add model files
Browse files- README.md +324 -0
- exp/asr_stats_raw_it_char/train/feats_stats.npz +3 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/RESULTS.md +29 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/config.yaml +223 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/acc.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/backward_time.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/cer.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/cer_ctc.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/forward_time.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/gpu_max_cached_mem_GB.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/iter_time.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/loss.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/loss_att.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/loss_ctc.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/optim0_lr0.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/optim_step_time.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/train_time.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/wer.png +0 -0
- exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/valid.acc.ave_10best.pth +3 -0
- meta.yaml +8 -0
README.md
ADDED
@@ -0,0 +1,324 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
tags:
|
3 |
+
- espnet
|
4 |
+
- audio
|
5 |
+
- automatic-speech-recognition
|
6 |
+
language: it
|
7 |
+
datasets:
|
8 |
+
- voxforge
|
9 |
+
license: cc-by-4.0
|
10 |
+
---
|
11 |
+
|
12 |
+
## ESPnet2 ASR model
|
13 |
+
|
14 |
+
### `pyf98/voxforge_it_conformer_e15_linear1024`
|
15 |
+
|
16 |
+
This model was trained by Yifan Peng using voxforge recipe in [espnet](https://github.com/espnet/espnet/).
|
17 |
+
|
18 |
+
### Demo: How to use in ESPnet2
|
19 |
+
|
20 |
+
Follow the [ESPnet installation instructions](https://espnet.github.io/espnet/installation.html)
|
21 |
+
if you haven't done that already.
|
22 |
+
|
23 |
+
```bash
|
24 |
+
cd espnet
|
25 |
+
git checkout bf8c8f00194bdfed8ca388d8b20d14791b7d270e
|
26 |
+
pip install -e .
|
27 |
+
cd egs2/voxforge/asr1
|
28 |
+
./run.sh --skip_data_prep false --skip_train true --download_model pyf98/voxforge_it_conformer_e15_linear1024
|
29 |
+
```
|
30 |
+
|
31 |
+
<!-- Generated by scripts/utils/show_asr_result.sh -->
|
32 |
+
# RESULTS
|
33 |
+
## Environments
|
34 |
+
- date: `Thu Dec 29 01:59:25 EST 2022`
|
35 |
+
- python version: `3.9.15 (main, Nov 24 2022, 14:31:59) [GCC 11.2.0]`
|
36 |
+
- espnet version: `espnet 202211`
|
37 |
+
- pytorch version: `pytorch 1.12.1`
|
38 |
+
- Git hash: `bf8c8f00194bdfed8ca388d8b20d14791b7d270e`
|
39 |
+
- Commit date: `Wed Dec 28 22:43:13 2022 -0500`
|
40 |
+
|
41 |
+
## asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse
|
42 |
+
### WER
|
43 |
+
|
44 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
45 |
+
|---|---|---|---|---|---|---|---|---|
|
46 |
+
|decode_asr_asr_model_valid.acc.ave/dt_it|1035|12587|70.2|24.6|5.2|3.3|33.1|94.7|
|
47 |
+
|decode_asr_asr_model_valid.acc.ave/et_it|1103|13699|71.9|23.3|4.8|2.9|31.0|92.4|
|
48 |
+
|
49 |
+
### CER
|
50 |
+
|
51 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
52 |
+
|---|---|---|---|---|---|---|---|---|
