update model card README.md
Browse files
README.md
CHANGED
@@ -1,8 +1,6 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
tags:
|
4 |
-
- automatic-speech-recognition
|
5 |
-
- gary109/AI_Light_Dance
|
6 |
- generated_from_trainer
|
7 |
model-index:
|
8 |
- name: ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1
|
@@ -14,10 +12,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
14 |
|
15 |
# ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1
|
16 |
|
17 |
-
This model is a fine-tuned version of [gary109/ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1](https://huggingface.co/gary109/ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1) on the
|
18 |
It achieves the following results on the evaluation set:
|
19 |
-
- Loss: 0.
|
20 |
-
- Wer: 0.
|
21 |
|
22 |
## Model description
|
23 |
|
@@ -36,72 +34,72 @@ More information needed
|
|
36 |
### Training hyperparameters
|
37 |
|
38 |
The following hyperparameters were used during training:
|
39 |
-
- learning_rate:
|
40 |
-
- train_batch_size:
|
41 |
-
- eval_batch_size:
|
42 |
- seed: 42
|
43 |
-
- gradient_accumulation_steps:
|
44 |
-
- total_train_batch_size:
|
45 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
46 |
- lr_scheduler_type: linear
|
47 |
-
- lr_scheduler_warmup_steps:
|
48 |
- num_epochs: 50.0
|
49 |
- mixed_precision_training: Native AMP
|
50 |
|
51 |
### Training results
|
52 |
|
53 |
-
| Training Loss | Epoch | Step
|
54 |
-
|
55 |
-
| 0.
|
56 |
-
| 0.
|
57 |
-
| 0.
|
58 |
-
| 0.
|
59 |
-
| 0.
|
60 |
-
| 0.
|
61 |
-
| 0.
|
62 |
-
| 0.
|
63 |
-
| 0.
|
64 |
-
| 0.
|
65 |
-
| 0.
|
66 |
-
| 0.
|
67 |
-
| 0.
|
68 |
-
| 0.
|
69 |
-
| 0.
|
70 |
-
| 0.
|
71 |
-
| 0.
|
72 |
-
| 0.
|
73 |
-
| 0.
|
74 |
-
| 0.
|
75 |
-
| 0.
|
76 |
-
| 0.
|
77 |
-
| 0.
|
78 |
-
| 0.
|
79 |
-
| 0.
|
80 |
-
| 0.
|
81 |
-
| 0.
|
82 |
-
| 0.
|
83 |
-
| 0.
|
84 |
-
| 0.
|
85 |
-
| 0.
|
86 |
-
| 0.
|
87 |
-
| 0.
|
88 |
-
| 0.
|
89 |
-
| 0.
|
90 |
-
| 0.
|
91 |
-
| 0.
|
92 |
-
| 0.
|
93 |
-
| 0.
|
94 |
-
| 0.
|
95 |
-
| 0.
|
96 |
-
| 0.
|
97 |
-
| 0.
|
98 |
-
| 0.
|
99 |
-
| 0.
|
100 |
-
| 0.
|
101 |
-
| 0.
|
102 |
-
| 0.
|
103 |
-
| 0.
|
104 |
-
| 0.
|
105 |
|
106 |
|
107 |
### Framework versions
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
tags:
|
|
|
|
|
4 |
- generated_from_trainer
|
5 |
model-index:
|
6 |
- name: ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1
|
|
|
12 |
|
13 |
# ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1
|
14 |
|
15 |
+
This model is a fine-tuned version of [gary109/ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1](https://huggingface.co/gary109/ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1) on the None dataset.
