: descending
|
81 |
+
sort_batch: descending
|
82 |
+
multiple_iterator: false
|
83 |
+
chunk_length: 500
|
84 |
+
chunk_shift_ratio: 0.5
|
85 |
+
num_cache_chunks: 1024
|
86 |
+
train_data_path_and_name_and_type:
|
87 |
+
- - dump/raw/train/wav.scp
|
88 |
+
- speech
|
89 |
+
- sound
|
90 |
+
- - dump/raw/train/text
|
91 |
+
- text
|
92 |
+
- text
|
93 |
+
valid_data_path_and_name_and_type:
|
94 |
+
- - dump/raw/devel/wav.scp
|
95 |
+
- speech
|
96 |
+
- sound
|
97 |
+
- - dump/raw/devel/text
|
98 |
+
- text
|
99 |
+
- text
|
100 |
+
allow_variable_data_keys: false
|
101 |
+
max_cache_size: 0.0
|
102 |
+
max_cache_fd: 32
|
103 |
+
valid_max_cache_size: null
|
104 |
+
optim: adam
|
105 |
+
optim_conf:
|
106 |
+
lr: 0.0002
|
107 |
+
scheduler: warmuplr
|
108 |
+
scheduler_conf:
|
109 |
+
warmup_steps: 25000
|
110 |
+
token_list:
|
111 |
+
- <blank>
|
112 |
+
- <unk>
|
113 |
+
- ▁the
|
114 |
+
- s
|
115 |
+
- ▁to
|
116 |
+
- ▁i
|
117 |
+
- ▁me
|
118 |
+
- ▁you
|
119 |
+
- ▁what
|
120 |
+
- ▁a
|
121 |
+
- ▁is
|
122 |
+
- a
|
123 |
+
- ▁my
|
124 |
+
- ▁please
|
125 |
+
- y
|
126 |
+
- ''''
|
127 |
+
- ▁in
|
128 |
+
- ing
|
129 |
+
- ▁s
|
130 |
+
- e
|
131 |
+
- o
|
132 |
+
- ▁for
|
133 |
+
- i
|
134 |
+
- ▁on
|
135 |
+
- d
|
136 |
+
- t
|
137 |
+
- u
|
138 |
+
- er
|
139 |
+
- p
|
140 |
+
- ▁of
|
141 |
+
- es
|
142 |
+
- re
|
143 |
+
- l
|
144 |
+
- ▁it
|
145 |
+
- ▁p
|
146 |
+
- le
|
147 |
+
- ▁f
|
148 |
+
- ▁m
|
149 |
+
- ▁email
|
150 |
+
- ▁d
|
151 |
+
- m
|
152 |
+
- ▁c
|
153 |
+
- ▁b
|
154 |
+
- st
|
155 |
+
- r
|
156 |
+
- n
|
157 |
+
- ar
|
158 |
+
- ▁t
|
159 |
+
- ▁h
|
160 |
+
- b
|
161 |
+
- ▁that
|
162 |
+
- c
|
163 |
+
- ▁this
|
164 |
+
- h
|
165 |
+
- an
|
166 |
+
- email_query
|
167 |
+
- ▁play
|
168 |
+
- ▁re
|
169 |
+
- ▁do
|
170 |
+
- ▁can
|
171 |
+
- at
|
172 |
+
- ▁have
|
173 |
+
- g
|
174 |
+
- ▁from
|
175 |
+
- ▁and
|
176 |
+
- en
|
177 |
+
- email_sendemail
|
178 |
+
- ▁olly
|
179 |
+
- 'on'
|
180 |
+
- ▁new
|
181 |
+
- it
|
182 |
+
- qa_factoid
|
183 |
+
- calendar_set
|
184 |
+
- ▁any
|
185 |
+
- or
|
186 |
+
- ▁g
|
187 |
+
- ent
|
188 |
+
- ▁how
|
189 |
+
- ▁tell
|
190 |
+
- ch
|
191 |
+
- ▁not
|
192 |
+
- ▁about
|
193 |
+
- ▁at
|
194 |
+
- ate
|
195 |
+
- general_negate
|
196 |
+
- f
|
197 |
+
- ▁today
|
198 |
+
- ▁e
|
199 |
+
- ed
|
200 |
+
- ▁list
|
201 |
+
- ▁r
|
202 |
+
- in
|
203 |
+
- k
|
204 |
+
- ic
|
205 |
+
- social_post
|
206 |
+
- ▁are
