zrjin commited on
Commit
3769e1c
1 Parent(s): b336aad

Upload 81 files

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/49epoch.pth +3 -0
  2. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/50epoch.pth +3 -0
  3. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/53epoch.pth +3 -0
  4. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/54epoch.pth +3 -0
  5. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/55epoch.pth +3 -0
  6. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/56epoch.pth +3 -0
  7. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/57epoch.pth +3 -0
  8. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/58epoch.pth +3 -0
  9. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/59epoch.pth +3 -0
  10. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/60epoch.pth +3 -0
  11. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/RESULTS.md +196 -0
  12. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/checkpoint.pth +3 -0
  13. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/config.yaml +227 -0
  14. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/acc.png +0 -0
  15. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/backward_time.png +0 -0
  16. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/cer.png +0 -0
  17. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/clip.png +0 -0
  18. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/forward_time.png +0 -0
  19. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/gpu_max_cached_mem_GB.png +0 -0
  20. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/grad_norm.png +0 -0
  21. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/iter_time.png +0 -0
  22. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/loss.png +0 -0
  23. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/loss_att.png +0 -0
  24. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/loss_scale.png +0 -0
  25. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/optim0_lr0.png +0 -0
  26. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/optim_step_time.png +0 -0
  27. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/train_time.png +0 -0
  28. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/wer.png +0 -0
  29. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/latest.pth +3 -0
  30. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/run.sh +1 -0
  31. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700307214.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1753232.0 +3 -0
  32. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700310367.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2045481.0 +3 -0
  33. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700412292.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1823743.0 +3 -0
  34. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700536345.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2992215.0 +3 -0
  35. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700307214.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1753232.1 +3 -0
  36. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700310367.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2045481.1 +3 -0
  37. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700412292.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1823743.1 +3 -0
  38. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700536345.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2992215.1 +3 -0
  39. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.1.log +773 -0
  40. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.2.log +0 -0
  41. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.3.log +1057 -0
  42. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.log +0 -0
  43. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/valid.acc.ave.pth +3 -0
  44. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/valid.acc.ave_10best.pth +3 -0
  45. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/valid.acc.best.pth +3 -0
  46. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/48epoch.pth +3 -0
  47. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/52epoch.pth +3 -0
  48. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/53epoch.pth +3 -0
  49. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/54epoch.pth +3 -0
  50. medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/55epoch.pth +3 -0
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/49epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:13942d8df6c2326cc0f4f5df1745ba9a6ee792b54ca65b85c322e9ebdad81439
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/50epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2a034944163cbf78cb911dd3e9e7e5561b8516429ce774c7669808a12a8f459b
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/53epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bc4a2014dacb245d48ec2e038cb352ff0902f18bc0156ce2b94eb1d321f2a5ab
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/54epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:50c499c23cef8489938aeb927367d479b55b109e852fde28fa17a14af64a8a76
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/55epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2930d5d0bb9b0029c8906d01748174b5663905ab969ee2137f019f2a72a02c7c
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/56epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d295b116b9891ef8520e7987d0682f79636305c96362e7067cfa63096f9e2e2e
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/57epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec090b65d90f017b160f5718d059cbbec01424491ae31ddb62e16255f64cc80a
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/58epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:07d4fb13b7ec8fcbc0e79b48c2f2e3ca03db84a62c6aecb96db6ec70d992daee
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/59epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63fced0fdc8817cd4f376fe8e33ffc3b07a8615ce2151f6b5bd6be0990bb75ff
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/60epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ab99d72231e8b471e6f66d902a88c36315286ecfc4bb77cf1efcd73bfe200db
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/RESULTS.md ADDED
@@ -0,0 +1,196 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <!-- Generated by scripts/utils/show_asr_result.sh -->
2
+ # RESULTS
3
+ ## Environments
4
+ - date: `Tue Mar 5 09:47:19 CST 2024`
5
+ - python version: `3.9.18 (main, Sep 11 2023, 13:41:44) [GCC 11.2.0]`
6
+ - espnet version: `espnet 202308`
7
+ - pytorch version: `pytorch 1.12.1+cu116`
8
+ - Git hash: `884659f9ee95374811015381c976fa3b4f6e01db`
9
+ - Commit date: `Thu Nov 23 00:23:29 2023 +0800`
10
+
11
+ ## exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new
12
+ ### WER
13
+
14
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
15
+ |---|---|---|---|---|---|---|---|---|
16
+ |decode_sot_asr_model_30epoch/test_clean_kaldi_fmt|961|64007|80.7|15.1|4.2|6.7|26.0|98.9|
17
+ |decode_sot_asr_model_30epoch/test_other_kaldi_fmt|992|80370|75.0|19.6|5.4|7.8|32.9|99.6|
18
+ |decode_sot_asr_model_35epoch/test_clean_kaldi_fmt|961|64007|81.5|14.5|4.0|7.4|26.0|98.6|
19
+ |decode_sot_asr_model_35epoch/test_other_kaldi_fmt|992|80370|75.7|18.8|5.5|9.0|33.2|99.4|
20
+ |decode_sot_asr_model_40epoch/test_clean_kaldi_fmt|961|64007|80.4|14.2|5.4|4.8|24.4|97.7|
21
+ |decode_sot_asr_model_40epoch/test_other_kaldi_fmt|992|80370|74.8|18.1|7.1|5.7|30.9|99.5|
22
+ |decode_sot_asr_model_8epoch/test_clean_kaldi_fmt|961|64007|70.4|23.7|5.8|7.2|36.7|99.1|
23
+ |decode_sot_asr_model_8epoch/test_other_kaldi_fmt|992|80370|63.7|28.7|7.5|8.6|44.8|99.6|
24
+ |decode_sot_asr_model_valid.acc.best/dev|3000|126853|54.3|32.2|13.4|28.9|74.6|100.0|
25
+ |decode_sot_asr_model_valid.acc.best/dev_2spk|3315|226216|76.8|11.5|11.6|12.4|35.5|99.2|
26
+ |decode_sot_asr_model_valid.acc.best/dev_2spk_sys7_8khz_spk1|1606|135101|36.6|21.7|41.7|3.2|66.7|100.0|
27
+ |decode_sot_asr_model_valid.acc.best/dev_2spk_sys7_8khz_spk2|1606|135101|35.5|21.5|43.0|3.4|67.9|100.0|
28
+ |decode_sot_asr_model_valid.acc.best/dev_3spk|2059|209679|63.9|20.3|15.8|10.4|46.5|100.0|
29
+ |decode_sot_asr_model_valid.acc.best/dev_3spk_kaldi_fmt|1004|124462|67.6|17.1|15.2|7.2|39.6|100.0|
30
+ |decode_sot_asr_model_valid.acc.best/dev_4spk|1467|200029|52.0|27.4|20.7|11.8|59.9|100.0|
31
+ |decode_sot_asr_model_valid.acc.best/dev_4spk_kaldi_fmt|721|119166|55.5|22.9|21.6|11.3|55.8|100.0|
32
+ |decode_sot_asr_model_valid.acc.best/dev_oracle|544|10798|85.9|12.2|1.9|88.7|102.9|92.8|
33
+ |decode_sot_asr_model_valid.acc.best/eval_oracle|4479|96585|84.7|13.0|2.4|88.2|103.6|94.5|
34
+ |decode_sot_asr_model_valid.acc.best/sot_sdm1_dev|2382|35243|0.0|0.0|100.0|0.0|100.0|100.0|
35
+ |decode_sot_asr_model_valid.acc.best/test-clean_2spk|4570|301042|77.5|10.9|11.6|11.3|33.8|99.5|
36
+ |decode_sot_asr_model_valid.acc.best/test-clean_3spk|2072|212871|64.4|19.3|16.3|11.8|47.4|100.0|
37
+ |decode_sot_asr_model_valid.acc.best/test-clean_4spk|1326|185394|53.2|26.2|20.6|10.6|57.4|100.0|
38
+ |decode_sot_asr_model_valid.acc.best/test-other_2spk|4663|336490|75.9|13.3|10.8|11.6|35.6|99.9|
39
+ |decode_sot_asr_model_valid.acc.best/test-other_3spk|2453|266074|60.5|23.6|15.9|11.5|51.0|100.0|
40
+ |decode_sot_asr_model_valid.acc.best/test-other_4spk|1795|259138|49.2|30.2|20.6|11.1|61.9|100.0|
41
+ |decode_sot_asr_model_valid.acc.best/test|3000|114243|55.6|30.7|13.7|32.9|77.3|99.9|
42
+ |decode_sot_asr_model_valid.acc.best/test_clean_2spk_kaldi_fmt|2180|178761|80.2|9.0|10.8|7.5|27.2|98.9|
43
+ |decode_sot_asr_model_valid.acc.best/test_clean_2spk_sys7_8khz_spk1|2180|178761|33.1|21.4|45.4|3.0|69.9|100.0|
44
+ |decode_sot_asr_model_valid.acc.best/test_clean_2spk_sys7_8khz_spk2|2180|178761|31.6|24.4|44.0|3.0|71.4|100.0|
45
+ |decode_sot_asr_model_valid.acc.best/test_clean_3spk_kaldi_fmt|977|124741|66.8|17.5|15.7|9.0|42.2|100.0|
46
+ |decode_sot_asr_model_valid.acc.best/test_clean_4spk_kaldi_fmt|632|109072|56.3|22.4|21.3|8.8|52.5|100.0|
47
+ |decode_sot_asr_model_valid.acc.best/test_clean_kaldi_fmt|961|64007|81.4|13.3|5.3|5.9|24.5|97.5|
48
+ |decode_sot_asr_model_valid.acc.best/test_other_2spk_kaldi_fmt|2363|205496|78.6|11.7|9.6|7.8|29.1|99.8|
49
+ |decode_sot_asr_model_valid.acc.best/test_other_2spk_sys7_8khz_spk1|2363|205496|28.2|26.2|45.6|3.1|74.9|100.0|
50
+ |decode_sot_asr_model_valid.acc.best/test_other_2spk_sys7_8khz_spk2|2363|205496|35.6|22.3|42.1|3.1|67.5|100.0|
51
+ |decode_sot_asr_model_valid.acc.best/test_other_3spk_kaldi_fmt|1246|162996|62.5|21.9|15.6|9.1|46.6|100.0|
52
+ |decode_sot_asr_model_valid.acc.best/test_other_4spk_kaldi_fmt|901|157123|51.9|26.0|22.1|10.4|58.5|100.0|
53
+ |decode_sot_asr_model_valid.acc.best/test_other_kaldi_fmt|992|80370|75.5|17.6|6.9|6.8|31.3|99.3|
54
+ |decode_sot_css_asr_model_valid.acc.best/dev_oracle|544|10798|85.0|12.9|2.1|48.0|63.0|92.8|
55
+ |decode_sot_css_asr_model_valid.acc.best/eval_oracle|4479|96585|84.1|13.2|2.7|49.1|65.0|94.5|
56
+
57
+ ### CER
58
+
59
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
60
+ |---|---|---|---|---|---|---|---|---|
61
+ |decode_sot_asr_model_30epoch/test_clean_kaldi_fmt|961|329390|89.3|6.0|4.6|6.7|17.4|98.9|
62
+ |decode_sot_asr_model_30epoch/test_other_kaldi_fmt|992|416899|85.6|8.2|6.3|7.7|22.2|99.6|
63
+ |decode_sot_asr_model_35epoch/test_clean_kaldi_fmt|961|329390|89.8|5.8|4.4|7.4|17.6|98.6|
64
+ |decode_sot_asr_model_35epoch/test_other_kaldi_fmt|992|416899|85.9|8.0|6.1|8.5|22.6|99.4|
65
+ |decode_sot_asr_model_40epoch/test_clean_kaldi_fmt|961|329390|88.6|5.9|5.5|4.5|16.0|97.7|
66
+ |decode_sot_asr_model_40epoch/test_other_kaldi_fmt|992|416899|84.5|8.0|7.5|5.2|20.7|99.5|
67
+ |decode_sot_asr_model_8epoch/test_clean_kaldi_fmt|961|329390|83.8|9.3|6.9|7.2|23.5|99.1|
68
+ |decode_sot_asr_model_8epoch/test_other_kaldi_fmt|992|416899|78.7|12.1|9.2|8.2|29.5|99.6|
69
+ |decode_sot_asr_model_valid.acc.best/dev|3000|673222|71.1|13.3|15.6|28.5|57.5|100.0|
70
+ |decode_sot_asr_model_valid.acc.best/dev_2spk|3315|1230801|83.7|6.2|10.1|10.8|27.2|99.2|
71
+ |decode_sot_asr_model_valid.acc.best/dev_2spk_sys7_8khz_spk1|1606|735694|47.9|9.1|43.0|3.6|55.7|100.0|
72
+ |decode_sot_asr_model_valid.acc.best/dev_2spk_sys7_8khz_spk2|1606|735694|47.1|9.1|43.8|3.7|56.5|100.0|
73
+ |decode_sot_asr_model_valid.acc.best/dev_3spk|2059|1140428|74.6|10.3|15.2|9.3|34.7|100.0|
74
+ |decode_sot_asr_model_valid.acc.best/dev_3spk_kaldi_fmt|1004|677017|76.8|8.2|15.0|6.2|29.4|100.0|
75
+ |decode_sot_asr_model_valid.acc.best/dev_4spk|1467|1087409|65.5|13.1|21.3|10.3|44.7|100.0|
76
+ |decode_sot_asr_model_valid.acc.best/dev_4spk_kaldi_fmt|721|647884|67.3|10.5|22.2|8.9|41.6|100.0|
77
+ |decode_sot_asr_model_valid.acc.best/dev_oracle|544|57590|94.8|3.2|2.0|86.2|91.5|92.8|
78
+ |decode_sot_asr_model_valid.acc.best/eval_oracle|4479|522239|93.8|3.6|2.6|85.2|91.4|94.5|
79
+ |decode_sot_asr_model_valid.acc.best/sot_sdm1_dev|2382|169857|0.0|0.0|100.0|0.0|100.0|100.0|
80
+ |decode_sot_asr_model_valid.acc.best/test-clean_2spk|4570|1550429|84.0|6.1|10.0|9.9|25.9|99.5|
81
+ |decode_sot_asr_model_valid.acc.best/test-clean_3spk|2072|1084475|74.6|10.4|15.1|10.2|35.7|100.0|
82
+ |decode_sot_asr_model_valid.acc.best/test-clean_4spk|1326|938467|66.3|13.1|20.7|9.9|43.6|100.0|
83