|
16 |
It achieves the following results on the evaluation set:
|
17 |
+
- Loss: 0.5403
|
18 |
+
- Wer: 0.2461
|
19 |
|
20 |
## Model description
|
21 |
|
|
|
34 |
### Training hyperparameters
|
35 |
|
36 |
The following hyperparameters were used during training:
|
37 |
+
- learning_rate: 2e-06
|
38 |
+
- train_batch_size: 8
|
39 |
+
- eval_batch_size: 8
|
40 |
- seed: 42
|
41 |
+
- gradient_accumulation_steps: 8
|
42 |
+
- total_train_batch_size: 64
|
43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
44 |
- lr_scheduler_type: linear
|
45 |
+
- lr_scheduler_warmup_steps: 50
|
46 |
- num_epochs: 50.0
|
47 |
- mixed_precision_training: Native AMP
|
48 |
|
49 |
### Training results
|
50 |
|
51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
52 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
53 |
+
| 0.3818 | 1.0 | 72 | 0.5569 | 0.2535 |
|
54 |
+
| 0.3686 | 2.0 | 144 | 0.5535 | 0.2501 |
|
55 |
+
| 0.3562 | 3.0 | 216 | 0.5526 | 0.2501 |
|
56 |
+
| 0.3506 | 4.0 | 288 | 0.5460 | 0.2520 |
|
57 |
+
| 0.369 | 5.0 | 360 | 0.5390 | 0.2484 |
|
58 |
+
| 0.3683 | 6.0 | 432 | 0.5426 | 0.2474 |
|
59 |
+
| 0.3541 | 7.0 | 504 | 0.5452 | 0.2495 |
|
60 |
+
| 0.369 | 8.0 | 576 | 0.5468 | 0.2490 |
|
61 |
+
| 0.358 | 9.0 | 648 | 0.5471 | 0.2453 |
|
62 |
+
| 0.3765 | 10.0 | 720 | 0.5376 | 0.2459 |
|
63 |
+
| 0.3654 | 11.0 | 792 | 0.5407 | 0.2486 |
|
64 |
+
| 0.373 | 12.0 | 864 | 0.5390 | 0.2475 |
|
65 |
+
| 0.3606 | 13.0 | 936 | 0.5441 | 0.2472 |
|
66 |
+
| 0.369 | 14.0 | 1008 | 0.5388 | 0.2490 |
|
67 |
+
| 0.3762 | 15.0 | 1080 | 0.5389 | 0.2497 |
|
68 |
+
| 0.3821 | 16.0 | 1152 | 0.5352 | 0.2490 |
|
69 |
+
| 0.3605 | 17.0 | 1224 | 0.5434 | 0.2492 |
|
70 |
+
| 0.3703 | 18.0 | 1296 | 0.5415 | 0.2500 |
|
71 |
+
| 0.3667 | 19.0 | 1368 | 0.5499 | 0.2487 |
|
72 |
+
| 0.3703 | 20.0 | 1440 | 0.5482 | 0.2494 |
|
73 |
+
| 0.369 | 21.0 | 1512 | 0.5418 | 0.2504 |
|
74 |
+
| 0.3708 | 22.0 | 1584 | 0.5437 | 0.2479 |
|
75 |
+
| 0.3609 | 23.0 | 1656 | 0.5453 | 0.2480 |
|
76 |
+
| 0.3534 | 24.0 | 1728 | 0.5393 | 0.2484 |
|
77 |
+
| 0.3656 | 25.0 | 1800 | 0.5363 | 0.2477 |
|
78 |
+
| 0.3713 | 26.0 | 1872 | 0.5406 | 0.2461 |
|
79 |
+
| 0.3572 | 27.0 | 1944 | 0.5369 | 0.2465 |
|
80 |
+
| 0.3665 | 28.0 | 2016 | 0.5375 | 0.2466 |
|
81 |
+
| 0.3748 | 29.0 | 2088 | 0.5367 | 0.2480 |
|
82 |
+
| 0.3769 | 30.0 | 2160 | 0.5359 | 0.2459 |
|
83 |
+
| 0.3634 | 31.0 | 2232 | 0.5417 | 0.2471 |
|
84 |
+
| 0.3627 | 32.0 | 2304 | 0.5398 | 0.2473 |
|
85 |
+
| 0.3497 | 33.0 | 2376 | 0.5428 | 0.2483 |
|
86 |
+
| 0.3479 | 34.0 | 2448 | 0.5390 | 0.2481 |
|
87 |
+
| 0.363 | 35.0 | 2520 | 0.5388 | 0.2478 |
|
88 |
+
| 0.3622 | 36.0 | 2592 | 0.5396 | 0.2490 |
|
89 |
+
| 0.3685 | 37.0 | 2664 | 0.5408 | 0.2479 |
|
90 |
+
| 0.356 | 38.0 | 2736 | 0.5385 | 0.2459 |
|
91 |
+
| 0.3529 | 39.0 | 2808 | 0.5389 | 0.2467 |
|
92 |
+
| 0.3702 | 40.0 | 2880 | 0.5392 | 0.2482 |
|
93 |
+
| 0.3645 | 41.0 | 2952 | 0.5408 | 0.2467 |
|
94 |
+
| 0.3489 | 42.0 | 3024 | 0.5406 | 0.2474 |
|
95 |
+
| 0.3556 | 43.0 | 3096 | 0.5402 | 0.2472 |
|
96 |
+
| 0.356 | 44.0 | 3168 | 0.5386 | 0.2468 |
|
97 |
+
| 0.3632 | 45.0 | 3240 | 0.5402 | 0.2463 |
|
98 |
+
| 0.3693 | 46.0 | 3312 | 0.5401 | 0.2472 |
|
99 |
+
| 0.3593 | 47.0 | 3384 | 0.5390 | 0.2463 |
|
100 |
+
| 0.3515 | 48.0 | 3456 | 0.5399 | 0.2468 |
|
101 |
+
| 0.3485 | 49.0 | 3528 | 0.5392 | 0.2461 |
|
102 |
+
| 0.3591 | 50.0 | 3600 | 0.5403 | 0.2461 |
|
103 |
|
104 |
|
105 |
### Framework versions
|