|
207 |
+
- play_music
|
208 |
+
- general_quirky
|
209 |
+
- ▁l
|
210 |
+
- al
|
211 |
+
- v
|
212 |
+
- ▁n
|
213 |
+
- ▁be
|
214 |
+
- ▁an
|
215 |
+
- ▁st
|
216 |
+
- et
|
217 |
+
- ▁am
|
218 |
+
- general_praise
|
219 |
+
- ▁time
|
220 |
+
- weather_query
|
221 |
+
- ▁up
|
222 |
+
- ▁check
|
223 |
+
- calendar_query
|
224 |
+
- ▁w
|
225 |
+
- om
|
226 |
+
- ur
|
227 |
+
- ▁send
|
228 |
+
- ▁with
|
229 |
+
- ly
|
230 |
+
- w
|
231 |
+
- general_explain
|
232 |
+
- ad
|
233 |
+
- ▁th
|
234 |
+
- news_query
|
235 |
+
- ▁one
|
236 |
+
- ▁emails
|
237 |
+
- day
|
238 |
+
- ▁sh
|
239 |
+
- ce
|
240 |
+
- ▁
|
241 |
+
- ▁last
|
242 |
+
- ve
|
243 |
+
- ▁he
|
244 |
+
- z
|
245 |
+
- ▁ch
|
246 |
+
- ▁will
|
247 |
+
- ▁set
|
248 |
+
- ▁would
|
249 |
+
- ▁was
|
250 |
+
- x
|
251 |
+
- general_repeat
|
252 |
+
- ▁add
|
253 |
+
- ▁again
|
254 |
+
- ou
|
255 |
+
- ▁ex
|
256 |
+
- is
|
257 |
+
- ct
|
258 |
+
- general_affirm
|
259 |
+
- general_confirm
|
260 |
+
- ▁song
|
261 |
+
- ▁next
|
262 |
+
- ▁j
|
263 |
+
- ▁meeting
|
264 |
+
- um
|
265 |
+
- ation
|
266 |
+
- ▁turn
|
267 |
+
- ▁did
|
268 |
+
- if
|
269 |
+
- ▁alarm
|
270 |
+
- am
|
271 |
+
- ▁like
|
272 |
+
- datetime_query
|
273 |
+
- ter
|
274 |
+
- ▁remind
|
275 |
+
- ▁o
|
276 |
+
- qa_definition
|
277 |
+
- ▁said
|
278 |
+
- ▁calendar
|
279 |
+
- ll
|
280 |
+
- se
|
281 |
+
- ers
|
282 |
+
- ▁pr
|
283 |
+
- th
|
284 |
+
- ▁get
|
285 |
+
- our
|
286 |
+
- ▁need
|
287 |
+
- ▁all
|
288 |
+
- ot
|
289 |
+
- ▁want
|
290 |
+
- ▁off
|
291 |
+
- and
|
292 |
+
- ▁right
|
293 |
+
- ▁de
|
294 |
+
- ▁tr
|
295 |
+
- ut
|
296 |
+
- general_dontcare
|
297 |
+
- as
|
298 |
+
- ▁week
|
299 |
+
- ▁tweet
|
300 |
+
- ight
|
301 |
+
- ir
|
302 |
+
- ▁your
|
303 |
+
- ▁event
|
304 |
+
- ▁news
|
305 |
+
- ▁se
|
306 |
+
- ay
|
307 |
+
- ion
|
308 |
+
- ▁com
|
309 |
+
- ▁there
|
310 |
+
- ▁ye
|
311 |
+
- ▁weather
|
312 |
+
- un
|
313 |
+
- ▁confirm
|
314 |
+
- ld
|
315 |
+
- calendar_remove
|
316 |
+
- ▁y
|
317 |
+
- ▁lights
|
318 |
+
- ▁more
|
319 |
+
- ▁v
|
320 |
+
- play_radio
|
321 |
+
- ▁does
|
322 |
+
- ▁po
|
323 |
+
- ▁now
|
324 |
+
- id
|
325 |
+
- email_querycontact
|
326 |
+
- ▁show
|
327 |
+
- ▁could
|
328 |
+
- ery
|
329 |
+
- op
|
330 |
+
- ▁day
|
331 |
+
- ▁pm
|
332 |
+
- ▁music
|
333 |
+
- ▁tomorrow
|
334 |
+
- ▁train
|
335 |
+
- ▁u
|
336 |
+
- ine
|
337 |
+
- ▁or
|
338 |
+
- ange
|
339 |
+
- qa_currency
|
340 |
+
- ice
|
341 |
