Hugging Face's logo

Padomin
/
t5-base-TEDxJP-3front-1body-0rear

Text2Text Generation
Transformers
PyTorch
TensorBoard
t5
Generated from Trainer
text-generation-inference
Model card Files Metrics Community
Padomin commited on
Commit
9bc1cda
·
1 Parent(s): d500a5d

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: 0.4681
20
- - Wer: 0.1753
21
- - Mer: 0.1694
22
- - Wil: 0.2570
23
- - Wip: 0.7430
24
- - Hits: 55547
25
- - Substitutions: 6465
26
- - Deletions: 2575
27
- - Insertions: 2285
28
- - Cer: 0.1373
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
- | 0.6131 | 1.0 | 1457 | 0.4967 | 0.2415 | 0.2225 | 0.3113 | 0.6887 | 54519 | 6769 | 3299 | 5532 | 0.2182 |
61
- | 0.5697 | 2.0 | 2914 | 0.4554 | 0.1948 | 0.1857 | 0.2737 | 0.7263 | 55187 | 6553 | 2847 | 3184 | 0.1621 |
62
- | 0.4783 | 3.0 | 4371 | 0.4414 | 0.1801 | 0.1736 | 0.2615 | 0.7385 | 55365 | 6489 | 2733 | 2408 | 0.1444 |
63
- | 0.3879 | 4.0 | 5828 | 0.4442 | 0.1763 | 0.1704 | 0.2582 | 0.7418 | 55428 | 6468 | 2691 | 2229 | 0.1373 |
64
- | 0.3965 | 5.0 | 7285 | 0.4476 | 0.1757 | 0.1698 | 0.2569 | 0.7431 | 55490 | 6414 | 2683 | 2249 | 0.1372 |
65
- | 0.3939 | 6.0 | 8742 | 0.4456 | 0.1748 | 0.1690 | 0.2558 | 0.7442 | 55513 | 6385 | 2689 | 2214 | 0.1365 |
66
- | 0.3542 | 7.0 | 10199 | 0.4510 | 0.1761 | 0.1701 | 0.2575 | 0.7425 | 55490 | 6444 | 2653 | 2275 | 0.1384 |
67
- | 0.3376 | 8.0 | 11656 | 0.4576 | 0.1758 | 0.1698 | 0.2572 | 0.7428 | 55500 | 6438 | 2649 | 2267 | 0.1389 |
68
- | 0.306 | 9.0 | 13113 | 0.4657 | 0.1753 | 0.1692 | 0.2569 | 0.7431 | 55582 | 6471 | 2534 | 2318 | 0.1375 |
69
- | 0.344 | 10.0 | 14570 | 0.4681 | 0.1753 | 0.1694 | 0.2570 | 0.7430 | 55547 | 6465 | 2575 | 2285 | 0.1373 |
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: 0.4641
20
+ - Wer: 0.1743
21
+ - Mer: 0.1684
22
+ - Wil: 0.2557
23
+ - Wip: 0.7443
24
+ - Hits: 55594
25
+ - Substitutions: 6428
26
+ - Deletions: 2565
27
+ - Insertions: 2267
28
+ - Cer: 0.1368
29
 
30
  ## Model description
31
 
 
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | Wer | Mer | Wil | Wip | Hits | Substitutions | Deletions | Insertions | Cer |
59
  |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:------:|:-----:|:-------------:|:---------:|:----------:|:------:|
60
+ | 0.6567 | 1.0 | 1457 | 0.4959 | 0.2072 | 0.1966 | 0.2877 | 0.7123 | 54688 | 6836 | 3063 | 3486 | 0.1936 |
61
+ | 0.5486 | 2.0 | 2914 | 0.4504 | 0.1870 | 0.1796 | 0.2677 | 0.7323 | 55158 | 6518 | 2911 | 2647 | 0.1528 |
62
+ | 0.4957 | 3.0 | 4371 | 0.4410 | 0.1764 | 0.1705 | 0.2578 | 0.7422 | 55412 | 6429 | 2746 | 2216 | 0.1375 |
63
+ | 0.4371 | 4.0 | 5828 | 0.4379 | 0.1761 | 0.1702 | 0.2572 | 0.7428 | 55447 | 6407 | 2733 | 2232 | 0.1377 |
64
+ | 0.387 | 5.0 | 7285 | 0.4408 | 0.1756 | 0.1696 | 0.2562 | 0.7438 | 55510 | 6372 | 2705 | 2263 | 0.1399 |
65
+ | 0.3589 | 6.0 | 8742 | 0.4466 | 0.1737 | 0.1681 | 0.2552 | 0.7448 | 55532 | 6406 | 2649 | 2165 | 0.1359 |
66
+ | 0.3876 | 7.0 | 10199 | 0.4532 | 0.1746 | 0.1689 | 0.2563 | 0.7437 | 55491 | 6436 | 2660 | 2179 | 0.1363 |
67
+ | 0.3199 | 8.0 | 11656 | 0.4591 | 0.1738 | 0.1681 | 0.2554 | 0.7446 | 55568 | 6431 | 2588 | 2208 | 0.1362 |
68
+ | 0.3079 | 9.0 | 13113 | 0.4625 | 0.1743 | 0.1685 | 0.2557 | 0.7443 | 55579 | 6425 | 2583 | 2252 | 0.1366 |
69
+ | 0.3124 | 10.0 | 14570 | 0.4641 | 0.1743 | 0.1684 | 0.2557 | 0.7443 | 55594 | 6428 | 2565 | 2267 | 0.1368 |
70
 
71
 
72
  ### Framework versions