+ |decode_sot_asr_model_valid.acc.best/test-other_2spk|4663|1742136|83.5|7.0|9.4|10.0|26.5|99.9|
84
+ |decode_sot_asr_model_valid.acc.best/test-other_3spk|2453|1381987|72.6|12.0|15.4|10.0|37.4|100.0|
85
+ |decode_sot_asr_model_valid.acc.best/test-other_4spk|1795|1346646|63.9|14.4|21.7|9.9|46.0|100.0|
86
+ |decode_sot_asr_model_valid.acc.best/test|3000|608408|71.7|12.5|15.8|32.4|60.7|99.9|
87
+ |decode_sot_asr_model_valid.acc.best/test_clean_2spk_kaldi_fmt|2180|921344|85.5|5.0|9.5|6.1|20.6|98.9|
88
+ |decode_sot_asr_model_valid.acc.best/test_clean_2spk_sys7_8khz_spk1|2180|921344|44.6|8.9|46.5|3.5|58.8|100.0|
89
+ |decode_sot_asr_model_valid.acc.best/test_clean_2spk_sys7_8khz_spk2|2180|921344|44.9|10.2|44.9|3.7|58.9|100.0|
90
+ |decode_sot_asr_model_valid.acc.best/test_clean_3spk_kaldi_fmt|977|635802|76.0|9.0|15.0|7.7|31.7|100.0|
91
+ |decode_sot_asr_model_valid.acc.best/test_clean_4spk_kaldi_fmt|632|552325|67.8|10.6|21.6|7.8|39.9|100.0|
92
+ |decode_sot_asr_model_valid.acc.best/test_clean_kaldi_fmt|961|329390|89.2|5.3|5.4|5.8|16.6|97.5|
93
+ |decode_sot_asr_model_valid.acc.best/test_other_2spk_kaldi_fmt|2363|1064868|85.3|6.1|8.6|6.4|21.1|99.8|
94
+ |decode_sot_asr_model_valid.acc.best/test_other_2spk_sys7_8khz_spk1|2363|1064868|42.2|11.1|46.7|3.8|61.6|100.0|
95
+ |decode_sot_asr_model_valid.acc.best/test_other_2spk_sys7_8khz_spk2|2363|1064868|47.8|9.2|43.0|3.7|55.9|100.0|
96
+ |decode_sot_asr_model_valid.acc.best/test_other_3spk_kaldi_fmt|1246|847159|73.8|10.5|15.6|7.9|34.1|100.0|
97
+ |decode_sot_asr_model_valid.acc.best/test_other_4spk_kaldi_fmt|901|817228|65.0|11.9|23.1|8.7|43.7|100.0|
98
+ |decode_sot_asr_model_valid.acc.best/test_other_kaldi_fmt|992|416899|85.1|7.3|7.6|6.8|21.7|99.3|
99
+ |decode_sot_css_asr_model_valid.acc.best/dev_oracle|544|57590|93.7|4.0|2.4|46.5|52.8|92.8|
100
+ |decode_sot_css_asr_model_valid.acc.best/eval_oracle|4479|522239|92.9|4.1|3.0|47.2|54.3|94.5|
101
+
102
+ ### TER
103
+
104
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
105
+ |---|---|---|---|---|---|---|---|---|
106
+ ## exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/decode_sot_asr_model_30epoch
107
+ ### WER
108
+
109
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
110
+ |---|---|---|---|---|---|---|---|---|
111
+ |org/dev_kaldi_fmt|605|47659|78.4|16.5|5.1|5.5|27.1|97.2|
112
+
113
+ ### CER
114
+
115
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
116
+ |---|---|---|---|---|---|---|---|---|
117
+ |org/dev_kaldi_fmt|605|258151|87.5|6.5|6.1|5.6|18.1|97.2|
118
+
119
+ ### TER
120
+
121
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
122
+ |---|---|---|---|---|---|---|---|---|
123
+ ## exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/decode_sot_asr_model_35epoch
124
+ ### WER
125
+
126
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
127
+ |---|---|---|---|---|---|---|---|---|
128
+ |org/dev_kaldi_fmt|605|47659|79.4|16.2|4.4|6.8|27.4|98.5|
129
+
130
+ ### CER
131
+
132
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
133
+ |---|---|---|---|---|---|---|---|---|
134
+ |org/dev_kaldi_fmt|605|258151|88.2|6.5|5.3|6.7|18.5|98.5|
135
+
136
+ ### TER
137
+
138
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
139
+ |---|---|---|---|---|---|---|---|---|
140
+ ## exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/decode_sot_asr_model_40epoch
141
+ ### WER
142
+
143
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
144
+ |---|---|---|---|---|---|---|---|---|
145
+ |org/dev_kaldi_fmt|605|47659|78.5|15.6|5.8|4.0|25.5|97.5|
146
+
147
+ ### CER
148
+
149
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
150
+ |---|---|---|---|---|---|---|---|---|
151
+ |org/dev_kaldi_fmt|605|258151|87.0|6.3|6.7|3.9|16.9|97.5|
152
+
153
+ ### TER
154
+
155
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
156
+ |---|---|---|---|---|---|---|---|---|
157
+ ## exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/decode_sot_asr_model_8epoch
158
+ ### WER
159
+
160
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
161
+ |---|---|---|---|---|---|---|---|---|
162
+ |org/dev_kaldi_fmt|605|47659|68.4|24.9|6.7|6.7|38.2|98.3|
163
+
164
+ ### CER
165
+
166
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
167
+ |---|---|---|---|---|---|---|---|---|
168
+ |org/dev_kaldi_fmt|605|258151|81.6|9.9|8.5|6.7|25.2|98.3|
169
+
170
+ ### TER
171
+
172
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
173
+ |---|---|---|---|---|---|---|---|---|
174
+ ## exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/decode_sot_asr_model_valid.acc.best
175
+ ### WER
176
+
177
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
178
+ |---|---|---|---|---|---|---|---|---|
179
+ |org/dev_2spk_kaldi_fmt|1606|135101|79.1|9.8|11.1|8.4|29.2|98.4|
180
+ |org/dev_kaldi_fmt|605|47659|78.8|15.0|6.2|5.7|26.9|97.5|
181
+ |org/sot_sdm1_eval|2385|37529|33.4|53.9|12.7|102.0|168.6|100.0|
182
+ |org/tt_mix_clean_reverb_max_16k|3000|3000|0.0|100.0|0.0|4047.2|4147.2|100.0|
183
+
184
+ ### CER
185
+
186
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
187
+ |---|---|---|---|---|---|---|---|---|
188
+ |org/dev_2spk_kaldi_fmt|1606|735694|84.8|5.3|9.9|6.9|22.1|98.4|
189
+ |org/dev_kaldi_fmt|605|258151|87.1|6.0|6.9|5.7|18.6|97.5|
190
+ |org/sot_sdm1_eval|2385|183036|62.5|22.7|14.8|97.8|135.2|100.0|
191
+ |org/tt_mix_clean_reverb_max_16k|3000|143026|17.3|82.6|0.1|411.1|493.8|100.0|
192
+
193
+ ### TER
194
+
195
+ |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
196
+ |---|---|---|---|---|---|---|---|---|
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/checkpoint.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14eb244b2d0dfc6d2dc23b8b9a10e1814df51f163d680c62439ff31374e6bd2e
3
+ size 516820547
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/config.yaml ADDED
@@ -0,0 +1,227 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ config: conf/tuning/train_sot_asr_conformer_medium.yaml
2
+ print_config: false
3
+ log_level: INFO
4
+ drop_last_iter: false
5
+ dry_run: false
6
+ iterator_type: sequence
7
+ valid_iterator_type: null
8
+ output_dir: exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new
9
+ ngpu: 1
10
+ seed: 0
11
+ num_workers: 16
12
+ num_att_plot: 3
13
+ dist_backend: nccl
14
+ dist_init_method: env://
15
+ dist_world_size: 4
16
+ dist_rank: 0
17
+ local_rank: 0
18
+ dist_master_addr: localhost
19
+ dist_master_port: 44319
20
+ dist_launcher: null
21
+ multiprocessing_distributed: true
22
+ unused_parameters: false
23
+ sharded_ddp: false
24
+ cudnn_enabled: true
25
+ cudnn_benchmark: false
26
+ cudnn_deterministic: true
27
+ collect_stats: false
28
+ write_collected_feats: false
29
+ max_epoch: 60
30
+ patience: null
31
+ val_scheduler_criterion:
32
+ - valid
33
+ - loss
34
+ early_stopping_criterion:
35
+ - valid
36
+ - loss
37
+ - min
38
+ best_model_criterion:
39
+ - - valid
40
+ - acc
41
+ - max
42
+ keep_nbest_models: 10
43
+ nbest_averaging_interval: 0
44
+ grad_clip: 5.0
45
+ grad_clip_type: 2.0
46
+ grad_noise: false
47
+ accum_grad: 4
48
+ no_forward_run: false
49
+ resume: true
50
+ train_dtype: float32
51
+ use_amp: false
52
+ log_interval: null
53
+ use_matplotlib: true
54
+ use_tensorboard: true
55
+ create_graph_in_tensorboard: false
56
+ use_wandb: false
57
+ wandb_project: null
58
+ wandb_id: null
59
+ wandb_entity: null
60
+ wandb_name: null
61
+ wandb_model_log_interval: -1
62
+ detect_anomaly: false
63
+ pretrain_path: null
64
+ init_param:
65
+ - /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth
66
+ ignore_init_mismatch: false
67
+ freeze_param: []
68
+ num_iters_per_epoch: null
69
+ batch_size: 20
70
+ valid_batch_size: null
71
+ batch_bins: 16000000
72
+ valid_batch_bins: null
73
+ train_shape_file:
74
+ - exp/asr_stats_raw_en_char_sp/train/speech_shape
75
+ - exp/asr_stats_raw_en_char_sp/train/text_shape.char
76
+ valid_shape_file:
77
+ - exp/asr_stats_raw_en_char_sp/valid/speech_shape
78
+ - exp/asr_stats_raw_en_char_sp/valid/text_shape.char
79
+ batch_type: numel
80
+ valid_batch_type: null
81
+ fold_length:
82
+ - 80000
83
+ - 150
84
+ sort_in_batch: descending
85
+ shuffle_within_batch: false
86
+ sort_batch: descending
87
+ multiple_iterator: false
88
+ chunk_length: 500
89
+ chunk_shift_ratio: 0.5
90
+ num_cache_chunks: 1024
91
+ chunk_excluded_key_prefixes: []
92
+ train_data_path_and_name_and_type:
93
+ - - dump/raw/train_medium_kaldi_fmt_sp/wav.scp
94
+ - speech
95
+ - kaldi_ark
96
+ - - dump/raw/train_medium_kaldi_fmt_sp/text
97
+ - text
98
+ - text
99
+ valid_data_path_and_name_and_type:
100
+ - - dump/raw/dev_kaldi_fmt/wav.scp
101
+ - speech
102
+ - kaldi_ark
103
+ - - dump/raw/dev_kaldi_fmt/text
104
+ - text
105
+ - text
106
+ allow_variable_data_keys: false
107
+ max_cache_size: 0.0
108
+ max_cache_fd: 32
109
+ valid_max_cache_size: null
110
+ exclude_weight_decay: false
111
+ exclude_weight_decay_conf: {}
112
+ optim: adam
113
+ optim_conf:
114
+ lr: 0.002
115
+ weight_decay: 1.0e-06
116
+ scheduler: warmuplr
117
+ scheduler_conf:
118
+ warmup_steps: 20000
119
+ token_list:
120
+ - <blank>
121
+ - <unk>
122
+ - <sc>
123
+ - <space>
124
+ - E
125
+ - T
126
+ - A
127
+ - O
128
+ - N
129
+ - I
130
+ - H
131
+ - S
132
+ - R
133
+ - D
134
+ - L
135
+ - U
136
+ - M
137
+ - C
138
+ - W
139
+ - F
140
+ - G
141
+ - Y
142
+ - P
143
+ - B
144
+ - V
145
+ - K
146
+ - ''''
147
+ - X
148
+ - J
149
+ - Q
150
+ - Z
151
+ - <sos/eos>
152
+ init: null
153
+ input_size: null
154
+ ctc_conf:
155
+ dropout_rate: 0.0
156
+ ctc_type: builtin
157
+ reduce: true
158
+ ignore_nan_grad: null
159
+ zero_infinity: true
160
+ joint_net_conf: null
161
+ use_preprocessor: true
162
+ token_type: char
163
+ bpemodel: null
164
+ non_linguistic_symbols: null
165
+ cleaner: null
166
+ g2p: null
167
+ speech_volume_normalize: null
168
+ rir_scp: null
169
+ rir_apply_prob: 1.0
170
+ noise_scp: null
171
+ noise_apply_prob: 1.0
172
+ noise_db_range: '13_15'
173
+ short_noise_thres: 0.5
174
+ aux_ctc_tasks: []
175
+ frontend: default
176
+ frontend_conf:
177
+ fs: 16k
178
+ specaug: null
179
+ specaug_conf: {}
180
+ normalize: global_mvn
181
+ normalize_conf:
182
+ stats_file: exp/asr_stats_raw_en_char_sp/train/feats_stats.npz
183
+ model: espnet
184
+ model_conf:
185
+ ctc_weight: 0.0
186
+ lsm_weight: 0.1
187
+ length_normalized_loss: false
188
+ preencoder: null
189
+ preencoder_conf: {}
190
+ encoder: conformer
191
+ encoder_conf:
192
+ output_size: 256
193
+ attention_heads: 4
194
+ linear_units: 2048
195
+ num_blocks: 12
196
+ dropout_rate: 0.1
197
+ positional_dropout_rate: 0.1
198
+ attention_dropout_rate: 0.1
199
+ input_layer: conv2d
200
+ normalize_before: true
201
+ macaron_style: true
202
+ rel_pos_type: latest
203
+ pos_enc_layer_type: rel_pos
204
+ selfattention_layer_type: rel_selfattn
205
+ activation_type: swish
206
+ use_cnn_module: true
207
+ cnn_module_kernel: 31
208
+ postencoder: null
209
+ postencoder_conf: {}
210
+ decoder: transformer
211
+ decoder_conf:
212
+ attention_heads: 4
213
+ linear_units: 2048
214
+ num_blocks: 6
215
+ dropout_rate: 0.1
216
+ positional_dropout_rate: 0.1
217
+ self_attention_dropout_rate: 0.1
218
+ src_attention_dropout_rate: 0.1
219
+ preprocessor: multi
220
+ preprocessor_conf:
221
+ speaker_change_symbol:
222
+ - <sc>
223
+ required:
224
+ - output_dir
225
+ - token_list
226
+ version: '202308'
227
+ distributed: true
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/acc.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/backward_time.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/cer.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/clip.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/forward_time.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/gpu_max_cached_mem_GB.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/grad_norm.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/iter_time.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/loss.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/loss_att.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/loss_scale.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/optim0_lr0.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/optim_step_time.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/train_time.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/images/wer.png ADDED
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/latest.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1ab99d72231e8b471e6f66d902a88c36315286ecfc4bb77cf1efcd73bfe200db
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/run.sh ADDED
@@ -0,0 +1 @@
 
 
1
+ ./asr.sh --lang en --audio_format flac.ark --stage 12 --feats_type raw --token_type char --sot_asr true --max_wav_duration 50 --speed_perturb_factors '0.9 1.0 1.1' --feats_normalize global_mvn --use_lm false --pretrained_model /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth --asr_config conf/tuning/train_sot_asr_conformer_medium.yaml --lm_config conf/tuning/train_lm_transformer.yaml --inference_config conf/tuning/decode_sot.yaml --train_set train_medium_kaldi_fmt --valid_set dev_kaldi_fmt --test_sets 'dev_kaldi_fmt test_clean_kaldi_fmt test_other_kaldi_fmt' --ngpu 4 --asr_tag train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new --lm_train_text data/local/other_text/text --bpe_train_text data/train_medium_kaldi_fmt/text --stage 12 "$@"; exit $?