+
- ▁contact
|
342 |
+
- ▁just
|
343 |
+
- ▁jo
|
344 |
+
- ▁think
|
345 |
+
- qa_stock
|
346 |
+
- end
|
347 |
+
- ss
|
348 |
+
- ber
|
349 |
+
- ▁tw
|
350 |
+
- ▁command
|
351 |
+
- ▁make
|
352 |
+
- ▁no
|
353 |
+
- ▁mo
|
354 |
+
- pe
|
355 |
+
- ▁find
|
356 |
+
- general_commandstop
|
357 |
+
- ▁when
|
358 |
+
- social_query
|
359 |
+
- ▁so
|
360 |
+
- ong
|
361 |
+
- ▁co
|
362 |
+
- ant
|
363 |
+
- ow
|
364 |
+
- q
|
365 |
+
- ▁much
|
366 |
+
- ▁where
|
367 |
+
- ue
|
368 |
+
- ul
|
369 |
+
- ri
|
370 |
+
- ake
|
371 |
+
- ap
|
372 |
+
- ▁start
|
373 |
+
- ▁mar
|
374 |
+
- ▁by
|
375 |
+
- one
|
376 |
+
- ▁know
|
377 |
+
- ▁wor
|
378 |
+
- oo
|
379 |
+
- ▁give
|
380 |
+
- ▁let
|
381 |
+
- ▁events
|
382 |
+
- der
|
383 |
+
- ▁ro
|
384 |
+
- ▁pl
|
385 |
+
- play_podcasts
|
386 |
+
- art
|
387 |
+
- us
|
388 |
+
- ▁work
|
389 |
+
- ▁current
|
390 |
+
- ol
|
391 |
+
- cooking_recipe
|
392 |
+
- nt
|
393 |
+
- ▁correct
|
394 |
+
- transport_query
|
395 |
+
- ia
|
396 |
+
- ▁stock
|
397 |
+
- ▁br
|
398 |
+
- ive
|
399 |
+
- ▁app
|
400 |
+
- ▁two
|
401 |
+
- ▁latest
|
402 |
+
- lists_query
|
403 |
+
- recommendation_events
|
404 |
+
- ab
|
405 |
+
- ▁go
|
406 |
+
- ▁but
|
407 |
+
- ook
|
408 |
+
- ▁some
|
409 |
+
- ke
|
410 |
+
- alarm_set
|
411 |
+
- play_audiobook
|
412 |
+
- ▁k
|
413 |
+
- ▁response
|
414 |
+
- ▁wr
|
415 |
+
- cast
|
416 |
+
- ▁open
|
417 |
+
- ▁cle
|
418 |
+
- ▁done
|
419 |
+
- ▁got
|
420 |
+
- ▁ca
|
421 |
+
- ite
|
422 |
+
- ase
|
423 |
+
- ▁thank
|
424 |
+
- iv
|
425 |
+
- ag
|
426 |
+
- ah
|
427 |
+
- ▁answer
|
428 |
+
- ie
|
429 |
+
- ▁five
|
430 |
+
- ▁book
|
431 |
+
- ▁rec
|
432 |
+
- ore
|
433 |
+
- ▁john
|
434 |
+
- ist
|
435 |
+
- ment
|
436 |
+
- ▁appreci
|
437 |
+
- ▁fri
|
438 |
+
- ack
|
439 |
+
- ▁remove
|
440 |
+
- ated
|
441 |
+
- ock
|
442 |
+
- ree
|
443 |
+
- j
|
444 |
+
- ▁good
|
445 |
+
- ▁many
|
446 |
+
- orn
|
447 |
+
- fe
|
448 |
+
- ▁radio
|
449 |
+
- ▁we
|
450 |
+
- int
|
451 |
+
- ▁facebook
|
452 |
+
- ▁cl
|
453 |
+
- ▁sev
|
454 |
+
- ▁schedule
|
455 |
+
- ard
|
456 |
+
- ▁per
|
457 |
+
- ▁li
|
458 |
+
- ▁going
|
459 |
+
- nd
|
460 |
+
- ain
|
461 |
+
- recommendation_locations
|
462 |
+
- ▁post
|
463 |
+
- lists_createoradd
|
464 |
+
- ff
|
465 |
+
- ▁su
|
466 |
+
- red
|
467 |
+
- iot_hue_lightoff
|
468 |
+
- lists_remove
|
469 |
+
- ▁ar
|
470 |
+
- een
|
471 |
+
- ▁say
|
472 |
+
- ro
|
473 |
+
- ▁volume
|
474 |
+
- ▁le
|