|
53 |
+
|decode_asr_asr_model_valid.acc.ave/dt_it|1035|75494|92.9|3.9|3.2|1.8|9.0|94.7|
|
54 |
+
|decode_asr_asr_model_valid.acc.ave/et_it|1103|81228|93.6|3.6|2.8|1.7|8.1|92.4|
|
55 |
+
|
56 |
+
### TER
|
57 |
+
|
58 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
59 |
+
|---|---|---|---|---|---|---|---|---|
|
60 |
+
|
61 |
+
## ASR config
|
62 |
+
|
63 |
+
<details><summary>expand</summary>
|
64 |
+
|
65 |
+
```
|
66 |
+
config: conf/tuning/train_asr_conformer_e15_linear1024.yaml
|
67 |
+
print_config: false
|
68 |
+
log_level: INFO
|
69 |
+
dry_run: false
|
70 |
+
iterator_type: sequence
|
71 |
+
output_dir: exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse
|
72 |
+
ngpu: 1
|
73 |
+
seed: 0
|
74 |
+
num_workers: 4
|
75 |
+
num_att_plot: 3
|
76 |
+
dist_backend: nccl
|
77 |
+
dist_init_method: env://
|
78 |
+
dist_world_size: null
|
79 |
+
dist_rank: null
|
80 |
+
local_rank: 0
|
81 |
+
dist_master_addr: null
|
82 |
+
dist_master_port: null
|
83 |
+
dist_launcher: null
|
84 |
+
multiprocessing_distributed: false
|
85 |
+
unused_parameters: false
|
86 |
+
sharded_ddp: false
|
87 |
+
cudnn_enabled: true
|
88 |
+
cudnn_benchmark: false
|
89 |
+
cudnn_deterministic: true
|
90 |
+
collect_stats: false
|
91 |
+
write_collected_feats: false
|
92 |
+
max_epoch: 100
|
93 |
+
patience: null
|
94 |
+
val_scheduler_criterion:
|
95 |
+
- valid
|
96 |
+
- loss
|
97 |
+
early_stopping_criterion:
|
98 |
+
- valid
|
99 |
+
- loss
|
100 |
+
- min
|
101 |
+
best_model_criterion:
|
102 |
+
- - valid
|
103 |
+
- acc
|
104 |
+
- max
|
105 |
+
keep_nbest_models: 10
|
106 |
+
nbest_averaging_interval: 0
|
107 |
+
grad_clip: 5.0
|
108 |
+
grad_clip_type: 2.0
|
109 |
+
grad_noise: false
|
110 |
+
accum_grad: 1
|
111 |
+
no_forward_run: false
|
112 |
+
resume: true
|
113 |
+
train_dtype: float32
|
114 |
+
use_amp: true
|
115 |
+
log_interval: null
|
116 |
+
use_matplotlib: true
|
117 |
+
use_tensorboard: true
|
118 |
+
create_graph_in_tensorboard: false
|
119 |
+
use_wandb: false
|
120 |
+
wandb_project: null
|
121 |
+
wandb_id: null
|
122 |
+
wandb_entity: null
|
123 |
+
wandb_name: null
|
124 |
+
wandb_model_log_interval: -1
|
125 |
+
detect_anomaly: false
|
126 |
+
pretrain_path: null
|
127 |
+
init_param: []
|
128 |
+
ignore_init_mismatch: false
|
129 |
+
freeze_param: []
|
130 |
+
num_iters_per_epoch: null
|
131 |
+
batch_size: 128
|
132 |
+
valid_batch_size: null
|
133 |
+
batch_bins: 1000000
|
134 |
+
valid_batch_bins: null
|
135 |
+
train_shape_file:
|
136 |
+
- exp/asr_stats_raw_it_char/train/speech_shape
|
137 |
+
- exp/asr_stats_raw_it_char/train/text_shape.char
|
138 |
+
valid_shape_file:
|
139 |
+
- exp/asr_stats_raw_it_char/valid/speech_shape
|
140 |
+
- exp/asr_stats_raw_it_char/valid/text_shape.char
|
141 |
+
batch_type: folded
|
142 |
+
valid_batch_type: null
|
143 |
+
fold_length:
|
144 |
+
- 80000
|
145 |
+
- 150
|
146 |
+
sort_in_batch: descending
|
147 |
+
sort_batch: descending
|
148 |
+
multiple_iterator: false
|
149 |
+
chunk_length: 500
|
150 |
+
chunk_shift_ratio: 0.5
|
151 |
+
num_cache_chunks: 1024
|
152 |
+
train_data_path_and_name_and_type:
|
153 |
+
- - dump/raw/tr_it/wav.scp
|
154 |
+
- speech
|
155 |
+
- sound
|
156 |
+
- - dump/raw/tr_it/text
|
157 |
