saattrupdan
commited on
Commit
•
0a00b58
1
Parent(s):
a3d38d7
update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,117 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: cc0-1.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- common_voice
|
7 |
+
model-index:
|
8 |
+
- name: kblab-voxrex-wav2vec2-large-cv8-da
|
9 |
+
results: []
|
10 |
+
---
|
11 |
+
|
12 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
13 |
+
should probably proofread and complete it, then remove this comment. -->
|
14 |
+
|
15 |
+
# kblab-voxrex-wav2vec2-large-cv8-da
|
16 |
+
|
17 |
+
This model is a fine-tuned version of [KBLab/wav2vec2-large-voxrex](https://huggingface.co/KBLab/wav2vec2-large-voxrex) on the common_voice dataset.
|
18 |
+
It achieves the following results on the evaluation set:
|
19 |
+
- Loss: 334.1232
|
20 |
+
- Wer: 0.3729
|
21 |
+
|
22 |
+
## Model description
|
23 |
+
|
24 |
+
More information needed
|
25 |
+
|
26 |
+
## Intended uses & limitations
|
27 |
+
|
28 |
+
More information needed
|
29 |
+
|
30 |
+
## Training and evaluation data
|
31 |
+
|
32 |
+
More information needed
|
33 |
+
|
34 |
+
## Training procedure
|
35 |
+
|
36 |
+
### Training hyperparameters
|
37 |
+
|
38 |
+
The following hyperparameters were used during training:
|
39 |
+
- learning_rate: 4e-05
|
40 |
+
- train_batch_size: 4
|
41 |
+
- eval_batch_size: 8
|
42 |
+
- seed: 4242
|
43 |
+
- gradient_accumulation_steps: 8
|
44 |
+
- total_train_batch_size: 32
|
45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
46 |
+
- lr_scheduler_type: linear
|
47 |
+
- lr_scheduler_warmup_steps: 500
|
48 |
+
- num_epochs: 500
|
49 |
+
- mixed_precision_training: Native AMP
|
50 |
+
|
51 |
+
### Training results
|
52 |
+
|
53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
54 |
+
|:-------------:|:------:|:-----:|:---------------:|:------:|
|
55 |
+
| 745.9947 | 5.55 | 300 | 1354.9309 | 1.0 |
|
56 |
+
| 576.053 | 11.11 | 600 | 1264.9503 | 1.0 |
|
57 |
+
| 506.6065 | 16.66 | 900 | 999.9504 | 0.9999 |
|
58 |
+
| 278.9146 | 22.22 | 1200 | 497.2084 | 0.8253 |
|
59 |
+
| 203.6943 | 27.77 | 1500 | 382.0594 | 0.6598 |
|
60 |
+
| 164.3587 | 33.33 | 1800 | 332.7126 | 0.5789 |
|
61 |
+
| 136.6555 | 38.88 | 2100 | 314.8907 | 0.5433 |
|
62 |
+
| 116.4133 | 44.44 | 2400 | 298.0489 | 0.5105 |
|
63 |
+
| 103.4391 | 49.99 | 2700 | 288.4926 | 0.4862 |
|
64 |
+
| 94.8171 | 55.55 | 3000 | 286.2634 | 0.4675 |
|
65 |
+
| 83.2515 | 61.11 | 3300 | 282.2958 | 0.4596 |
|
66 |
+
| 75.3823 | 66.66 | 3600 | 283.7801 | 0.4481 |
|
67 |
+
| 68.4964 | 72.22 | 3900 | 288.9401 | 0.4460 |
|
68 |
+