475 |
+
- ▁reply
|
476 |
+
- ▁complaint
|
477 |
+
- ▁delete
|
478 |
+
- ▁out
|
479 |
+
- lly
|
480 |
+
- ame
|
481 |
+
- ▁ne
|
482 |
+
- ▁detail
|
483 |
+
- ▁if
|
484 |
+
- im
|
485 |
+
- ▁happ
|
486 |
+
- orr
|
487 |
+
- ich
|
488 |
+
- em
|
489 |
+
- ▁ev
|
490 |
+
- ction
|
491 |
+
- ▁dollar
|
492 |
+
- ▁as
|
493 |
+
- alarm_query
|
494 |
+
- audio_volume_mute
|
495 |
+
- ac
|
496 |
+
- music_query
|
497 |
+
- ▁mon
|
498 |
+
- ther
|
499 |
+
- ▁thanks
|
500 |
+
- cel
|
501 |
+
- ▁who
|
502 |
+
- ave
|
503 |
+
- ▁service
|
504 |
+
- ▁mail
|
505 |
+
- ▁hear
|
506 |
+
- ty
|
507 |
+
- de
|
508 |
+
- ▁si
|
509 |
+
- ▁wh
|
510 |
+
- ood
|
511 |
+
- ell
|
512 |
+
- ▁con
|
513 |
+
- icket
|
514 |
+
- ▁once
|
515 |
+
- ound
|
516 |
+
- ▁don
|
517 |
+
- ▁loc
|
518 |
+
- ▁light
|
519 |
+
- ▁birthday
|
520 |
+
- ▁inf
|
521 |
+
- ffe
|
522 |
+
- ▁has
|
523 |
+
- ▁playlist
|
524 |
+
- ort
|
525 |
+
- el
|
526 |
+
- ening
|
527 |
+
- ▁us
|
528 |
+
- ▁un
|
529 |
+
- own
|
530 |
+
- ▁inc
|
531 |
+
- ai
|
532 |
+
- ▁speak
|
533 |
+
- age
|
534 |
+
- ▁mess
|
535 |
+
- ast
|
536 |
+
- ci
|
537 |
+
- ver
|
538 |
+
- ▁ten
|
539 |
+
- ▁underst
|
540 |
+
- gh
|
541 |
+
- audio_volume_up
|
542 |
+
- ome
|
543 |
+
- transport_ticket
|
544 |
+
- ind
|
545 |
+
- iot_hue_lightchange
|
546 |
+
- iot_coffee
|
547 |
+
- pp
|
548 |
+
- ▁res
|
549 |
+
- plain
|
550 |
+
- io
|
551 |
+
- lar
|
552 |
+
- takeaway_query
|
553 |
+
- ge
|
554 |
+
- takeaway_order
|
555 |
+
- email_addcontact
|
556 |
+
- play_game
|
557 |
+
- ak
|
558 |
+
- ▁fa
|
559 |
+
- transport_traffic
|
560 |
+
- music_likeness
|
561 |
+
- ▁rep
|
562 |
+
- act
|
563 |
+
- ust
|
564 |
+
- transport_taxi
|
565 |
+
- iot_hue_lightdim
|
566 |
+
- ▁mu
|
567 |
+
- ▁ti
|
568 |
+
- ick
|
569 |
+
- ▁ha
|
570 |
+
- ould
|
571 |
+
- general_joke
|
572 |
+
- '1'
|
573 |
+
- qa_maths
|
574 |
+
- ▁lo
|
575 |
+
- iot_cleaning
|
576 |
+
- ill
|
577 |
+
- her
|
578 |
+
- iot_hue_lightup
|
579 |
+
- pl
|
580 |
+
- '2'
|
581 |
+
- alarm_remove
|
582 |
+
- orrect
|
583 |
+
- ▁cont
|
584 |
+
- mail
|
585 |
+
- out
|
586 |
+
- audio_volume_down
|
587 |
+
- book
|
588 |
+
- ail
|
589 |
+
- recommendation_movies
|
590 |
+
- ck
|
591 |
+
- ▁man
|
592 |
+
- ▁mus
|
593 |
+
- ▁che
|
594 |
+
- me
|
595 |
+
- ume
|
596 |
+
- ▁answ
|
597 |
+
- datetime_convert
|
598 |
+
- ▁late
|
599 |
+
- iot_wemo_on
|
600 |
+
- ▁twe
|
601 |
+
- music_settings
|
602 |
+
- iot_wemo_off