+
- text
|
158 |
+
- text
|
159 |
+
valid_data_path_and_name_and_type:
|
160 |
+
- - dump/raw/dt_it/wav.scp
|
161 |
+
- speech
|
162 |
+
- sound
|
163 |
+
- - dump/raw/dt_it/text
|
164 |
+
- text
|
165 |
+
- text
|
166 |
+
allow_variable_data_keys: false
|
167 |
+
max_cache_size: 0.0
|
168 |
+
max_cache_fd: 32
|
169 |
+
valid_max_cache_size: null
|
170 |
+
optim: adam
|
171 |
+
optim_conf:
|
172 |
+
lr: 0.002
|
173 |
+
scheduler: warmuplr
|
174 |
+
scheduler_conf:
|
175 |
+
warmup_steps: 10000
|
176 |
+
token_list:
|
177 |
+
- <blank>
|
178 |
+
- <unk>
|
179 |
+
- <space>
|
180 |
+
- A
|
181 |
+
- E
|
182 |
+
- I
|
183 |
+
- O
|
184 |
+
- R
|
185 |
+
- N
|
186 |
+
- L
|
187 |
+
- S
|
188 |
+
- T
|
189 |
+
- C
|
190 |
+
- D
|
191 |
+
- U
|
192 |
+
- M
|
193 |
+
- P
|
194 |
+
- V
|
195 |
+
- G
|
196 |
+
- F
|
197 |
+
- H
|
198 |
+
- B
|
199 |
+
- Q
|
200 |
+
- Z
|
201 |
+
- ''''
|
202 |
+
- Ò
|
203 |
+
- À
|
204 |
+
- È
|
205 |
+
- Ú
|
206 |
+
- X
|
207 |
+
- W
|
208 |
+
- Í
|
209 |
+
- É
|
210 |
+
- Y
|
211 |
+
- K
|
212 |
+
- J
|
213 |
+
- '1'
|
214 |
+
- <sos/eos>
|
215 |
+
init: null
|
216 |
+
input_size: null
|
217 |
+
ctc_conf:
|
218 |
+
dropout_rate: 0.0
|
219 |
+
ctc_type: builtin
|
220 |
+
reduce: true
|
221 |
+
ignore_nan_grad: null
|
222 |
+
zero_infinity: true
|
223 |
+
joint_net_conf: null
|
224 |
+
use_preprocessor: true
|
225 |
+
token_type: char
|
226 |
+
bpemodel: null
|
227 |
+
non_linguistic_symbols: null
|
228 |
+
cleaner: null
|
229 |
+
g2p: null
|
230 |
+
speech_volume_normalize: null
|
231 |
+
rir_scp: null
|
232 |
+
rir_apply_prob: 1.0
|
233 |
+
noise_scp: null
|
234 |
+
noise_apply_prob: 1.0
|
235 |
+
noise_db_range: '13_15'
|
236 |
+
short_noise_thres: 0.5
|
237 |
+
frontend: default
|
238 |
+
frontend_conf:
|
239 |
+
fs: 16k
|
240 |
+
specaug: null
|
241 |
+
specaug_conf: {}
|
242 |
+
normalize: global_mvn
|
243 |
+
normalize_conf:
|
244 |
+
stats_file: exp/asr_stats_raw_it_char/train/feats_stats.npz
|
245 |
+
norm_vars: false
|
246 |
+
model: espnet
|
247 |
+
model_conf:
|
248 |
+
ctc_weight: 0.3
|
249 |
+
lsm_weight: 0.1
|
250 |
+
length_normalized_loss: false
|
251 |
+
preencoder: null
|
252 |
+
preencoder_conf: {}
|
253 |
+
encoder: conformer
|
254 |
+
encoder_conf:
|
255 |
+
output_size: 256
|
256 |
+
attention_heads: 4
|
257 |
+
linear_units: 1024
|
258 |
+
num_blocks: 15
|
259 |
+
dropout_rate: 0.1
|
260 |
+
positional_dropout_rate: 0.1
|
261 |
+
attention_dropout_rate: 0.1
|
262 |
+
input_layer: conv2d
|
263 |
+
normalize_before: true
|
264 |
+
macaron_style: true
|
265 |
+
rel_pos_type: latest
|
266 |
+
pos_enc_layer_type: rel_pos
|
267 |
+
selfattention_layer_type: rel_selfattn
|
268 |
+
activation_type: swish
|
269 |
+
use_cnn_module: true
|
270 |
+
cnn_module_kernel: 31
|
271 |
+
postencoder: null
|
272 |
+
postencoder_conf: {}
|
273 |
+
decoder: transformer
|
274 |
+
decoder_conf:
|
275 |
+
attention_heads: 4
|
276 |
+
linear_units: 2048
|
277 |
+
num_blocks: 6
|
278 |
+
dropout_rate: 0.1
|
279 |
+
positional_dropout_rate: 0.1
|
280 |
+
self_attention_dropout_rate: 0.0
|
281 |
+
src_attention_dropout_rate: 0.0
|
282 |
+
preprocessor: default
|
283 |
+
preprocessor_conf: {}
|
284 |
+
required:
|
285 |
+
- output_dir
|
286 |
+
- token_list
