Hugging Face's logo

Padomin
/
t5-base-TEDxJP-8front-1body-0rear

Transformers
PyTorch
TensorBoard
t5
text2text-generation
Generated from Trainer
text-generation-inference
Model card Files
xet
Metrics Community
Padomin commited on
Commit
03f6433
·
1 Parent(s): ba1d63c

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +20 -20
README.md CHANGED
@@ -16,16 +16,16 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  This model is a fine-tuned version of [sonoisa/t5-base-japanese](https://huggingface.co/sonoisa/t5-base-japanese) on the te_dx_jp dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.4587
20
- - Wer: 0.1742
21
- - Mer: 0.1681
22
- - Wil: 0.2560
23
- - Wip: 0.7440
24
- - Hits: 55670
25
- - Substitutions: 6489
26
- - Deletions: 2428
27
- - Insertions: 2332
28
- - Cer: 0.1366
29
 
30
  ## Model description
31
 
@@ -57,16 +57,16 @@ The following hyperparameters were used during training:
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | Wer | Mer | Wil | Wip | Hits | Substitutions | Deletions | Insertions | Cer |
59
  |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:------:|:-----:|:-------------:|:---------:|:----------:|:------:|
60
- | 0.6291 | 1.0 | 1457 | 0.4916 | 0.2265 | 0.2109 | 0.3006 | 0.6994 | 54712 | 6799 | 3076 | 4751 | 0.1974 |
61
- | 0.5771 | 2.0 | 2914 | 0.4435 | 0.1887 | 0.1808 | 0.2689 | 0.7311 | 55222 | 6539 | 2826 | 2820 | 0.1620 |
62
- | 0.4774 | 3.0 | 4371 | 0.4331 | 0.1762 | 0.1704 | 0.2582 | 0.7418 | 55428 | 6473 | 2686 | 2224 | 0.1371 |
63
- | 0.4182 | 4.0 | 5828 | 0.4298 | 0.1752 | 0.1693 | 0.2571 | 0.7429 | 55509 | 6470 | 2608 | 2238 | 0.1368 |
64
- | 0.3797 | 5.0 | 7285 | 0.4331 | 0.1729 | 0.1675 | 0.2545 | 0.7455 | 55518 | 6392 | 2677 | 2100 | 0.1351 |
65
- | 0.3733 | 6.0 | 8742 | 0.4395 | 0.1738 | 0.1679 | 0.2556 | 0.7444 | 55635 | 6474 | 2478 | 2273 | 0.1346 |
66
- | 0.3236 | 7.0 | 10199 | 0.4467 | 0.1740 | 0.1679 | 0.2552 | 0.7448 | 55668 | 6439 | 2480 | 2317 | 0.1357 |
67
- | 0.3304 | 8.0 | 11656 | 0.4511 | 0.1747 | 0.1686 | 0.2566 | 0.7434 | 55645 | 6500 | 2442 | 2342 | 0.1364 |
68
- | 0.3032 | 9.0 | 13113 | 0.4534 | 0.1738 | 0.1678 | 0.2552 | 0.7448 | 55675 | 6450 | 2462 | 2311 | 0.1365 |
69
- | 0.3024 | 10.0 | 14570 | 0.4587 | 0.1742 | 0.1681 | 0.2560 | 0.7440 | 55670 | 6489 | 2428 | 2332 | 0.1366 |
70
 
71
 
72
  ### Framework versions
 
16
 
17
  This model is a fine-tuned version of [sonoisa/t5-base-japanese](https://huggingface.co/sonoisa/t5-base-japanese) on the te_dx_jp dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.4552
20
+ - Wer: 0.1738
21
+ - Mer: 0.1679
22
+ - Wil: 0.2549
23
+ - Wip: 0.7451
24
+ - Hits: 55655
25
+ - Substitutions: 6417
26
+ - Deletions: 2515
27
+ - Insertions: 2295
28
+ - Cer: 0.1358
29
 
30
  ## Model description
31
 
 
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | Wer | Mer | Wil | Wip | Hits | Substitutions | Deletions | Insertions | Cer |
59
  |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:------:|:-----:|:-------------:|:---------:|:----------:|:------:|
60
+ | 0.6066 | 1.0 | 1457 | 0.4857 | 0.2409 | 0.2216 | 0.3093 | 0.6907 | 54673 | 6687 | 3227 | 5647 | 0.2134 |
61
+ | 0.5605 | 2.0 | 2914 | 0.4415 | 0.1932 | 0.1838 | 0.2713 | 0.7287 | 55433 | 6531 | 2623 | 3327 | 0.1579 |
62
+ | 0.4629 | 3.0 | 4371 | 0.4327 | 0.1789 | 0.1721 | 0.2603 | 0.7397 | 55580 | 6534 | 2473 | 2548 | 0.1440 |
63
+ | 0.3676 | 4.0 | 5828 | 0.4309 | 0.1729 | 0.1671 | 0.2533 | 0.7467 | 55663 | 6337 | 2587 | 2243 | 0.1358 |
64
+ | 0.3862 | 5.0 | 7285 | 0.4359 | 0.1737 | 0.1677 | 0.2539 | 0.7461 | 55698 | 6350 | 2539 | 2330 | 0.1352 |
65
+ | 0.3699 | 6.0 | 8742 | 0.4355 | 0.1722 | 0.1665 | 0.2527 | 0.7473 | 55686 | 6340 | 2561 | 2221 | 0.1349 |
66
+ | 0.3474 | 7.0 | 10199 | 0.4415 | 0.1732 | 0.1673 | 0.2540 | 0.7460 | 55686 | 6390 | 2511 | 2286 | 0.1351 |
67
+ | 0.3278 | 8.0 | 11656 | 0.4466 | 0.1744 | 0.1684 | 0.2557 | 0.7443 | 55629 | 6441 | 2517 | 2308 | 0.1356 |
68
+ | 0.2957 | 9.0 | 13113 | 0.4541 | 0.1740 | 0.1680 | 0.2551 | 0.7449 | 55656 | 6422 | 2509 | 2310 | 0.1363 |
69
+ | 0.3434 | 10.0 | 14570 | 0.4552 | 0.1738 | 0.1679 | 0.2549 | 0.7451 | 55655 | 6417 | 2515 | 2295 | 0.1358 |
70
 
71
 
72
  ### Framework versions