|
603 |
+
- orre
|
604 |
+
- ith
|
605 |
+
- ▁tom
|
606 |
+
- ▁fr
|
607 |
+
- ere
|
608 |
+
- ▁ad
|
609 |
+
- xt
|
610 |
+
- ▁ab
|
611 |
+
- ank
|
612 |
+
- general_greet
|
613 |
+
- now
|
614 |
+
- ▁meet
|
615 |
+
- ▁curre
|
616 |
+
- ▁respon
|
617 |
+
- ▁ag
|
618 |
+
- audio_volume_other
|
619 |
+
- ink
|
620 |
+
- ▁spe
|
621 |
+
- iot_hue_lighton
|
622 |
+
- ght
|
623 |
+
- ▁rem
|
624 |
+
- '?'
|
625 |
+
- urn
|
626 |
+
- ▁op
|
627 |
+
- ▁complain
|
628 |
+
- ▁comm
|
629 |
+
- let
|
630 |
+
- music_dislikeness
|
631 |
+
- ove
|
632 |
+
- ▁sch
|
633 |
+
- ather
|
634 |
+
- ▁rad
|
635 |
+
- edule
|
636 |
+
- ▁under
|
637 |
+
- lease
|
638 |
+
- ▁bir
|
639 |
+
- erv
|
640 |
+
- ▁birth
|
641 |
+
- ▁face
|
642 |
+
- ▁cur
|
643 |
+
- sw
|
644 |
+
- ▁serv
|
645 |
+
- ek
|
646 |
+
- aid
|
647 |
+
- '9'
|
648 |
+
- ▁vol
|
649 |
+
- edu
|
650 |
+
- '5'
|
651 |
+
- cooking_query
|
652 |
+
- lete
|
653 |
+
- ▁joh
|
654 |
+
- ▁det
|
655 |
+
- firm
|
656 |
+
- nder
|
657 |
+
- '0'
|
658 |
+
- _
|
659 |
+
- irm
|
660 |
+
- '8'
|
661 |
+
- '&'
|
662 |
+
- list
|
663 |
+
- pon
|
664 |
+
- qa_query
|
665 |
+
- '7'
|
666 |
+
- '3'
|
667 |
+
- '-'
|
668 |
+
- N
|
669 |
+
- A
|
670 |
+
- M
|
671 |
+
- E
|
672 |
+
- ']'
|
673 |
+
- '['
|
674 |
+
- ':'
|
675 |
+
- reci
|
676 |
+
- ▁doll
|
677 |
+
- <sos/eos>
|
678 |
+
init: null
|
679 |
+
input_size: null
|
680 |
+
ctc_conf:
|
681 |
+
dropout_rate: 0.0
|
682 |
+
ctc_type: builtin
|
683 |
+
reduce: true
|
684 |
+
ignore_nan_grad: true
|
685 |
+
joint_net_conf: null
|
686 |
+
model_conf:
|
687 |
+
ctc_weight: 0.3
|
688 |
+
lsm_weight: 0.1
|
689 |
+
length_normalized_loss: false
|
690 |
+
extract_feats_in_collect_stats: false
|
691 |
+
use_preprocessor: true
|
692 |
+
token_type: word
|
693 |
+
bpemodel: null
|
694 |
+
non_linguistic_symbols: null
|
695 |
+
cleaner: null
|
696 |
+
g2p: null
|
697 |
+
speech_volume_normalize: null
|
698 |
+
rir_scp: null
|
699 |
+
rir_apply_prob: 1.0
|
700 |
+
noise_scp: null
|
701 |
+
noise_apply_prob: 1.0
|
702 |
+
noise_db_range: '13_15'
|
703 |
+
frontend: default
|
704 |
+
frontend_conf:
|
705 |
+
fs: 16k
|
706 |
+
specaug: specaug
|
707 |
+
specaug_conf:
|
708 |
+
apply_time_warp: true
|
709 |
+
time_warp_window: 5
|
710 |
+
time_warp_mode: bicubic
|
711 |
+
apply_freq_mask: true
|
712 |
+
freq_mask_width_range:
|
713 |
+
- 0
|
714 |
+
- 30
|
715 |
+
num_freq_mask: 2
|
716 |
+
apply_time_mask: true