|
287 |
+
version: '202211'
|
288 |
+
distributed: false
|
289 |
+
```
|
290 |
+
|
291 |
+
</details>
|
292 |
+
|
293 |
+
|
294 |
+
|
295 |
+
### Citing ESPnet
|
296 |
+
|
297 |
+
```BibTex
|
298 |
+
@inproceedings{watanabe2018espnet,
|
299 |
+
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},
|
300 |
+
title={{ESPnet}: End-to-End Speech Processing Toolkit},
|
301 |
+
year={2018},
|
302 |
+
booktitle={Proceedings of Interspeech},
|
303 |
+
pages={2207--2211},
|
304 |
+
doi={10.21437/Interspeech.2018-1456},
|
305 |
+
url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
|
306 |
+
}
|
307 |
+
|
308 |
+
|
309 |
+
|
310 |
+
|
311 |
+
```
|
312 |
+
|
313 |
+
or arXiv:
|
314 |
+
|
315 |
+
```bibtex
|
316 |
+
@misc{watanabe2018espnet,
|
317 |
+
title={ESPnet: End-to-End Speech Processing Toolkit},
|
318 |
+
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},
|
319 |
+
year={2018},
|
320 |
+
eprint={1804.00015},
|
321 |
+
archivePrefix={arXiv},
|
322 |
+
primaryClass={cs.CL}
|
323 |
+
}
|
324 |
+
```
|
exp/asr_stats_raw_it_char/train/feats_stats.npz
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8ac4c6cee9a2bbf0175bb92bc350e21038c874b9d01b537b2688f5c8e1bae3d6
|
3 |
+
size 1402
|
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/RESULTS.md
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
<!-- Generated by scripts/utils/show_asr_result.sh -->
|
2 |
+
# RESULTS
|
3 |
+
## Environments
|
4 |
+
- date: `Thu Dec 29 01:59:25 EST 2022`
|
5 |
+
- python version: `3.9.15 (main, Nov 24 2022, 14:31:59) [GCC 11.2.0]`
|
6 |
+
- espnet version: `espnet 202211`
|
7 |
+
- pytorch version: `pytorch 1.12.1`
|
8 |
+
- Git hash: `bf8c8f00194bdfed8ca388d8b20d14791b7d270e`
|
9 |
+
- Commit date: `Wed Dec 28 22:43:13 2022 -0500`
|
10 |
+
|
11 |
+
## asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse
|
12 |
+
### WER
|
13 |
+
|
14 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
15 |
+
|---|---|---|---|---|---|---|---|---|
|
16 |
+
|decode_asr_asr_model_valid.acc.ave/dt_it|1035|12587|70.2|24.6|5.2|3.3|33.1|94.7|
|
17 |
+
|decode_asr_asr_model_valid.acc.ave/et_it|1103|13699|71.9|23.3|4.8|2.9|31.0|92.4|
|
18 |
+
|
19 |
+
### CER
|
20 |
+
|
21 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
22 |
+
|---|---|---|---|---|---|---|---|---|
|
23 |
+
|decode_asr_asr_model_valid.acc.ave/dt_it|1035|75494|92.9|3.9|3.2|1.8|9.0|94.7|
|
24 |
+
|decode_asr_asr_model_valid.acc.ave/et_it|1103|81228|93.6|3.6|2.8|1.7|8.1|92.4|
|
25 |
+
|
26 |
+
### TER
|
27 |
+
|
28 |
+
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|
29 |
+
|---|---|---|---|---|---|---|---|---|
|
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/config.yaml
ADDED
@@ -0,0 +1,223 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
config: conf/tuning/train_asr_conformer_e15_linear1024.yaml
|
2 |
+
print_config: false
|
3 |
+
log_level: INFO
|
4 |
+
dry_run: false
|
5 |
+
iterator_type: sequence
|
6 |
+
output_dir: exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse
|
7 |
+
ngpu: 1
|
8 |
+
seed: 0
|
9 |
+
num_workers: 4
|
10 |
+
num_att_plot: 3
|
11 |
+
dist_backend: nccl
|
12 |
+
dist_init_method: env://
|
13 |
+
dist_world_size: null
|
14 |
+
dist_rank: null
|
15 |
+
local_rank: 0
|
16 |
+
