Hugging Face's logo

Padomin
/
t5-base-TEDxJP-0front-1body-10rear

Transformers
PyTorch
TensorBoard
t5
text2text-generation
Generated from Trainer
text-generation-inference
Model card Files
xet
Metrics Community
Padomin commited on
Commit
0d88f6d
·
1 Parent(s): 67eb3c7

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: 0.4651
20
- - Wer: 0.1755
21
- - Mer: 0.1696
22
- - Wil: 0.2578
23
- - Wip: 0.7422
24
- - Hits: 55485
25
- - Substitutions: 6506
26
- - Deletions: 2596
27
- - Insertions: 2232
28
- - Cer: 0.1385
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
- | 0.6039 | 1.0 | 1457 | 0.4876 | 0.2540 | 0.2305 | 0.3180 | 0.6820 | 54757 | 6739 | 3091 | 6572 | 0.2244 |
61
- | 0.5521 | 2.0 | 2914 | 0.4520 | 0.1914 | 0.1828 | 0.2711 | 0.7289 | 55273 | 6572 | 2742 | 3049 | 0.1583 |
62
- | 0.4726 | 3.0 | 4371 | 0.4400 | 0.1760 | 0.1704 | 0.2588 | 0.7412 | 55329 | 6508 | 2750 | 2107 | 0.1385 |
63
- | 0.3703 | 4.0 | 5828 | 0.4438 | 0.1751 | 0.1693 | 0.2568 | 0.7432 | 55483 | 6442 | 2662 | 2203 | 0.1370 |
64
- | 0.383 | 5.0 | 7285 | 0.4406 | 0.1788 | 0.1723 | 0.2591 | 0.7409 | 55462 | 6398 | 2727 | 2421 | 0.1413 |
65
- | 0.3882 | 6.0 | 8742 | 0.4425 | 0.1763 | 0.1705 | 0.2585 | 0.7415 | 55407 | 6486 | 2694 | 2205 | 0.1387 |
66
- | 0.3487 | 7.0 | 10199 | 0.4525 | 0.1744 | 0.1687 | 0.2562 | 0.7438 | 55517 | 6446 | 2624 | 2194 | 0.1360 |
67
- | 0.3298 | 8.0 | 11656 | 0.4562 | 0.1753 | 0.1696 | 0.2580 | 0.7420 | 55457 | 6519 | 2611 | 2195 | 0.1378 |
68
- | 0.2947 | 9.0 | 13113 | 0.4655 | 0.1752 | 0.1694 | 0.2576 | 0.7424 | 55482 | 6512 | 2593 | 2208 | 0.1377 |
69
- | 0.329 | 10.0 | 14570 | 0.4651 | 0.1755 | 0.1696 | 0.2578 | 0.7422 | 55485 | 6506 | 2596 | 2232 | 0.1385 |
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: 0.4646
20
+ - Wer: 0.1756
21
+ - Mer: 0.1698
22
+ - Wil: 0.2581
23
+ - Wip: 0.7419
24
+ - Hits: 55450
25
+ - Substitutions: 6518
26
+ - Deletions: 2619
27
+ - Insertions: 2202
28
+ - Cer: 0.1383
29
 
30
  ## Model description
31
 
 
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | Wer | Mer | Wil | Wip | Hits | Substitutions | Deletions | Insertions | Cer |
59
  |:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:------:|:-----:|:-------------:|:---------:|:----------:|:------:|
60
+ | 0.6456 | 1.0 | 1457 | 0.4955 | 0.2298 | 0.2127 | 0.3032 | 0.6968 | 54935 | 6936 | 2716 | 5190 | 0.2008 |
61
+ | 0.5333 | 2.0 | 2914 | 0.4472 | 0.1817 | 0.1755 | 0.2646 | 0.7354 | 55142 | 6584 | 2861 | 2291 | 0.1431 |
62
+ | 0.4864 | 3.0 | 4371 | 0.4420 | 0.1774 | 0.1714 | 0.2600 | 0.7400 | 55396 | 6542 | 2649 | 2266 | 0.1397 |
63
+ | 0.429 | 4.0 | 5828 | 0.4374 | 0.1764 | 0.1704 | 0.2587 | 0.7413 | 55446 | 6512 | 2629 | 2249 | 0.1389 |
64
+ | 0.3741 | 5.0 | 7285 | 0.4413 | 0.1744 | 0.1687 | 0.2559 | 0.7441 | 55518 | 6416 | 2653 | 2198 | 0.1383 |
65
+ | 0.3467 | 6.0 | 8742 | 0.4467 | 0.1742 | 0.1686 | 0.2564 | 0.7436 | 55493 | 6466 | 2628 | 2159 | 0.1390 |
66
+ | 0.3761 | 7.0 | 10199 | 0.4524 | 0.1754 | 0.1696 | 0.2577 | 0.7423 | 55471 | 6498 | 2618 | 2210 | 0.1380 |
67
+ | 0.3102 | 8.0 | 11656 | 0.4557 | 0.1751 | 0.1695 | 0.2574 | 0.7426 | 55412 | 6478 | 2697 | 2133 | 0.1395 |
68
+ | 0.3008 | 9.0 | 13113 | 0.4632 | 0.1758 | 0.1700 | 0.2584 | 0.7416 | 55421 | 6516 | 2650 | 2189 | 0.1387 |
69
+ | 0.3051 | 10.0 | 14570 | 0.4646 | 0.1756 | 0.1698 | 0.2581 | 0.7419 | 55450 | 6518 | 2619 | 2202 | 0.1383 |
70
 
71
 
72
  ### Framework versions