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700307214.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1753232.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c30cf181e3c82c4cc12d7aebfb77b819aa2da1fd08a2d567519ad1b5196e3171
3
+ size 88
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700310367.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2045481.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0848df7dac3f565c375a3f38dc7a661fd41628f61bffb3d75e70e3641706faec
3
+ size 107387014
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700412292.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1823743.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c40b67a8b2a99f63326ec1b7b8685ca88cc3ed776a10e39c5273e164c8982b67
3
+ size 8306
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/train/events.out.tfevents.1700536345.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2992215.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:08432dbcf98ebaad27388a1b90b781f1aef124b1ac2c3c4676f8a387c81c0c1a
3
+ size 743299213
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700307214.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1753232.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05d778de3f4cc8bdfe77155d9f1e2d6782358f94fdcf930fd16cb4b391e0f81f
3
+ size 88
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700310367.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2045481.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3ead35607a81bfff4e5bab9f6dd496480516f7350927ad98a7a4db115507b2ee
3
+ size 2338
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700412292.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.1823743.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e422f849535779325246177b90cb5193453b58ac6cceb2e15640105036c361a8
3
+ size 88
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/tensorboard/valid/events.out.tfevents.1700536345.de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb.2992215.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:143371316cfd5edd14fe5c0cc235d84980859e4dd329af2b61fcdd64eb7b93fb
3
+ size 14752
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.1.log ADDED
@@ -0,0 +1,773 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python3 -m espnet2.bin.asr_train --use_preprocessor true --bpemodel none --token_type char --token_list data/en_token_list/char/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/wav.scp,speech,kaldi_ark --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/speech_shape --resume true --init_param /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth --ignore_init_mismatch false --fold_length 80000 --output_dir exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new --config conf/tuning/train_sot_asr_conformer_medium.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/asr_stats_raw_en_char_sp/train/feats_stats.npz --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/wav.scp,speech,kaldi_ark --train_shape_file exp/asr_stats_raw_en_char_sp/train/speech_shape --fold_length 150 --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/text,text,text --train_shape_file exp/asr_stats_raw_en_char_sp/train/text_shape.char --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/text,text,text --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/text_shape.char --ngpu 4 --multiprocessing_distributed True
2
+ # Started at Mon Nov 20 00:43:12 CST 2023
3
+ #
4
+ /star-home/jinzengrui/lib/miniconda3/envs/dev39/bin/python3 /star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py --use_preprocessor true --bpemodel none --token_type char --token_list data/en_token_list/char/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/wav.scp,speech,kaldi_ark --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/speech_shape --resume true --init_param /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth --ignore_init_mismatch false --fold_length 80000 --output_dir exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new --config conf/tuning/train_sot_asr_conformer_medium.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/asr_stats_raw_en_char_sp/train/feats_stats.npz --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/wav.scp,speech,kaldi_ark --train_shape_file exp/asr_stats_raw_en_char_sp/train/speech_shape --fold_length 150 --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/text,text,text --train_shape_file exp/asr_stats_raw_en_char_sp/train/text_shape.char --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/text,text,text --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/text_shape.char --ngpu 4 --multiprocessing_distributed True
5
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:23,119 (distributed_c10d:228) INFO: Added key: store_based_barrier_key:1 to store for rank: 0
6
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:23,119 (distributed_c10d:262) INFO: Rank 0: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes.
7
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:23,174 (asr:490) INFO: Vocabulary size: 32
8
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:31,610 (abs_task:1229) INFO: pytorch.version=1.11.0+cu102, cuda.available=True, cudnn.version=7605, cudnn.benchmark=False, cudnn.deterministic=True
9
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:31,624 (abs_task:1230) INFO: Model structure:
10
+ ESPnetASRModel(
11
+ (frontend): DefaultFrontend(
12
+ (stft): Stft(n_fft=512, win_length=512, hop_length=128, center=True, normalized=False, onesided=True)
13
+ (frontend): Frontend()
14
+ (logmel): LogMel(sr=16000, n_fft=512, n_mels=80, fmin=0, fmax=8000.0, htk=False)
15
+ )
16
+ (normalize): GlobalMVN(stats_file=exp/asr_stats_raw_en_char_sp/train/feats_stats.npz, norm_means=True, norm_vars=True)
17
+ (encoder): ConformerEncoder(
18
+ (embed): Conv2dSubsampling(
19
+ (conv): Sequential(
20
+ (0): Conv2d(1, 256, kernel_size=(3, 3), stride=(2, 2))
21
+ (1): ReLU()
22
+ (2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2))
23
+ (3): ReLU()
24
+ )
25
+ (out): Sequential(
26
+ (0): Linear(in_features=4864, out_features=256, bias=True)
27
+ (1): RelPositionalEncoding(
28
+ (dropout): Dropout(p=0.1, inplace=False)
29
+ )
30
+ )
31
+ )
32
+ (encoders): MultiSequential(
33
+ (0): EncoderLayer(
34
+ (self_attn): RelPositionMultiHeadedAttention(
35
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
36
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
37
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
38
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
39
+ (dropout): Dropout(p=0.1, inplace=False)
40
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
41
+ )
42
+ (feed_forward): PositionwiseFeedForward(
43
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
44
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
45
+ (dropout): Dropout(p=0.1, inplace=False)
46
+ (activation): Swish()
47
+ )
48
+ (feed_forward_macaron): PositionwiseFeedForward(
49
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
50
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
51
+ (dropout): Dropout(p=0.1, inplace=False)
52
+ (activation): Swish()
53
+ )
54
+ (conv_module): ConvolutionModule(
55
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
56
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
57
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
58
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
59
+ (activation): Swish()
60
+ )
61
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
62
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
63
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
64
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
65
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
66
+ (dropout): Dropout(p=0.1, inplace=False)
67
+ )
68
+ (1): EncoderLayer(
69
+ (self_attn): RelPositionMultiHeadedAttention(
70
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
71
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
72
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
73
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
74
+ (dropout): Dropout(p=0.1, inplace=False)
75
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
76
+ )
77
+ (feed_forward): PositionwiseFeedForward(
78
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
79
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
80
+ (dropout): Dropout(p=0.1, inplace=False)
81
+ (activation): Swish()
82
+ )
83
+ (feed_forward_macaron): PositionwiseFeedForward(
84
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
85
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
86
+ (dropout): Dropout(p=0.1, inplace=False)
87
+ (activation): Swish()
88
+ )
89
+ (conv_module): ConvolutionModule(
90
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
91
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
92
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
93
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
94
+ (activation): Swish()
95
+ )
96
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
97
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
98
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
99
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
100
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
101
+ (dropout): Dropout(p=0.1, inplace=False)
102
+ )
103
+ (2): EncoderLayer(
104
+ (self_attn): RelPositionMultiHeadedAttention(
105
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
106
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
107
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
108
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
109
+ (dropout): Dropout(p=0.1, inplace=False)
110
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
111
+ )
112
+ (feed_forward): PositionwiseFeedForward(
113
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
114
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
115
+ (dropout): Dropout(p=0.1, inplace=False)
116
+ (activation): Swish()
117
+ )
118
+ (feed_forward_macaron): PositionwiseFeedForward(
119
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
120
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
121
+ (dropout): Dropout(p=0.1, inplace=False)
122
+ (activation): Swish()
123
+ )
124
+ (conv_module): ConvolutionModule(
125
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
126
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
127
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
128
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
129
+ (activation): Swish()
130
+ )
131
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
132
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
133
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
134
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
135
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
136
+ (dropout): Dropout(p=0.1, inplace=False)
137
+ )
138
+ (3): EncoderLayer(
139
+ (self_attn): RelPositionMultiHeadedAttention(
140
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
141
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
142
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
143
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
144
+ (dropout): Dropout(p=0.1, inplace=False)
145
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
146
+ )
147
+ (feed_forward): PositionwiseFeedForward(
148
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
149
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
150
+ (dropout): Dropout(p=0.1, inplace=False)
151
+ (activation): Swish()
152
+ )
153
+ (feed_forward_macaron): PositionwiseFeedForward(
154
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
155
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
156
+ (dropout): Dropout(p=0.1, inplace=False)
157
+ (activation): Swish()
158
+ )
159
+ (conv_module): ConvolutionModule(
160
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
161
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
162
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
163
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
164
+ (activation): Swish()
165
+ )
166
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
167
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
168
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
169
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
170
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
171
+ (dropout): Dropout(p=0.1, inplace=False)
172
+ )
173
+ (4): EncoderLayer(
174
+ (self_attn): RelPositionMultiHeadedAttention(
175
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
176
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
177
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
178
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
179
+ (dropout): Dropout(p=0.1, inplace=False)
180
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
181
+ )
182
+ (feed_forward): PositionwiseFeedForward(
183
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
184
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
185
+ (dropout): Dropout(p=0.1, inplace=False)
186
+ (activation): Swish()
187
+ )
188
+ (feed_forward_macaron): PositionwiseFeedForward(
189
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
190
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
191
+ (dropout): Dropout(p=0.1, inplace=False)
192
+ (activation): Swish()
193
+ )
194
+ (conv_module): ConvolutionModule(
195
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
196
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
197
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
198
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
199
+ (activation): Swish()
200
+ )
201
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
202
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
203
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
204
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
205
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
206
+ (dropout): Dropout(p=0.1, inplace=False)
207
+ )
208
+ (5): EncoderLayer(
209
+ (self_attn): RelPositionMultiHeadedAttention(
210
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
211
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
212
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
213
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
214
+ (dropout): Dropout(p=0.1, inplace=False)
215
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
216
+ )
217
+ (feed_forward): PositionwiseFeedForward(
218
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
219
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
220
+ (dropout): Dropout(p=0.1, inplace=False)
221
+ (activation): Swish()
222
+ )
223
+ (feed_forward_macaron): PositionwiseFeedForward(
224
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
225
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
226
+ (dropout): Dropout(p=0.1, inplace=False)
227
+ (activation): Swish()
228
+ )
229
+ (conv_module): ConvolutionModule(
230
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
231
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
232
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
233
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
234
+ (activation): Swish()
235
+ )
236
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
237
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
238
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
239
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
240
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
241
+ (dropout): Dropout(p=0.1, inplace=False)
242
+ )
243
+ (6): EncoderLayer(
244
+ (self_attn): RelPositionMultiHeadedAttention(
245
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
246
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
247
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
248
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
249
+ (dropout): Dropout(p=0.1, inplace=False)
250
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
251
+ )
252
+ (feed_forward): PositionwiseFeedForward(
253
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
254
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
255
+ (dropout): Dropout(p=0.1, inplace=False)
256
+ (activation): Swish()
257
+ )
258
+ (feed_forward_macaron): PositionwiseFeedForward(
259
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
260
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
261
+ (dropout): Dropout(p=0.1, inplace=False)
262
+ (activation): Swish()
263
+ )
264
+ (conv_module): ConvolutionModule(
265
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
266
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
267
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
268
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
269
+ (activation): Swish()
270
+ )
271
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
272
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
273
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
274
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
275
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
276
+ (dropout): Dropout(p=0.1, inplace=False)
277
+ )
278
+ (7): EncoderLayer(
279
+ (self_attn): RelPositionMultiHeadedAttention(
280
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
281
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
282
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
283
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
284
+ (dropout): Dropout(p=0.1, inplace=False)
285
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
286
+ )
287
+ (feed_forward): PositionwiseFeedForward(
288
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
289
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
290
+ (dropout): Dropout(p=0.1, inplace=False)
291
+ (activation): Swish()
292
+ )
293
+ (feed_forward_macaron): PositionwiseFeedForward(
294
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
295
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
296
+ (dropout): Dropout(p=0.1, inplace=False)
297
+ (activation): Swish()
298
+ )
299
+ (conv_module): ConvolutionModule(
300
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
301
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
302
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
303