|
717 |
+
time_mask_width_range:
|
718 |
+
- 0
|
719 |
+
- 40
|
720 |
+
num_time_mask: 2
|
721 |
+
normalize: utterance_mvn
|
722 |
+
normalize_conf: {}
|
723 |
+
preencoder: null
|
724 |
+
preencoder_conf: {}
|
725 |
+
encoder: conformer
|
726 |
+
encoder_conf:
|
727 |
+
output_size: 512
|
728 |
+
attention_heads: 8
|
729 |
+
linear_units: 2048
|
730 |
+
num_blocks: 12
|
731 |
+
dropout_rate: 0.1
|
732 |
+
positional_dropout_rate: 0.1
|
733 |
+
attention_dropout_rate: 0.1
|
734 |
+
input_layer: conv2d
|
735 |
+
normalize_before: true
|
736 |
+
macaron_style: true
|
737 |
+
pos_enc_layer_type: rel_pos
|
738 |
+
selfattention_layer_type: rel_selfattn
|
739 |
+
activation_type: swish
|
740 |
+
use_cnn_module: true
|
741 |
+
cnn_module_kernel: 31
|
742 |
+
postencoder: null
|
743 |
+
postencoder_conf: {}
|
744 |
+
decoder: transformer
|
745 |
+
decoder_conf:
|
746 |
+
attention_heads: 8
|
747 |
+
linear_units: 2048
|
748 |
+
num_blocks: 6
|
749 |
+
dropout_rate: 0.1
|
750 |
+
positional_dropout_rate: 0.1
|
751 |
+
self_attention_dropout_rate: 0.1
|
752 |
+
src_attention_dropout_rate: 0.1
|
753 |
+
required:
|
754 |
+
- output_dir
|
755 |
+
- token_list
|
756 |
+
version: 0.10.7a1
|
757 |
+
distributed: true
|
exp/asr_train_asr_raw_en_word/images/acc.png
ADDED
exp/asr_train_asr_raw_en_word/images/backward_time.png
ADDED
exp/asr_train_asr_raw_en_word/images/cer.png
ADDED
exp/asr_train_asr_raw_en_word/images/cer_ctc.png
ADDED
exp/asr_train_asr_raw_en_word/images/forward_time.png
ADDED
exp/asr_train_asr_raw_en_word/images/gpu_max_cached_mem_GB.png
ADDED
exp/asr_train_asr_raw_en_word/images/iter_time.png
ADDED
exp/asr_train_asr_raw_en_word/images/loss.png
ADDED
exp/asr_train_asr_raw_en_word/images/loss_att.png
ADDED
exp/asr_train_asr_raw_en_word/images/loss_ctc.png
ADDED
exp/asr_train_asr_raw_en_word/images/optim0_lr0.png
ADDED
exp/asr_train_asr_raw_en_word/images/optim_step_time.png
ADDED
exp/asr_train_asr_raw_en_word/images/train_time.png
ADDED
exp/asr_train_asr_raw_en_word/images/wer.png
ADDED
meta.yaml
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
espnet: 0.10.7a1
|
2 |
+
files:
|
3 |
+
asr_model_file: exp/asr_train_asr_raw_en_word/35epoch.pth
|
4 |
+
python: "3.8.12 (default, Oct 12 2021, 13:49:34) \n[GCC 7.5.0]"
|
5 |
+
timestamp: 1654349933.152364
|
6 |
+
torch: 1.9.0
|
7 |
+
yaml_files:
|
8 |
+
asr_train_config: exp/asr_train_asr_raw_en_word/config.yaml
|