dist_master_addr: null
|
17 |
+
dist_master_port: null
|
18 |
+
dist_launcher: null
|
19 |
+
multiprocessing_distributed: false
|
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: 100
|
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: true
|
50 |
+
log_interval: null
|
51 |
+
use_matplotlib: true
|
52 |
+
use_tensorboard: true
|
53 |
+
create_graph_in_tensorboard: false
|
54 |
+
use_wandb: false
|
55 |
+
wandb_project: null
|
56 |
+
wandb_id: null
|
57 |
+
wandb_entity: null
|
58 |
+
wandb_name: null
|
59 |
+
wandb_model_log_interval: -1
|
60 |
+
detect_anomaly: false
|
61 |
+
pretrain_path: null
|
62 |
+
init_param: []
|
63 |
+
ignore_init_mismatch: false
|
64 |
+
freeze_param: []
|
65 |
+
num_iters_per_epoch: null
|
66 |
+
batch_size: 128
|
67 |
+
valid_batch_size: null
|
68 |
+
batch_bins: 1000000
|
69 |
+
valid_batch_bins: null
|
70 |
+
train_shape_file:
|
71 |
+
- exp/asr_stats_raw_it_char/train/speech_shape
|
72 |
+
- exp/asr_stats_raw_it_char/train/text_shape.char
|
73 |
+
valid_shape_file:
|
74 |
+
- exp/asr_stats_raw_it_char/valid/speech_shape
|
75 |
+
- exp/asr_stats_raw_it_char/valid/text_shape.char
|
76 |
+
batch_type: folded
|
77 |
+
valid_batch_type: null
|
78 |
+
fold_length:
|
79 |
+
- 80000
|
80 |
+
- 150
|
81 |
+
sort_in_batch: descending
|
82 |
+
sort_batch: descending
|
83 |
+
multiple_iterator: false
|
84 |
+
chunk_length: 500
|
85 |
+
chunk_shift_ratio: 0.5
|
86 |
+
num_cache_chunks: 1024
|
87 |
+
train_data_path_and_name_and_type:
|
88 |
+
- - dump/raw/tr_it/wav.scp
|
89 |
+
- speech
|
90 |
+
- sound
|
91 |
+
- - dump/raw/tr_it/text
|
92 |
+
- text
|
93 |
+
- text
|
94 |
+
valid_data_path_and_name_and_type:
|
95 |
+
- - dump/raw/dt_it/wav.scp
|
96 |
+
- speech
|
97 |
+
- sound
|
98 |
+
- - dump/raw/dt_it/text
|
99 |
+
- text
|
100 |
+
- text
|
101 |
+
allow_variable_data_keys: false
|
102 |
+
max_cache_size: 0.0
|
103 |
+
max_cache_fd: 32
|
104 |
+
valid_max_cache_size: null
|
105 |
+
optim: adam
|
106 |
+
optim_conf:
|
107 |
+
lr: 0.002
|
108 |
+
scheduler: warmuplr
|
109 |
+
scheduler_conf:
|
110 |
+
warmup_steps: 10000
|
111 |
+
token_list:
|
112 |
+
- <blank>
|
113 |
+
- <unk>
|
114 |
+
- <space>
|
115 |
+
- A
|
116 |
+
- E
|
117 |
+
- I
|
118 |
+
- O
|
119 |
+
- R
|
120 |
+
- N
|
121 |
+
- L
|
122 |
+
- S
|
123 |
+
- T
|
124 |
+
- C
|
125 |
+
- D
|
126 |
+
- U
|
127 |
+
- M
|
128 |
+
- P
|
129 |
+
- V
|
130 |
+
- G
|
131 |
+
- F
|
132 |
+
- H
|
133 |
+
- B
|
134 |
+
- Q
|
135 |
+
- Z
|
136 |
+
- ''''
|
137 |
+
- Ò
|
138 |
+
- À
|
139 |
+
- È
|
140 |
+
- Ú
|
141 |
+
- X
|
142 |
+
- W
|
143 |
+
- Í
|
144 |
+
- É
|
145 |
+
- Y
|
146 |
+
- K
|
147 |
+
- J
|
148 |
+
- '1'
|
149 |
+
- <sos/eos>
|
150 |
+
init: null
|
151 |
+
input_size: null
|
152 |
+
ctc_conf:
|
153 |
+
dropout_rate: 0.0
|
154 |
+
ctc_type: builtin
|
155 |
+
reduce: true
|
156 |
+
ignore_nan_grad: null
|
157 |
+
zero_infinity: true
|
158 |
+
joint_net_conf: null
|
159 |
+
use_preprocessor: true
|
160 |
+
token_type: char
|
161 |
+
bpemodel: null
|
162 |
+
non_linguistic_symbols: null
|
163 |
+
cleaner: null
|
164 |
+
g2p: null
|
165 |
+
speech_volume_normalize: null
|
166 |
+
rir_scp: null
|
167 |
+
rir_apply_prob: 1.0
|
168 |
+
noise_scp: null
|
169 |
+
noise_apply_prob: 1.0