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
304
+ (activation): Swish()
305
+ )
306
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
307
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
308
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
309
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
310
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
311
+ (dropout): Dropout(p=0.1, inplace=False)
312
+ )
313
+ (8): EncoderLayer(
314
+ (self_attn): RelPositionMultiHeadedAttention(
315
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
316
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
317
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
318
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
319
+ (dropout): Dropout(p=0.1, inplace=False)
320
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
321
+ )
322
+ (feed_forward): PositionwiseFeedForward(
323
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
324
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
325
+ (dropout): Dropout(p=0.1, inplace=False)
326
+ (activation): Swish()
327
+ )
328
+ (feed_forward_macaron): PositionwiseFeedForward(
329
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
330
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
331
+ (dropout): Dropout(p=0.1, inplace=False)
332
+ (activation): Swish()
333
+ )
334
+ (conv_module): ConvolutionModule(
335
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
336
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
337
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
338
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
339
+ (activation): Swish()
340
+ )
341
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
342
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
343
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
344
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
345
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
346
+ (dropout): Dropout(p=0.1, inplace=False)
347
+ )
348
+ (9): EncoderLayer(
349
+ (self_attn): RelPositionMultiHeadedAttention(
350
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
351
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
352
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
353
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
354
+ (dropout): Dropout(p=0.1, inplace=False)
355
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
356
+ )
357
+ (feed_forward): PositionwiseFeedForward(
358
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
359
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
360
+ (dropout): Dropout(p=0.1, inplace=False)
361
+ (activation): Swish()
362
+ )
363
+ (feed_forward_macaron): PositionwiseFeedForward(
364
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
365
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
366
+ (dropout): Dropout(p=0.1, inplace=False)
367
+ (activation): Swish()
368
+ )
369
+ (conv_module): ConvolutionModule(
370
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
371
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
372
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
373
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
374
+ (activation): Swish()
375
+ )
376
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
377
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
378
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
379
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
380
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
381
+ (dropout): Dropout(p=0.1, inplace=False)
382
+ )
383
+ (10): EncoderLayer(
384
+ (self_attn): RelPositionMultiHeadedAttention(
385
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
386
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
387
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
388
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
389
+ (dropout): Dropout(p=0.1, inplace=False)
390
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
391
+ )
392
+ (feed_forward): PositionwiseFeedForward(
393
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
394
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
395
+ (dropout): Dropout(p=0.1, inplace=False)
396
+ (activation): Swish()
397
+ )
398
+ (feed_forward_macaron): PositionwiseFeedForward(
399
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
400
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
401
+ (dropout): Dropout(p=0.1, inplace=False)
402
+ (activation): Swish()
403
+ )
404
+ (conv_module): ConvolutionModule(
405
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
406
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
407
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
408
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
409
+ (activation): Swish()
410
+ )
411
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
412
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
413
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
414
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
415
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
416
+ (dropout): Dropout(p=0.1, inplace=False)
417
+ )
418
+ (11): EncoderLayer(
419
+ (self_attn): RelPositionMultiHeadedAttention(
420
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
421
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
422
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
423
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
424
+ (dropout): Dropout(p=0.1, inplace=False)
425
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
426
+ )
427
+ (feed_forward): PositionwiseFeedForward(
428
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
429
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
430
+ (dropout): Dropout(p=0.1, inplace=False)
431
+ (activation): Swish()
432
+ )
433
+ (feed_forward_macaron): PositionwiseFeedForward(
434
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
435
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
436
+ (dropout): Dropout(p=0.1, inplace=False)
437
+ (activation): Swish()
438
+ )
439
+ (conv_module): ConvolutionModule(
440
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
441
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
442
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
443
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
444
+ (activation): Swish()
445
+ )
446
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
447
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
448
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
449
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
450
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
451
+ (dropout): Dropout(p=0.1, inplace=False)
452
+ )
453
+ )
454
+ (after_norm): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
455
+ )
456
+ (decoder): TransformerDecoder(
457
+ (embed): Sequential(
458
+ (0): Embedding(32, 256)
459
+ (1): PositionalEncoding(
460
+ (dropout): Dropout(p=0.1, inplace=False)
461
+ )
462
+ )
463
+ (after_norm): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
464
+ (output_layer): Linear(in_features=256, out_features=32, bias=True)
465
+ (decoders): MultiSequential(
466
+ (0): DecoderLayer(
467
+ (self_attn): MultiHeadedAttention(
468
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
469
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
470
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
471
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
472
+ (dropout): Dropout(p=0.1, inplace=False)
473
+ )
474
+ (src_attn): MultiHeadedAttention(
475
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
476
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
477
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
478
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
479
+ (dropout): Dropout(p=0.1, inplace=False)
480
+ )
481
+ (feed_forward): PositionwiseFeedForward(
482
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
483
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
484
+ (dropout): Dropout(p=0.1, inplace=False)
485
+ (activation): ReLU()
486
+ )
487
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
488
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
489
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
490
+ (dropout): Dropout(p=0.1, inplace=False)
491
+ )
492
+ (1): DecoderLayer(
493
+ (self_attn): MultiHeadedAttention(
494
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
495
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
496
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
497
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
498
+ (dropout): Dropout(p=0.1, inplace=False)
499
+ )
500
+ (src_attn): MultiHeadedAttention(
501
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
502
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
503
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
504
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
505
+ (dropout): Dropout(p=0.1, inplace=False)
506
+ )
507
+ (feed_forward): PositionwiseFeedForward(
508
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
509
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
510
+ (dropout): Dropout(p=0.1, inplace=False)
511
+ (activation): ReLU()
512
+ )
513
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
514
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
515
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
516
+ (dropout): Dropout(p=0.1, inplace=False)
517
+ )
518
+ (2): DecoderLayer(
519
+ (self_attn): MultiHeadedAttention(
520
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
521
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
522
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
523
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
524
+ (dropout): Dropout(p=0.1, inplace=False)
525
+ )
526
+ (src_attn): MultiHeadedAttention(
527
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
528
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
529
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
530
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
531
+ (dropout): Dropout(p=0.1, inplace=False)
532
+ )
533
+ (feed_forward): PositionwiseFeedForward(
534
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
535
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
536
+ (dropout): Dropout(p=0.1, inplace=False)
537
+ (activation): ReLU()
538
+ )
539
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
540
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
541
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
542
+ (dropout): Dropout(p=0.1, inplace=False)
543
+ )
544
+ (3): DecoderLayer(
545
+ (self_attn): MultiHeadedAttention(
546
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
547
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
548
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
549
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
550
+ (dropout): Dropout(p=0.1, inplace=False)
551
+ )
552
+ (src_attn): MultiHeadedAttention(
553
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
554
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
555
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
556
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
557
+ (dropout): Dropout(p=0.1, inplace=False)
558
+ )
559
+ (feed_forward): PositionwiseFeedForward(
560
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
561
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
562
+ (dropout): Dropout(p=0.1, inplace=False)
563
+ (activation): ReLU()
564
+ )
565
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
566
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
567
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
568
+ (dropout): Dropout(p=0.1, inplace=False)
569
+ )
570
+ (4): DecoderLayer(
571
+ (self_attn): MultiHeadedAttention(
572
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
573
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
574
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
575
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
576
+ (dropout): Dropout(p=0.1, inplace=False)
577
+ )
578
+ (src_attn): MultiHeadedAttention(
579
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
580
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
581
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
582
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
583
+ (dropout): Dropout(p=0.1, inplace=False)
584
+ )
585
+ (feed_forward): PositionwiseFeedForward(
586
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
587
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
588
+ (dropout): Dropout(p=0.1, inplace=False)
589
+ (activation): ReLU()
590
+ )
591
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
592
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
593
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
594
+ (dropout): Dropout(p=0.1, inplace=False)
595
+ )
596
+ (5): DecoderLayer(
597
+ (self_attn): MultiHeadedAttention(
598
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
599
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
600
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
601
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
602
+ (dropout): Dropout(p=0.1, inplace=False)
603
+ )
604
+ (src_attn): MultiHeadedAttention(
605
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
606
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
607
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
608
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
609
+ (dropout): Dropout(p=0.1, inplace=False)
610
+ )
611
+ (feed_forward): PositionwiseFeedForward(
612
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
613
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
614
+ (dropout): Dropout(p=0.1, inplace=False)
615
+ (activation): ReLU()
616
+ )
617
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
618
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
619
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
620
+ (dropout): Dropout(p=0.1, inplace=False)
621
+ )
622
+ )
623
+ )
624
+ (criterion_att): LabelSmoothingLoss(
625
+ (criterion): KLDivLoss()
626
+ )
627
+ )
628
+
629
+ Model summary:
630
+ Class Name: ESPnetASRModel
631
+ Total Number of model parameters: 43.00 M
632
+ Number of trainable parameters: 43.00 M (100.0%)
633
+ Size: 172.01 MB
634
+ Type: torch.float32
635
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:31,624 (abs_task:1233) INFO: Optimizer:
636
+ Adam (
637
+ Parameter Group 0
638
+ amsgrad: False
639
+ betas: (0.9, 0.999)
640
+ eps: 1e-08
641
+ initial_lr: 0.002
642
+ lr: 1e-07
643
+ maximize: False
644
+ weight_decay: 1e-06
645
+ )
646
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:31,624 (abs_task:1234) INFO: Scheduler: WarmupLR(warmup_steps=20000)
647
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:31,626 (abs_task:1243) INFO: Saving the configuration in exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/config.yaml
648
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:31,692 (abs_task:1304) INFO: Loading pretrained params from /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth
649
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:35,293 (asr:461) INFO: Optional Data Names: ('text_spk2', 'text_spk3', 'text_spk4')
650
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,615 (abs_task:1614) INFO: [train] dataset:
651
+ ESPnetDataset(
652
+ speech: {"path": "dump/raw/train_medium_kaldi_fmt_sp/wav.scp", "type": "kaldi_ark"}
653
+ text: {"path": "dump/raw/train_medium_kaldi_fmt_sp/text", "type": "text"}
654
+ preprocess: <espnet2.train.preprocessor.CommonPreprocessor_multi object at 0x7f4603b63730>)
655
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,616 (abs_task:1615) INFO: [train] Batch sampler: NumElementsBatchSampler(N-batch=10615, batch_bins=16000000, sort_in_batch=descending, sort_batch=descending)
656
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,620 (abs_task:1616) INFO: [train] mini-batch sizes summary: N-batch=10615, mean=53.4, min=7, max=201
657
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,737 (asr:461) INFO: Optional Data Names: ('text_spk2', 'text_spk3', 'text_spk4')
658
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,747 (abs_task:1614) INFO: [valid] dataset:
659
+ ESPnetDataset(
660
+ speech: {"path": "dump/raw/dev_kaldi_fmt/wav.scp", "type": "kaldi_ark"}
661
+ text: {"path": "dump/raw/dev_kaldi_fmt/text", "type": "text"}
662
+ preprocess: <espnet2.train.preprocessor.CommonPreprocessor_multi object at 0x7f45f1f856d0>)
663
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,747 (abs_task:1615) INFO: [valid] Batch sampler: NumElementsBatchSampler(N-batch=12, batch_bins=16000000, sort_in_batch=descending, sort_batch=descending)
664
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,747 (abs_task:1616) INFO: [valid] mini-batch sizes summary: N-batch=12, mean=50.4, min=17, max=82
665
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,752 (asr:461) INFO: Optional Data Names: ('text_spk2', 'text_spk3', 'text_spk4')
666
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,775 (abs_task:1614) INFO: [plot_att] dataset:
667
+ ESPnetDataset(
668
+ speech: {"path": "dump/raw/dev_kaldi_fmt/wav.scp", "type": "kaldi_ark"}
669
+ text: {"path": "dump/raw/dev_kaldi_fmt/text", "type": "text"}
670
+ preprocess: <espnet2.train.preprocessor.CommonPreprocessor_multi object at 0x7f45f1f85d60>)
671
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,775 (abs_task:1615) INFO: [plot_att] Batch sampler: UnsortedBatchSampler(N-batch=605, batch_size=1, key_file=exp/asr_stats_raw_en_char_sp/valid/speech_shape,
672
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:44,775 (abs_task:1616) INFO: [plot_att] mini-batch sizes summary: N-batch=3, mean=1.0, min=1, max=1
673
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:48,183 (trainer:159) INFO: The training was resumed using exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/checkpoint.pth
674
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823743 [0] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
675
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823743 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
676
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823743 [0] NCCL INFO NET/IB : No device found.