| 64.664 | 77.77 | 4200 | 286.0048 | 0.4411 |
|
69 |
+
| 60.8176 | 83.33 | 4500 | 281.5687 | 0.4313 |
|
70 |
+
| 55.4021 | 88.88 | 4800 | 285.4101 | 0.4250 |
|
71 |
+
| 52.1835 | 94.44 | 5100 | 289.8307 | 0.4223 |
|
72 |
+
| 50.4828 | 99.99 | 5400 | 283.2961 | 0.4138 |
|
73 |
+
| 48.3316 | 105.55 | 5700 | 291.5920 | 0.4116 |
|
74 |
+
| 43.5322 | 111.11 | 6000 | 291.0108 | 0.4129 |
|
75 |
+
| 41.9436 | 116.66 | 6300 | 289.5014 | 0.4082 |
|
76 |
+
| 38.9519 | 122.22 | 6600 | 292.2263 | 0.4102 |
|
77 |
+
| 39.0257 | 127.77 | 6900 | 299.3404 | 0.4016 |
|
78 |
+
| 37.7255 | 133.33 | 7200 | 302.4544 | 0.3999 |
|
79 |
+
| 35.9706 | 138.88 | 7500 | 294.0447 | 0.3934 |
|
80 |
+
| 34.1693 | 144.44 | 7800 | 295.0625 | 0.3936 |
|
81 |
+
| 32.4428 | 149.99 | 8100 | 300.1613 | 0.3935 |
|
82 |
+
| 33.4516 | 155.55 | 8400 | 310.9324 | 0.3951 |
|
83 |
+
| 31.039 | 161.11 | 8700 | 313.7309 | 0.3921 |
|
84 |
+
| 29.4764 | 166.66 | 9000 | 307.3336 | 0.3976 |
|
85 |
+
| 29.4936 | 172.22 | 9300 | 304.8356 | 0.3958 |
|
86 |
+
| 27.3962 | 177.77 | 9600 | 309.1968 | 0.3895 |
|
87 |
+
| 28.9029 | 183.33 | 9900 | 313.4612 | 0.3899 |
|
88 |
+
| 26.0951 | 188.88 | 10200 | 310.5530 | 0.3862 |
|
89 |
+
| 26.339 | 194.44 | 10500 | 312.8733 | 0.3850 |
|
90 |
+
| 25.0828 | 199.99 | 10800 | 319.1792 | 0.3904 |
|
91 |
+
| 25.2904 | 205.55 | 11100 | 305.1705 | 0.3896 |
|
92 |
+
| 24.3586 | 211.11 | 11400 | 316.6768 | 0.3895 |
|
93 |
+
| 22.0152 | 216.66 | 11700 | 325.8116 | 0.3890 |
|
94 |
+
| 23.443 | 222.22 | 12000 | 315.7543 | 0.3805 |
|
95 |
+
| 23.6808 | 227.77 | 12300 | 324.8992 | 0.3831 |
|
96 |
+
| 22.6985 | 233.33 | 12600 | 321.0045 | 0.3814 |
|
97 |
+
| 21.3545 | 238.88 | 12900 | 323.3702 | 0.3843 |
|
98 |
+
| 20.9784 | 244.44 | 13200 | 334.0336 | 0.3844 |
|
99 |
+
| 20.1037 | 249.99 | 13500 | 316.2501 | 0.3764 |
|
100 |
+
| 20.2854 | 255.55 | 13800 | 322.2803 | 0.3768 |
|
101 |
+
| 19.4469 | 261.11 | 14100 | 326.5724 | 0.3760 |
|
102 |
+
| 20.1176 | 266.66 | 14400 | 322.7105 | 0.3752 |
|
103 |
+
| 19.1208 | 272.22 | 14700 | 323.2078 | 0.3766 |
|
104 |
+
| 20.024 | 277.77 | 15000 | 317.7683 | 0.3696 |
|
105 |
+
| 18.8389 | 283.33 | 15300 | 323.7159 | 0.3702 |
|
106 |
+
| 19.5328 | 288.88 | 15600 | 324.3281 | 0.3713 |
|
107 |
+
| 18.4362 | 294.44 | 15900 | 329.0278 | 0.3749 |
|
108 |
+
| 16.8288 | 299.99 | 16200 | 327.6952 | 0.3727 |
|
109 |
+
| 17.6299 | 305.55 | 16500 | 334.1232 | 0.3729 |
|
110 |
+
|
111 |
+
|
112 |
+
### Framework versions
|
113 |
+
|
114 |
+
- Transformers 4.16.2
|
115 |
+
- Pytorch 1.10.2+cu102
|
116 |
+
- Datasets 1.18.3
|
117 |
+
- Tokenizers 0.11.0
|