|
170 |
+
noise_db_range: '13_15'
|
171 |
+
short_noise_thres: 0.5
|
172 |
+
frontend: default
|
173 |
+
frontend_conf:
|
174 |
+
fs: 16k
|
175 |
+
specaug: null
|
176 |
+
specaug_conf: {}
|
177 |
+
normalize: global_mvn
|
178 |
+
normalize_conf:
|
179 |
+
stats_file: exp/asr_stats_raw_it_char/train/feats_stats.npz
|
180 |
+
norm_vars: false
|
181 |
+
model: espnet
|
182 |
+
model_conf:
|
183 |
+
ctc_weight: 0.3
|
184 |
+
lsm_weight: 0.1
|
185 |
+
length_normalized_loss: false
|
186 |
+
preencoder: null
|
187 |
+
preencoder_conf: {}
|
188 |
+
encoder: conformer
|
189 |
+
encoder_conf:
|
190 |
+
output_size: 256
|
191 |
+
attention_heads: 4
|
192 |
+
linear_units: 1024
|
193 |
+
num_blocks: 15
|
194 |
+
dropout_rate: 0.1
|
195 |
+
positional_dropout_rate: 0.1
|
196 |
+
attention_dropout_rate: 0.1
|
197 |
+
input_layer: conv2d
|
198 |
+
normalize_before: true
|
199 |
+
macaron_style: true
|
200 |
+
rel_pos_type: latest
|
201 |
+
pos_enc_layer_type: rel_pos
|
202 |
+
selfattention_layer_type: rel_selfattn
|
203 |
+
activation_type: swish
|
204 |
+
use_cnn_module: true
|
205 |
+
cnn_module_kernel: 31
|
206 |
+
postencoder: null
|
207 |
+
postencoder_conf: {}
|
208 |
+
decoder: transformer
|
209 |
+
decoder_conf:
|
210 |
+
attention_heads: 4
|
211 |
+
linear_units: 2048
|
212 |
+
num_blocks: 6
|
213 |
+
dropout_rate: 0.1
|
214 |
+
positional_dropout_rate: 0.1
|
215 |
+
self_attention_dropout_rate: 0.0
|
216 |
+
src_attention_dropout_rate: 0.0
|
217 |
+
preprocessor: default
|
218 |
+
preprocessor_conf: {}
|
219 |
+
required:
|
220 |
+
- output_dir
|
221 |
+
- token_list
|
222 |
+
version: '202211'
|
223 |
+
distributed: false
|
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/acc.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/backward_time.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/cer.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/cer_ctc.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/forward_time.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/gpu_max_cached_mem_GB.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/iter_time.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/loss.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/loss_att.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/loss_ctc.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/optim0_lr0.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/optim_step_time.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/train_time.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/images/wer.png
ADDED
exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/valid.acc.ave_10best.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:557aba09cc97162703e876a6d3555aa976c4b67e19bbbe419a6b886c34c7e781
|
3 |
+
size 141098293
|
meta.yaml
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
espnet: '202211'
|
2 |
+
files:
|
3 |
+
asr_model_file: exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/valid.acc.ave_10best.pth
|
4 |
+
python: "3.9.15 (main, Nov 24 2022, 14:31:59) \n[GCC 11.2.0]"
|
5 |
+
timestamp: 1672297168.240548
|
6 |
+
torch: 1.12.1
|
7 |
+
yaml_files:
|
8 |
+
asr_train_config: exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse/config.yaml
|