677
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823743 [0] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
678
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823743 [0] NCCL INFO Using network Socket
679
+ NCCL version 2.10.3+cuda10.2
680
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823744 [1] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
681
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823745 [2] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
682
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823746 [3] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
683
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823744 [1] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
684
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823745 [2] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
685
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823746 [3] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
686
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823744 [1] NCCL INFO NET/IB : No device found.
687
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823744 [1] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
688
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823744 [1] NCCL INFO Using network Socket
689
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823746 [3] NCCL INFO NET/IB : No device found.
690
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823746 [3] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
691
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823746 [3] NCCL INFO Using network Socket
692
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823745 [2] NCCL INFO NET/IB : No device found.
693
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823745 [2] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
694
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823745 [2] NCCL INFO Using network Socket
695
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2
696
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Setting affinity for GPU 3 to 20000002
697
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1
698
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Setting affinity for GPU 2 to 20000002
699
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Channel 00/02 : 0 1 2 3
700
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0
701
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Channel 01/02 : 0 1 2 3
702
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Setting affinity for GPU 1 to 20000002
703
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1
704
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Setting affinity for GPU 0 to 20000002
705
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Channel 00 : 3[8000] -> 0[4000] via P2P/IPC
706
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Channel 00 : 2[7000] -> 3[8000] via P2P/IPC
707
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Channel 01 : 2[7000] -> 3[8000] via P2P/IPC
708
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Channel 00 : 1[6000] -> 2[7000] via P2P/IPC
709
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Channel 01 : 1[6000] -> 2[7000] via P2P/IPC
710
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Channel 00 : 0[4000] -> 1[6000] via P2P/IPC
711
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Channel 01 : 0[4000] -> 1[6000] via P2P/IPC
712
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Channel 01 : 3[8000] -> 0[4000] via P2P/IPC
713
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Connected all rings
714
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Connected all rings
715
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Channel 00 : 3[8000] -> 2[7000] via P2P/IPC
716
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Channel 01 : 3[8000] -> 2[7000] via P2P/IPC
717
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Connected all rings
718
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Channel 00 : 2[7000] -> 1[6000] via P2P/IPC
719
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Connected all rings
720
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Channel 00 : 1[6000] -> 0[4000] via P2P/IPC
721
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Channel 01 : 2[7000] -> 1[6000] via P2P/IPC
722
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO Connected all trees
723
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
724
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
725
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Channel 01 : 1[6000] -> 0[4000] via P2P/IPC
726
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO Connected all trees
727
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
728
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
729
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO Connected all trees
730
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
731
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
732
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO Connected all trees
733
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
734
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
735
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823746:1823875 [3] NCCL INFO comm 0x7f4294001200 rank 3 nranks 4 cudaDev 3 busId 8000 - Init COMPLETE
736
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823870 [0] NCCL INFO comm 0x7f450c001200 rank 0 nranks 4 cudaDev 0 busId 4000 - Init COMPLETE
737
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823745:1823876 [2] NCCL INFO comm 0x7f631c001200 rank 2 nranks 4 cudaDev 2 busId 7000 - Init COMPLETE
738
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823744:1823874 [1] NCCL INFO comm 0x7fa7ac001200 rank 1 nranks 4 cudaDev 1 busId 6000 - Init COMPLETE
739
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1823743:1823743 [0] NCCL INFO Launch mode Parallel
740
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:44:52,106 (trainer:284) INFO: 9/60epoch started
741
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:49:07,241 (distributed:948) INFO: Reducer buckets have been rebuilt in this iteration.
742
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 00:55:25,795 (trainer:732) INFO: 9epoch:train:1-530batch: iter_time=0.002, forward_time=0.209, loss_att=107.878, acc=0.904, loss=107.878, backward_time=0.325, grad_norm=103.272, clip=100.000, loss_scale=1.000, optim_step_time=0.085, optim0_lr0=0.002, train_time=4.789
743
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:01:45,507 (trainer:732) INFO: 9epoch:train:531-1060batch: iter_time=2.446e-04, forward_time=0.209, loss_att=112.672, acc=0.903, loss=112.672, backward_time=0.327, grad_norm=105.518, clip=100.000, loss_scale=1.000, optim_step_time=0.084, optim0_lr0=0.002, train_time=2.865
744
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:08:04,727 (trainer:732) INFO: 9epoch:train:1061-1590batch: iter_time=2.303e-04, forward_time=0.208, loss_att=109.250, acc=0.905, loss=109.250, backward_time=0.327, grad_norm=112.951, clip=100.000, loss_scale=1.000, optim_step_time=0.084, optim0_lr0=0.002, train_time=2.863
745
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:14:24,051 (trainer:732) INFO: 9epoch:train:1591-2120batch: iter_time=3.562e-04, forward_time=0.211, loss_att=109.274, acc=0.906, loss=109.274, backward_time=0.327, grad_norm=104.180, clip=100.000, loss_scale=1.000, optim_step_time=0.086, optim0_lr0=0.002, train_time=2.861
746
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:20:44,930 (trainer:732) INFO: 9epoch:train:2121-2650batch: iter_time=3.394e-04, forward_time=0.211, loss_att=110.221, acc=0.905, loss=110.221, backward_time=0.328, grad_norm=110.200, clip=100.000, loss_scale=1.000, optim_step_time=0.086, optim0_lr0=0.002, train_time=2.875
747
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:27:04,841 (trainer:732) INFO: 9epoch:train:2651-3180batch: iter_time=3.402e-04, forward_time=0.211, loss_att=111.572, acc=0.905, loss=111.572, backward_time=0.328, grad_norm=102.797, clip=100.000, loss_scale=1.000, optim_step_time=0.085, optim0_lr0=0.002, train_time=2.866
748
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:33:23,833 (trainer:732) INFO: 9epoch:train:3181-3710batch: iter_time=3.449e-04, forward_time=0.210, loss_att=108.396, acc=0.906, loss=108.396, backward_time=0.327, grad_norm=103.265, clip=100.000, loss_scale=1.000, optim_step_time=0.085, optim0_lr0=0.002, train_time=2.860
749
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:39:42,244 (trainer:732) INFO: 9epoch:train:3711-4240batch: iter_time=3.845e-04, forward_time=0.210, loss_att=105.457, acc=0.907, loss=105.457, backward_time=0.326, grad_norm=105.516, clip=100.000, loss_scale=1.000, optim_step_time=0.086, optim0_lr0=0.002, train_time=2.855
750
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:46:01,633 (trainer:732) INFO: 9epoch:train:4241-4770batch: iter_time=3.412e-04, forward_time=0.211, loss_att=110.350, acc=0.906, loss=110.350, backward_time=0.327, grad_norm=104.195, clip=100.000, loss_scale=1.000, optim_step_time=0.083, optim0_lr0=0.002, train_time=2.863
751
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:52:20,391 (trainer:732) INFO: 9epoch:train:4771-5300batch: iter_time=3.219e-04, forward_time=0.210, loss_att=106.497, acc=0.907, loss=106.497, backward_time=0.327, grad_norm=105.330, clip=100.000, loss_scale=1.000, optim_step_time=0.084, optim0_lr0=0.002, train_time=2.858
752
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 01:58:41,240 (trainer:732) INFO: 9epoch:train:5301-5830batch: iter_time=3.283e-04, forward_time=0.211, loss_att=108.151, acc=0.909, loss=108.151, backward_time=0.328, grad_norm=100.440, clip=100.000, loss_scale=1.000, optim_step_time=0.084, optim0_lr0=0.002, train_time=2.875
753
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 02:05:01,053 (trainer:732) INFO: 9epoch:train:5831-6360batch: iter_time=3.129e-04, forward_time=0.210, loss_att=107.337, acc=0.908, loss=107.337, backward_time=0.328, grad_norm=105.336, clip=100.000, loss_scale=1.000, optim_step_time=0.083, optim0_lr0=0.002, train_time=2.865
754
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 02:11:21,060 (trainer:732) INFO: 9epoch:train:6361-6890batch: iter_time=3.236e-04, forward_time=0.210, loss_att=106.133, acc=0.909, loss=106.133, backward_time=0.328, grad_norm=104.171, clip=100.000, loss_scale=1.000, optim_step_time=0.083, optim0_lr0=0.002, train_time=2.869
755
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-20 02:17:40,657 (trainer:732) INFO: 9epoch:train:6891-7420batch: iter_time=3.511e-04, forward_time=0.210, loss_att=107.243, acc=0.908, loss=107.243, backward_time=0.327, grad_norm=108.454, clip=100.000, loss_scale=1.000, optim_step_time=0.085, optim0_lr0=0.002, train_time=2.863
756
+ Traceback (most recent call last):
757
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 197, in _run_module_as_main
758
+ return _run_code(code, main_globals, None,
759
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 87, in _run_code
760
+ exec(code, run_globals)
761
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 23, in <module>
762
+ main()
763
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 19, in main
764
+ ASRTask.main(cmd=cmd)
765
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 1132, in main
766
+ while not ProcessContext(processes, error_queues).join():
767
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 140, in join
768
+ raise ProcessExitedException(
769
+ torch.multiprocessing.spawn.ProcessExitedException: process 0 terminated with signal SIGKILL
770
+ # Accounting: time=5952 threads=1
771
+ # Ended (code 1) at Mon Nov 20 02:22:24 CST 2023, elapsed time 5952 seconds
772
+ /star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/resource_tracker.py:216: UserWarning: resource_tracker: There appear to be 224 leaked semaphore objects to clean up at shutdown
773
+ warnings.warn('resource_tracker: There appear to be %d '
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.2.log ADDED
The diff for this file is too large to render. See raw diff
 
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.3.log ADDED
@@ -0,0 +1,1057 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # python3 -m espnet2.bin.asr_train --use_preprocessor true --bpemodel none --token_type char --token_list data/en_token_list/char/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/wav.scp,speech,kaldi_ark --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/speech_shape --resume true --init_param /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth --ignore_init_mismatch false --fold_length 80000 --output_dir exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new --config conf/tuning/train_sot_asr_conformer_medium.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/asr_stats_raw_en_char_sp/train/feats_stats.npz --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/wav.scp,speech,kaldi_ark --train_shape_file exp/asr_stats_raw_en_char_sp/train/speech_shape --fold_length 150 --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/text,text,text --train_shape_file exp/asr_stats_raw_en_char_sp/train/text_shape.char --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/text,text,text --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/text_shape.char --ngpu 4 --multiprocessing_distributed True
2
+ # Started at Sat Nov 18 19:32:54 CST 2023
3
+ #
4
+ /star-home/jinzengrui/lib/miniconda3/envs/dev39/bin/python3 /star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py --use_preprocessor true --bpemodel none --token_type char --token_list data/en_token_list/char/tokens.txt --non_linguistic_symbols none --cleaner none --g2p none --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/wav.scp,speech,kaldi_ark --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/speech_shape --resume true --init_param /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth --ignore_init_mismatch false --fold_length 80000 --output_dir exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new --config conf/tuning/train_sot_asr_conformer_medium.yaml --frontend_conf fs=16k --normalize=global_mvn --normalize_conf stats_file=exp/asr_stats_raw_en_char_sp/train/feats_stats.npz --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/wav.scp,speech,kaldi_ark --train_shape_file exp/asr_stats_raw_en_char_sp/train/speech_shape --fold_length 150 --train_data_path_and_name_and_type dump/raw/train_medium_kaldi_fmt_sp/text,text,text --train_shape_file exp/asr_stats_raw_en_char_sp/train/text_shape.char --valid_data_path_and_name_and_type dump/raw/dev_kaldi_fmt/text,text,text --valid_shape_file exp/asr_stats_raw_en_char_sp/valid/text_shape.char --ngpu 4 --multiprocessing_distributed True
5
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:14,486 (distributed_c10d:228) INFO: Added key: store_based_barrier_key:1 to store for rank: 0
6
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:14,486 (distributed_c10d:262) INFO: Rank 0: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes.
7
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:14,530 (asr:490) INFO: Vocabulary size: 32
8
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:22,927 (abs_task:1229) INFO: pytorch.version=1.11.0+cu102, cuda.available=True, cudnn.version=7605, cudnn.benchmark=False, cudnn.deterministic=True
9
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:22,941 (abs_task:1230) INFO: Model structure:
10
+ ESPnetASRModel(
11
+ (frontend): DefaultFrontend(
12
+ (stft): Stft(n_fft=512, win_length=512, hop_length=128, center=True, normalized=False, onesided=True)
13
+ (frontend): Frontend()
14
+ (logmel): LogMel(sr=16000, n_fft=512, n_mels=80, fmin=0, fmax=8000.0, htk=False)
15
+ )
16
+ (normalize): GlobalMVN(stats_file=exp/asr_stats_raw_en_char_sp/train/feats_stats.npz, norm_means=True, norm_vars=True)
17
+ (encoder): ConformerEncoder(
18
+ (embed): Conv2dSubsampling(
19
+ (conv): Sequential(
20
+ (0): Conv2d(1, 256, kernel_size=(3, 3), stride=(2, 2))
21
+ (1): ReLU()
22
+ (2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2))
23
+ (3): ReLU()
24
+ )
25
+ (out): Sequential(
26
+ (0): Linear(in_features=4864, out_features=256, bias=True)
27
+ (1): RelPositionalEncoding(
28
+ (dropout): Dropout(p=0.1, inplace=False)
29
+ )
30
+ )
31
+ )
32
+ (encoders): MultiSequential(
33
+ (0): EncoderLayer(
34
+ (self_attn): RelPositionMultiHeadedAttention(
35
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
36
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
37
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
38
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
39
+ (dropout): Dropout(p=0.1, inplace=False)
40
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
41
+ )
42
+ (feed_forward): PositionwiseFeedForward(
43
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
44
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
45
+ (dropout): Dropout(p=0.1, inplace=False)
46
+ (activation): Swish()
47
+ )
48
+ (feed_forward_macaron): PositionwiseFeedForward(
49
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
50
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
51
+ (dropout): Dropout(p=0.1, inplace=False)
52
+ (activation): Swish()
53
+ )
54
+ (conv_module): ConvolutionModule(
55
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
56
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
57
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
58
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
59
+ (activation): Swish()
60
+ )
61
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
62
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
63
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
64
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
65
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
66
+ (dropout): Dropout(p=0.1, inplace=False)
67
+ )
68
+ (1): EncoderLayer(
69
+ (self_attn): RelPositionMultiHeadedAttention(
70
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
71
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
72
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
73
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
74
+ (dropout): Dropout(p=0.1, inplace=False)
75
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
76
+ )
77
+ (feed_forward): PositionwiseFeedForward(
78
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
79
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
80
+ (dropout): Dropout(p=0.1, inplace=False)
81
+ (activation): Swish()
82
+ )
83
+ (feed_forward_macaron): PositionwiseFeedForward(
84
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
85
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
86
+ (dropout): Dropout(p=0.1, inplace=False)
87
+ (activation): Swish()
88
+ )
89
+ (conv_module): ConvolutionModule(
90
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
91
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
92
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
93
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
94
+ (activation): Swish()
95
+ )
96
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
97
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
98
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
99
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
100
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
101
+ (dropout): Dropout(p=0.1, inplace=False)
102
+ )
103
+ (2): EncoderLayer(
104
+ (self_attn): RelPositionMultiHeadedAttention(
105
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
106
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
107
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
108
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
109
+ (dropout): Dropout(p=0.1, inplace=False)
110
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
111
+ )
112
+ (feed_forward): PositionwiseFeedForward(
113
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
114
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
115
+ (dropout): Dropout(p=0.1, inplace=False)
116
+ (activation): Swish()
117
+ )
118
+ (feed_forward_macaron): PositionwiseFeedForward(
119
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
120
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
121
+ (dropout): Dropout(p=0.1, inplace=False)
122
+ (activation): Swish()
123
+ )
124
+ (conv_module): ConvolutionModule(
125
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
126
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
127
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
128
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
129
+ (activation): Swish()
130
+ )
131
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
132
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
133
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
134
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
135
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
136
+ (dropout): Dropout(p=0.1, inplace=False)
137
+ )
138
+ (3): EncoderLayer(
139
+ (self_attn): RelPositionMultiHeadedAttention(
140
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
141
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
142
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
143
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
144
+ (dropout): Dropout(p=0.1, inplace=False)
145
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
146
+ )
147
+ (feed_forward): PositionwiseFeedForward(
148
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
149
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
150
+ (dropout): Dropout(p=0.1, inplace=False)
151
+ (activation): Swish()
152
+ )
153
+ (feed_forward_macaron): PositionwiseFeedForward(
154
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
155
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
156
+ (dropout): Dropout(p=0.1, inplace=False)
157
+ (activation): Swish()
158
+ )
159
+ (conv_module): ConvolutionModule(
160
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
161
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
162
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
163
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
164
+ (activation): Swish()
165
+ )
166
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
167
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
168
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
169
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
170
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
171
+ (dropout): Dropout(p=0.1, inplace=False)
172
+ )
173
+ (4): EncoderLayer(
174
+ (self_attn): RelPositionMultiHeadedAttention(
175
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
176
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
177
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
178
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
179
+ (dropout): Dropout(p=0.1, inplace=False)
180
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
181
+ )
182
+ (feed_forward): PositionwiseFeedForward(
183
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
184
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
185
+ (dropout): Dropout(p=0.1, inplace=False)
186
+ (activation): Swish()
187
+ )
188
+ (feed_forward_macaron): PositionwiseFeedForward(
189
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
190
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
191
+ (dropout): Dropout(p=0.1, inplace=False)
192
+ (activation): Swish()
193
+ )
194
+ (conv_module): ConvolutionModule(
195
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
196
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
197
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
198
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
199
+ (activation): Swish()
200
+ )
201
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
202
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
203
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
204
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
205
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
206
+ (dropout): Dropout(p=0.1, inplace=False)
207
+ )
208
+ (5): EncoderLayer(
209
+ (self_attn): RelPositionMultiHeadedAttention(
210
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
211
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
212
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
213
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
214
+ (dropout): Dropout(p=0.1, inplace=False)
215
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
216
+ )
217
+ (feed_forward): PositionwiseFeedForward(
218
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
219
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
220
+ (dropout): Dropout(p=0.1, inplace=False)
221
+ (activation): Swish()
222
+ )
223
+ (feed_forward_macaron): PositionwiseFeedForward(
224
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
225
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
226
+ (dropout): Dropout(p=0.1, inplace=False)
227
+ (activation): Swish()
228
+ )
229
+ (conv_module): ConvolutionModule(
230
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
231
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
232
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
233
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
234
+ (activation): Swish()
235
+ )
236
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
237
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
238
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
239
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
240
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
241
+ (dropout): Dropout(p=0.1, inplace=False)
242
+ )
243
+ (6): EncoderLayer(
244
+ (self_attn): RelPositionMultiHeadedAttention(
245
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
246
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
247
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
248
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
249
+ (dropout): Dropout(p=0.1, inplace=False)
250
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
251
+ )
252
+ (feed_forward): PositionwiseFeedForward(
253
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
254
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
255
+ (dropout): Dropout(p=0.1, inplace=False)
256
+ (activation): Swish()
257
+ )
258
+ (feed_forward_macaron): PositionwiseFeedForward(
259
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
260
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
261
+ (dropout): Dropout(p=0.1, inplace=False)
262
+ (activation): Swish()
263
+ )
264
+ (conv_module): ConvolutionModule(
265
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
266
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
267
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
268
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
269
+ (activation): Swish()
270
+ )
271
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
272
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
273
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
274
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
275
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
276
+ (dropout): Dropout(p=0.1, inplace=False)
277
+ )
278
+ (7): EncoderLayer(
279
+ (self_attn): RelPositionMultiHeadedAttention(
280
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
281
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
282
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
283
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
284
+ (dropout): Dropout(p=0.1, inplace=False)
285
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
286
+ )
287
+ (feed_forward): PositionwiseFeedForward(
288
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
289
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
290
+ (dropout): Dropout(p=0.1, inplace=False)
291
+ (activation): Swish()
292
+ )
293
+ (feed_forward_macaron): PositionwiseFeedForward(
294
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
295
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
296
+ (dropout): Dropout(p=0.1, inplace=False)
297
+ (activation): Swish()
298
+ )
299
+ (conv_module): ConvolutionModule(
300
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
301
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
302
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
303
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
304
+ (activation): Swish()
305
+ )
306
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
307
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
308
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
309
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
310
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
311
+ (dropout): Dropout(p=0.1, inplace=False)
312
+ )
313
+ (8): EncoderLayer(
314
+ (self_attn): RelPositionMultiHeadedAttention(
315
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
316
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
317
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
318
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
319
+ (dropout): Dropout(p=0.1, inplace=False)
320
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
321
+ )
322
+ (feed_forward): PositionwiseFeedForward(
323
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
324
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
325
+ (dropout): Dropout(p=0.1, inplace=False)
326
+ (activation): Swish()
327
+ )
328
+ (feed_forward_macaron): PositionwiseFeedForward(
329
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
330
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
331
+ (dropout): Dropout(p=0.1, inplace=False)
332
+ (activation): Swish()
333
+ )
334
+ (conv_module): ConvolutionModule(
335
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
336
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
337
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
338
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
339
+ (activation): Swish()
340
+ )
341
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
342
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
343
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
344
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
345
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
346
+ (dropout): Dropout(p=0.1, inplace=False)
347
+ )
348
+ (9): EncoderLayer(
349
+ (self_attn): RelPositionMultiHeadedAttention(
350
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
351
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
352
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
353
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
354
+ (dropout): Dropout(p=0.1, inplace=False)
355
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
356
+ )
357
+ (feed_forward): PositionwiseFeedForward(
358
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
359
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
360
+ (dropout): Dropout(p=0.1, inplace=False)
361
+ (activation): Swish()
362
+ )
363
+ (feed_forward_macaron): PositionwiseFeedForward(
364
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
365
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
366
+ (dropout): Dropout(p=0.1, inplace=False)
367
+ (activation): Swish()
368
+ )
369
+ (conv_module): ConvolutionModule(
370
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
371
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
372
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
373
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
374
+ (activation): Swish()
375
+ )
376
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
377
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
378
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
379
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
380
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
381
+ (dropout): Dropout(p=0.1, inplace=False)
382
+ )
383
+ (10): EncoderLayer(
384
+ (self_attn): RelPositionMultiHeadedAttention(
385
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
386
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
387
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
388
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
389
+ (dropout): Dropout(p=0.1, inplace=False)
390
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
391
+ )
392
+ (feed_forward): PositionwiseFeedForward(
393
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
394
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
395
+ (dropout): Dropout(p=0.1, inplace=False)
396
+ (activation): Swish()
397
+ )
398
+ (feed_forward_macaron): PositionwiseFeedForward(
399
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
400
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
401
+ (dropout): Dropout(p=0.1, inplace=False)
402
+ (activation): Swish()
403
+ )
404
+ (conv_module): ConvolutionModule(
405
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
406
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
407
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
408
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
409
+ (activation): Swish()
410
+ )
411
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
412
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
413
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
414
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
415
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
416
+ (dropout): Dropout(p=0.1, inplace=False)
417
+ )
418
+ (11): EncoderLayer(
419
+ (self_attn): RelPositionMultiHeadedAttention(
420
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
421
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
422
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
423
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
424
+ (dropout): Dropout(p=0.1, inplace=False)
425
+ (linear_pos): Linear(in_features=256, out_features=256, bias=False)
426
+ )
427
+ (feed_forward): PositionwiseFeedForward(
428
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
429
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
430
+ (dropout): Dropout(p=0.1, inplace=False)
431
+ (activation): Swish()
432
+ )
433
+ (feed_forward_macaron): PositionwiseFeedForward(
434
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
435
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
436
+ (dropout): Dropout(p=0.1, inplace=False)
437
+ (activation): Swish()
438
+ )
439
+ (conv_module): ConvolutionModule(
440
+ (pointwise_conv1): Conv1d(256, 512, kernel_size=(1,), stride=(1,))
441
+ (depthwise_conv): Conv1d(256, 256, kernel_size=(31,), stride=(1,), padding=(15,), groups=256)
442
+ (norm): BatchNorm1d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
443
+ (pointwise_conv2): Conv1d(256, 256, kernel_size=(1,), stride=(1,))
444
+ (activation): Swish()
445
+ )
446
+ (norm_ff): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
447
+ (norm_mha): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
448
+ (norm_ff_macaron): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
449
+ (norm_conv): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
450
+ (norm_final): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
451
+ (dropout): Dropout(p=0.1, inplace=False)
452
+ )
453
+ )
454
+ (after_norm): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
455
+ )
456
+ (decoder): TransformerDecoder(
457
+ (embed): Sequential(
458
+ (0): Embedding(32, 256)
459
+ (1): PositionalEncoding(
460
+ (dropout): Dropout(p=0.1, inplace=False)
461
+ )
462
+ )
463
+ (after_norm): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
464
+ (output_layer): Linear(in_features=256, out_features=32, bias=True)
465
+ (decoders): MultiSequential(
466
+ (0): DecoderLayer(
467
+ (self_attn): MultiHeadedAttention(
468
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
469
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
470
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
471
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
472
+ (dropout): Dropout(p=0.1, inplace=False)
473
+ )
474
+ (src_attn): MultiHeadedAttention(
475
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
476
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
477
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
478
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
479
+ (dropout): Dropout(p=0.1, inplace=False)
480
+ )
481
+ (feed_forward): PositionwiseFeedForward(
482
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
483
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
484
+ (dropout): Dropout(p=0.1, inplace=False)
485
+ (activation): ReLU()
486
+ )
487
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
488
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
489
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
490
+ (dropout): Dropout(p=0.1, inplace=False)
491
+ )
492
+ (1): DecoderLayer(
493
+ (self_attn): MultiHeadedAttention(
494
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
495
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
496
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
497
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
498
+ (dropout): Dropout(p=0.1, inplace=False)
499
+ )
500
+ (src_attn): MultiHeadedAttention(
501
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
502
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
503
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
504
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
505
+ (dropout): Dropout(p=0.1, inplace=False)
506
+ )
507
+ (feed_forward): PositionwiseFeedForward(
508
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
509
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
510
+ (dropout): Dropout(p=0.1, inplace=False)
511
+ (activation): ReLU()
512
+ )
513
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
514
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
515
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
516
+ (dropout): Dropout(p=0.1, inplace=False)
517
+ )
518
+ (2): DecoderLayer(
519
+ (self_attn): MultiHeadedAttention(
520
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
521
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
522
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
523
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
524
+ (dropout): Dropout(p=0.1, inplace=False)
525
+ )
526
+ (src_attn): MultiHeadedAttention(
527
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
528
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
529
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
530
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
531
+ (dropout): Dropout(p=0.1, inplace=False)
532
+ )
533
+ (feed_forward): PositionwiseFeedForward(
534
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
535
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
536
+ (dropout): Dropout(p=0.1, inplace=False)
537
+ (activation): ReLU()
538
+ )
539
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
540
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
541
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
542
+ (dropout): Dropout(p=0.1, inplace=False)
543
+ )
544
+ (3): DecoderLayer(
545
+ (self_attn): MultiHeadedAttention(
546
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
547
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
548
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
549
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
550
+ (dropout): Dropout(p=0.1, inplace=False)
551
+ )
552
+ (src_attn): MultiHeadedAttention(
553
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
554
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
555
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
556
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
557
+ (dropout): Dropout(p=0.1, inplace=False)
558
+ )
559
+ (feed_forward): PositionwiseFeedForward(
560
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
561
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
562
+ (dropout): Dropout(p=0.1, inplace=False)
563
+ (activation): ReLU()
564
+ )
565
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
566
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
567
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
568
+ (dropout): Dropout(p=0.1, inplace=False)
569
+ )
570
+ (4): DecoderLayer(
571
+ (self_attn): MultiHeadedAttention(
572
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
573
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
574
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
575
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
576
+ (dropout): Dropout(p=0.1, inplace=False)
577
+ )
578
+ (src_attn): MultiHeadedAttention(
579
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
580
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
581
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
582
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
583
+ (dropout): Dropout(p=0.1, inplace=False)
584
+ )
585
+ (feed_forward): PositionwiseFeedForward(
586
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
587
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
588
+ (dropout): Dropout(p=0.1, inplace=False)
589
+ (activation): ReLU()
590
+ )
591
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
592
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
593
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
594
+ (dropout): Dropout(p=0.1, inplace=False)
595
+ )
596
+ (5): DecoderLayer(
597
+ (self_attn): MultiHeadedAttention(
598
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
599
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
600
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
601
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
602
+ (dropout): Dropout(p=0.1, inplace=False)
603
+ )
604
+ (src_attn): MultiHeadedAttention(
605
+ (linear_q): Linear(in_features=256, out_features=256, bias=True)
606
+ (linear_k): Linear(in_features=256, out_features=256, bias=True)
607
+ (linear_v): Linear(in_features=256, out_features=256, bias=True)
608
+ (linear_out): Linear(in_features=256, out_features=256, bias=True)
609
+ (dropout): Dropout(p=0.1, inplace=False)
610
+ )
611
+ (feed_forward): PositionwiseFeedForward(
612
+ (w_1): Linear(in_features=256, out_features=2048, bias=True)
613
+ (w_2): Linear(in_features=2048, out_features=256, bias=True)
614
+ (dropout): Dropout(p=0.1, inplace=False)
615
+ (activation): ReLU()
616
+ )
617
+ (norm1): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
618
+ (norm2): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
619
+ (norm3): LayerNorm((256,), eps=1e-12, elementwise_affine=True)
620
+ (dropout): Dropout(p=0.1, inplace=False)
621
+ )
622
+ )
623
+ )
624
+ (criterion_att): LabelSmoothingLoss(
625
+ (criterion): KLDivLoss()
626
+ )
627
+ )
628
+
629
+ Model summary:
630
+ Class Name: ESPnetASRModel
631
+ Total Number of model parameters: 43.00 M
632
+ Number of trainable parameters: 43.00 M (100.0%)
633
+ Size: 172.01 MB
634
+ Type: torch.float32
635
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:22,942 (abs_task:1233) INFO: Optimizer:
636
+ Adam (
637
+ Parameter Group 0
638
+ amsgrad: False
639
+ betas: (0.9, 0.999)
640
+ eps: 1e-08
641
+ initial_lr: 0.002
642
+ lr: 1e-07
643
+ maximize: False
644
+ weight_decay: 1e-06
645
+ )
646
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:22,942 (abs_task:1234) INFO: Scheduler: WarmupLR(warmup_steps=20000)
647
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:22,945 (abs_task:1243) INFO: Saving the configuration in exp/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/config.yaml
648
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:22,967 (abs_task:1304) INFO: Loading pretrained params from /star-home/jinzengrui/dev/espnet/egs2/librimix/sot_asr1_pretrain/exp/asr_train_sot_asr_conformer_raw_en_char_sp/45epoch.pth
649
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:25,928 (asr:461) INFO: Optional Data Names: ('text_spk2', 'text_spk3', 'text_spk4')
650
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,548 (abs_task:1614) INFO: [train] dataset:
651
+ ESPnetDataset(
652
+ speech: {"path": "dump/raw/train_medium_kaldi_fmt_sp/wav.scp", "type": "kaldi_ark"}
653
+ text: {"path": "dump/raw/train_medium_kaldi_fmt_sp/text", "type": "text"}
654
+ preprocess: <espnet2.train.preprocessor.CommonPreprocessor_multi object at 0x7f8e24d29490>)
655
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,548 (abs_task:1615) INFO: [train] Batch sampler: NumElementsBatchSampler(N-batch=9484, batch_bins=16000000, sort_in_batch=descending, sort_batch=descending)
656
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,552 (abs_task:1616) INFO: [train] mini-batch sizes summary: N-batch=9484, mean=41.8, min=10, max=170
657
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,657 (asr:461) INFO: Optional Data Names: ('text_spk2', 'text_spk3', 'text_spk4')
658
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,665 (abs_task:1614) INFO: [valid] dataset:
659
+ ESPnetDataset(
660
+ speech: {"path": "dump/raw/dev_kaldi_fmt/wav.scp", "type": "kaldi_ark"}
661
+ text: {"path": "dump/raw/dev_kaldi_fmt/text", "type": "text"}
662
+ preprocess: <espnet2.train.preprocessor.CommonPreprocessor_multi object at 0x7f8e12793700>)
663
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,665 (abs_task:1615) INFO: [valid] Batch sampler: NumElementsBatchSampler(N-batch=11, batch_bins=16000000, sort_in_batch=descending, sort_batch=descending)
664
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,665 (abs_task:1616) INFO: [valid] mini-batch sizes summary: N-batch=11, mean=41.1, min=22, max=66
665
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,671 (asr:461) INFO: Optional Data Names: ('text_spk2', 'text_spk3', 'text_spk4')
666
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,697 (abs_task:1614) INFO: [plot_att] dataset:
667
+ ESPnetDataset(
668
+ speech: {"path": "dump/raw/dev_kaldi_fmt/wav.scp", "type": "kaldi_ark"}
669
+ text: {"path": "dump/raw/dev_kaldi_fmt/text", "type": "text"}
670
+ preprocess: <espnet2.train.preprocessor.CommonPreprocessor_multi object at 0x7f8e12793c70>)
671
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,697 (abs_task:1615) INFO: [plot_att] Batch sampler: UnsortedBatchSampler(N-batch=452, batch_size=1, key_file=exp/asr_stats_raw_en_char_sp/valid/speech_shape,
672
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:32,697 (abs_task:1616) INFO: [plot_att] mini-batch sizes summary: N-batch=3, mean=1.0, min=1, max=1
673
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1753232 [0] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
674
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1753232 [0] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
675
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1753232 [0] NCCL INFO NET/IB : No device found.
676
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1753232 [0] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
677
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1753232 [0] NCCL INFO Using network Socket
678
+ NCCL version 2.10.3+cuda10.2
679
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1753233 [1] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
680
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1753234 [2] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
681
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1753233 [1] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
682
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1753234 [2] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
683
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1753235 [3] NCCL INFO Bootstrap : Using eth0:10.177.13.150<0>
684
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1753235 [3] NCCL INFO NET/Plugin : No plugin found (libnccl-net.so), using internal implementation
685
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1753233 [1] NCCL INFO NET/IB : No device found.
686
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1753233 [1] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
687
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1753233 [1] NCCL INFO Using network Socket
688
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1753234 [2] NCCL INFO NET/IB : No device found.
689
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1753234 [2] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
690
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1753234 [2] NCCL INFO Using network Socket
691
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1753235 [3] NCCL INFO NET/IB : No device found.
692
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1753235 [3] NCCL INFO NET/Socket : Using [0]eth0:10.177.13.150<0>
693
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1753235 [3] NCCL INFO Using network Socket
694
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Trees [0] 2/-1/-1->1->0 [1] 2/-1/-1->1->0
695
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Setting affinity for GPU 1 to 20000002
696
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Trees [0] 3/-1/-1->2->1 [1] 3/-1/-1->2->1
697
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Trees [0] -1/-1/-1->3->2 [1] -1/-1/-1->3->2
698
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Channel 00/02 : 0 1 2 3
699
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Setting affinity for GPU 2 to 20000002
700
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Setting affinity for GPU 5 to 20000002
701
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Channel 01/02 : 0 1 2 3
702
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Trees [0] 1/-1/-1->0->-1 [1] 1/-1/-1->0->-1
703
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Setting affinity for GPU 0 to 20000002
704
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Channel 00 : 2[7000] -> 3[e000] via direct shared memory
705
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Channel 01 : 2[7000] -> 3[e000] via direct shared memory
706
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Channel 00 : 1[6000] -> 2[7000] via P2P/IPC
707
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Channel 01 : 1[6000] -> 2[7000] via P2P/IPC
708
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Channel 00 : 0[4000] -> 1[6000] via P2P/IPC
709
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Channel 01 : 0[4000] -> 1[6000] via P2P/IPC
710
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Channel 00 : 3[e000] -> 0[4000] via direct shared memory
711
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Channel 01 : 3[e000] -> 0[4000] via direct shared memory
712
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Connected all rings
713
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Channel 00 : 1[6000] -> 0[4000] via P2P/IPC
714
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Channel 01 : 1[6000] -> 0[4000] via P2P/IPC
715
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Connected all rings
716
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Connected all rings
717
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO Connected all trees
718
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
719
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Connected all rings
720
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Channel 00 : 3[e000] -> 2[7000] via direct shared memory
721
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
722
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Channel 01 : 3[e000] -> 2[7000] via direct shared memory
723
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Channel 00 : 2[7000] -> 1[6000] via P2P/IPC
724
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Channel 01 : 2[7000] -> 1[6000] via P2P/IPC
725
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO Connected all trees
726
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
727
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
728
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO Connected all trees
729
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
730
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
731
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO Connected all trees
732
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO threadThresholds 8/8/64 | 32/8/64 | 8/8/512
733
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO 2 coll channels, 2 p2p channels, 2 p2p channels per peer
734
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753233:1756016 [1] NCCL INFO comm 0x7fc0a8001200 rank 1 nranks 4 cudaDev 1 busId 6000 - Init COMPLETE
735
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753235:1756018 [3] NCCL INFO comm 0x7f483c001200 rank 3 nranks 4 cudaDev 3 busId e000 - Init COMPLETE
736
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1756014 [0] NCCL INFO comm 0x7f8d5c001200 rank 0 nranks 4 cudaDev 0 busId 4000 - Init COMPLETE
737
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753232:1753232 [0] NCCL INFO Launch mode Parallel
738
+ de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:1753234:1756017 [2] NCCL INFO comm 0x7fbe0c001200 rank 2 nranks 4 cudaDev 2 busId 7000 - Init COMPLETE
739
+ [de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb:0/4] 2023-11-18 19:33:34,575 (trainer:284) INFO: 1/60epoch started
740
+ Traceback (most recent call last):
741
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 197, in _run_module_as_main
742
+ return _run_code(code, main_globals, None,
743
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 87, in _run_code
744
+ Traceback (most recent call last):
745
+ File "<string>", line 1, in <module>
746
+ Traceback (most recent call last):
747
+ File "<string>", line 1, in <module>
748
+ Traceback (most recent call last):
749
+ Traceback (most recent call last):
750
+ File "<string>", line 1, in <module>
751
+ File "<string>", line 1, in <module>
752
+ exec(code, run_globals)
753
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 23, in <module>
754
+ main()
755
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 19, in main
756
+ ASRTask.main(cmd=cmd)
757
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 1132, in main
758
+ while not ProcessContext(processes, error_queues).join():
759
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 109, in join
760
+ ready = multiprocessing.connection.wait(
761
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/connection.py", line 931, in wait
762
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 116, in spawn_main
763
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 116, in spawn_main
764
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 116, in spawn_main
765
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 116, in spawn_main
766
+ ready = selector.select(timeout)
767
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/selectors.py", line 416, in select
768
+ exitcode = _main(fd, parent_sentinel)
769
+ exitcode = _main(fd, parent_sentinel)
770
+ exitcode = _main(fd, parent_sentinel)
771
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 125, in _main
772
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 125, in _main
773
+ exitcode = _main(fd, parent_sentinel)
774
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 125, in _main
775
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 125, in _main
776
+ prepare(preparation_data)
777
+ prepare(preparation_data)
778
+ prepare(preparation_data)
779
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 234, in prepare
780
+ prepare(preparation_data)
781
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 234, in prepare
782
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 234, in prepare
783
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 234, in prepare
784
+ fd_event_list = self._selector.poll(timeout)
785
+ KeyboardInterrupt
786
+ _fixup_main_from_name(data['init_main_from_name'])
787
+ _fixup_main_from_name(data['init_main_from_name'])
788
+ _fixup_main_from_name(data['init_main_from_name'])
789
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 258, in _fixup_main_from_name
790
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 258, in _fixup_main_from_name
791
+ _fixup_main_from_name(data['init_main_from_name'])
792
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 258, in _fixup_main_from_name
793
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/spawn.py", line 258, in _fixup_main_from_name
794
+ main_content = runpy.run_module(mod_name,
795
+ main_content = runpy.run_module(mod_name,
796
+ main_content = runpy.run_module(mod_name,
797
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 225, in run_module
798
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 225, in run_module
799
+ main_content = runpy.run_module(mod_name,
800
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 225, in run_module
801
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 225, in run_module
802
+ return _run_module_code(code, init_globals, run_name, mod_spec)
803
+ return _run_module_code(code, init_globals, run_name, mod_spec)
804
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 97, in _run_module_code
805
+ return _run_module_code(code, init_globals, run_name, mod_spec)
806
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 97, in _run_module_code
807
+ return _run_module_code(code, init_globals, run_name, mod_spec)
808
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 97, in _run_module_code
809
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 97, in _run_module_code
810
+ _run_code(code, mod_globals, init_globals,
811
+ _run_code(code, mod_globals, init_globals,
812
+ _run_code(code, mod_globals, init_globals,
813
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 87, in _run_code
814
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 87, in _run_code
815
+ _run_code(code, mod_globals, init_globals,
816
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 87, in _run_code
817
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/runpy.py", line 87, in _run_code
818
+ exec(code, run_globals)
819
+ exec(code, run_globals)
820
+ exec(code, run_globals)
821
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 2, in <module>
822
+ exec(code, run_globals)
823
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 2, in <module>
824
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 2, in <module>
825
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/bin/asr_train.py", line 2, in <module>
826
+ from espnet2.tasks.asr import ASRTask
827
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/asr.py", line 73, in <module>
828
+ from espnet2.tasks.asr import ASRTask
829
+ from espnet2.tasks.asr import ASRTask
830
+ from espnet2.tasks.asr import ASRTask
831
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/asr.py", line 73, in <module>
832
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/asr.py", line 73, in <module>
833
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/asr.py", line 73, in <module>
834
+ from espnet2.tasks.abs_task import AbsTask
835
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 48, in <module>
836
+ from espnet2.tasks.abs_task import AbsTask
837
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 48, in <module>
838
+ from espnet2.tasks.abs_task import AbsTask
839
+ from espnet2.tasks.abs_task import AbsTask
840
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 48, in <module>
841
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 48, in <module>
842
+ from espnet2.train.dataset import DATA_TYPES, AbsDataset, ESPnetDataset
843
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/dataset.py", line 10, in <module>
844
+ from espnet2.train.dataset import DATA_TYPES, AbsDataset, ESPnetDataset
845
+ from espnet2.train.dataset import DATA_TYPES, AbsDataset, ESPnetDataset
846
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/dataset.py", line 10, in <module>
847
+ from espnet2.train.dataset import DATA_TYPES, AbsDataset, ESPnetDataset
848
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/dataset.py", line 10, in <module>
849
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/dataset.py", line 10, in <module>
850
+ import h5py
851
+ import h5py
852
+ import h5py
853
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/__init__.py", line 58, in <module>
854
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/__init__.py", line 58, in <module>
855
+ import h5py
856
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/__init__.py", line 58, in <module>
857
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/__init__.py", line 58, in <module>
858
+ from ._hl import filters
859
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/_hl/filters.py", line 44, in <module>
860
+ from ._hl import filters
861
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/_hl/filters.py", line 44, in <module>
862
+ from ._hl import filters
863
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/_hl/filters.py", line 44, in <module>
864
+ from ._hl import filters
865
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/h5py/_hl/filters.py", line 44, in <module>
866
+ from .base import product
867
+ File "<frozen importlib._bootstrap>", line 1007, in _find_and_load
868
+ from .base import product
869
+ File "<frozen importlib._bootstrap>", line 1007, in _find_and_load
870
+ from .base import product
871
+ File "<frozen importlib._bootstrap>", line 1007, in _find_and_load
872
+ from .base import product
873
+ File "<frozen importlib._bootstrap>", line 1007, in _find_and_load
874
+ File "<frozen importlib._bootstrap>", line 986, in _find_and_load_unlocked
875
+ File "<frozen importlib._bootstrap>", line 986, in _find_and_load_unlocked
876
+ File "<frozen importlib._bootstrap>", line 986, in _find_and_load_unlocked
877
+ File "<frozen importlib._bootstrap>", line 986, in _find_and_load_unlocked
878
+ File "<frozen importlib._bootstrap>", line 680, in _load_unlocked
879
+ File "<frozen importlib._bootstrap>", line 680, in _load_unlocked
880
+ File "<frozen importlib._bootstrap>", line 680, in _load_unlocked
881
+ File "<frozen importlib._bootstrap_external>", line 846, in exec_module
882
+ File "<frozen importlib._bootstrap>", line 680, in _load_unlocked
883
+ File "<frozen importlib._bootstrap_external>", line 846, in exec_module
884
+ File "<frozen importlib._bootstrap_external>", line 846, in exec_module
885
+ File "<frozen importlib._bootstrap_external>", line 941, in get_code
886
+ File "<frozen importlib._bootstrap_external>", line 941, in get_code
887
+ File "<frozen importlib._bootstrap_external>", line 941, in get_code
888
+ File "<frozen importlib._bootstrap_external>", line 1039, in get_data
889
+ File "<frozen importlib._bootstrap_external>", line 1039, in get_data
890
+ File "<frozen importlib._bootstrap_external>", line 846, in exec_module
891
+ KeyboardInterrupt
892
+ File "<frozen importlib._bootstrap_external>", line 1039, in get_data
893
+ File "<frozen importlib._bootstrap_external>", line 941, in get_code
894
+ KeyboardInterrupt
895
+ KeyboardInterrupt
896
+ File "<frozen importlib._bootstrap_external>", line 1039, in get_data
897
+ KeyboardInterrupt
898
+ Process SpawnProcess-3:
899
+ Process SpawnProcess-1:
900
+ Process SpawnProcess-4:
901
+ Process SpawnProcess-2:
902
+ Traceback (most recent call last):
903
+ Traceback (most recent call last):
904
+ Traceback (most recent call last):
905
+ Traceback (most recent call last):
906
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 315, in _bootstrap
907
+ self.run()
908
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 315, in _bootstrap
909
+ self.run()
910
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 108, in run
911
+ self._target(*self._args, **self._kwargs)
912
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 315, in _bootstrap
913
+ self.run()
914
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 108, in run
915
+ self._target(*self._args, **self._kwargs)
916
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 1391, in main_worker
917
+ cls.trainer.run(
918
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 315, in _bootstrap
919
+ self.run()
920
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 108, in run
921
+ self._target(*self._args, **self._kwargs)
922
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 1391, in main_worker
923
+ cls.trainer.run(
924
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 290, in run
925
+ all_steps_are_invalid = cls.train_one_epoch(
926
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 108, in run
927
+ self._target(*self._args, **self._kwargs)
928
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 1391, in main_worker
929
+ cls.trainer.run(
930
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 290, in run
931
+ all_steps_are_invalid = cls.train_one_epoch(
932
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 510, in train_one_epoch
933
+ for iiter, (utt_id, batch) in enumerate(
934
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/tasks/abs_task.py", line 1391, in main_worker
935
+ cls.trainer.run(
936
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 290, in run
937
+ all_steps_are_invalid = cls.train_one_epoch(
938
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 510, in train_one_epoch
939
+ for iiter, (utt_id, batch) in enumerate(
940
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/reporter.py", line 263, in measure_iter_time
941
+ iterator = iter(iterable)
942
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 290, in run
943
+ all_steps_are_invalid = cls.train_one_epoch(
944
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 510, in train_one_epoch
945
+ for iiter, (utt_id, batch) in enumerate(
946
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/reporter.py", line 263, in measure_iter_time
947
+ iterator = iter(iterable)
948
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 368, in __iter__
949
+ return self._get_iterator()
950
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/trainer.py", line 510, in train_one_epoch
951
+ for iiter, (utt_id, batch) in enumerate(
952
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/reporter.py", line 263, in measure_iter_time
953
+ iterator = iter(iterable)
954
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 368, in __iter__
955
+ return self._get_iterator()
956
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 314, in _get_iterator
957
+ return _MultiProcessingDataLoaderIter(self)
958
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/espnet-202308-py3.9.egg/espnet2/train/reporter.py", line 263, in measure_iter_time
959
+ iterator = iter(iterable)
960
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 368, in __iter__
961
+ return self._get_iterator()
962
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 314, in _get_iterator
963
+ return _MultiProcessingDataLoaderIter(self)
964
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 927, in __init__
965
+ w.start()
966
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 368, in __iter__
967
+ return self._get_iterator()
968
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 314, in _get_iterator
969
+ return _MultiProcessingDataLoaderIter(self)
970
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 927, in __init__
971
+ w.start()
972
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 121, in start
973
+ self._popen = self._Popen(self)
974
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 314, in _get_iterator
975
+ return _MultiProcessingDataLoaderIter(self)
976
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 927, in __init__
977
+ w.start()
978
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 121, in start
979
+ self._popen = self._Popen(self)
980
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 224, in _Popen
981
+ return _default_context.get_context().Process._Popen(process_obj)
982
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 927, in __init__
983
+ w.start()
984
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 121, in start
985
+ self._popen = self._Popen(self)
986
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 224, in _Popen
987
+ return _default_context.get_context().Process._Popen(process_obj)
988
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 224, in _Popen
989
+ return _default_context.get_context().Process._Popen(process_obj)
990
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 284, in _Popen
991
+ return Popen(process_obj)
992
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/process.py", line 121, in start
993
+ self._popen = self._Popen(self)
994
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 284, in _Popen
995
+ return Popen(process_obj)
996
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 284, in _Popen
997
+ return Popen(process_obj)
998
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 32, in __init__
999
+ super().__init__(process_obj)
1000
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 224, in _Popen
1001
+ return _default_context.get_context().Process._Popen(process_obj)
1002
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 32, in __init__
1003
+ super().__init__(process_obj)
1004
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 32, in __init__
1005
+ super().__init__(process_obj)
1006
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_fork.py", line 19, in __init__
1007
+ self._launch(process_obj)
1008
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_fork.py", line 19, in __init__
1009
+ self._launch(process_obj)
1010
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/context.py", line 284, in _Popen
1011
+ return Popen(process_obj)
1012
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_fork.py", line 19, in __init__
1013
+ self._launch(process_obj)
1014
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 62, in _launch
1015
+ f.write(fp.getbuffer())
1016
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 62, in _launch
1017
+ f.write(fp.getbuffer())
1018
+ KeyboardInterrupt
1019
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 32, in __init__
1020
+ super().__init__(process_obj)
1021
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 62, in _launch
1022
+ f.write(fp.getbuffer())
1023
+ KeyboardInterrupt
1024
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_fork.py", line 19, in __init__
1025
+ self._launch(process_obj)
1026
+ KeyboardInterrupt
1027
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/multiprocessing/popen_spawn_posix.py", line 62, in _launch
1028
+ f.write(fp.getbuffer())
1029
+ KeyboardInterrupt
1030
+ Exception ignored in: <function _MultiProcessingDataLoaderIter.__del__ at 0x7fc1763ce5e0>
1031
+ Traceback (most recent call last):
1032
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1358, in __del__
1033
+ Exception ignored in: <function _MultiProcessingDataLoaderIter.__del__ at 0x7f490af825e0>
1034
+ Traceback (most recent call last):
1035
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1358, in __del__
1036
+ Exception ignored in: <function _MultiProcessingDataLoaderIter.__del__ at 0x7fbedaa735e0>
1037
+ Traceback (most recent call last):
1038
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1358, in __del__
1039
+ self._shutdown_workers()
1040
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1316, in _shutdown_workers
1041
+ self._shutdown_workers()
1042
+ self._shutdown_workers()
1043
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1316, in _shutdown_workers
1044
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1316, in _shutdown_workers
1045
+ Exception ignored in: <function _MultiProcessingDataLoaderIter.__del__ at 0x7f8e352eb5e0>
1046
+ Traceback (most recent call last):
1047
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1358, in __del__
1048
+ if self._persistent_workers or self._workers_status[worker_id]:
1049
+ if self._persistent_workers or self._workers_status[worker_id]:
1050
+ if self._persistent_workers or self._workers_status[worker_id]:
1051
+ AttributeError: '_MultiProcessingDataLoaderIter' object has no attribute '_workers_status'
1052
+ AttributeError: '_MultiProcessingDataLoaderIter' object has no attribute '_workers_status'
1053
+ AttributeError: '_MultiProcessingDataLoaderIter' object has no attribute '_workers_status'
1054
+ self._shutdown_workers()
1055
+ File "/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/torch/utils/data/dataloader.py", line 1316, in _shutdown_workers
1056
+ if self._persistent_workers or self._workers_status[worker_id]:
1057
+ AttributeError: '_MultiProcessingDataLoaderIter' object has no attribute '_workers_status'
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/train.log ADDED
The diff for this file is too large to render. See raw diff
 
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/valid.acc.ave.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6942cd8e2a7d9015a74726b4164bd69324059cb230e97ebccf434f6b1afdee4c
3
+ size 172358249
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/valid.acc.ave_10best.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6942cd8e2a7d9015a74726b4164bd69324059cb230e97ebccf434f6b1afdee4c
3
+ size 172358249
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new/valid.acc.best.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63fced0fdc8817cd4f376fe8e33ffc3b07a8615ce2151f6b5bd6be0990bb75ff
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/48epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5a83dbcc31cbe211bf57d554185f00fdefa382c828fadc33a1202749a7cafbc
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/52epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8359149a4f68dfaf1fb54be64f57b199f0c2a2ce86c123ca9479a8f29854ab59
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/53epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f7928e027e6aaca6f547f40d08d36ed21ec877c81f13b6e8432ffd119a4f635
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/54epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2917e31143d7136e849de1dee47a99942384a481fe11ed989b17d7faf676170c
3
+ size 172367337
medium/asr_train_sot_asr_conformer_raw_en_char_sp_finetune_ls100_45epoch_new_whamr/55epoch.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:035c566070194c490d66447e070ae13154064c7d387b0e57a13693deabadd34d
3
+ size 172367337