update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- vivos_dataset
|
7 |
+
model-index:
|
8 |
+
- name: wav2vec2-dataset-vios
|
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 |
+
# wav2vec2-dataset-vios
|
16 |
+
|
17 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the vivos_dataset dataset.
|
18 |
+
It achieves the following results on the evaluation set:
|
19 |
+
- Loss: 0.5423
|
20 |
+
- Wer: 0.4075
|
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: 0.0003
|
40 |
+
- train_batch_size: 16
|
41 |
+
- eval_batch_size: 8
|
42 |
+
- seed: 42
|
43 |
+
- gradient_accumulation_steps: 2
|
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: 30
|
49 |
+
- mixed_precision_training: Native AMP
|
50 |
+
|
51 |
+
### Training results
|
52 |
+
|
53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
54 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
55 |
+
| 5.0963 | 1.1 | 400 | 1.1336 | 0.7374 |
|
56 |
+
| 0.6576 | 2.2 | 800 | 0.4716 | 0.3727 |
|
57 |
+
| 0.4099 | 3.3 | 1200 | 0.3907 | 0.3100 |
|
58 |
+
| 0.3332 | 4.4 | 1600 | 0.3735 | 0.2766 |
|
59 |
+
| 0.2976 | 5.49 | 2000 | 0.3932 | 0.2801 |
|
60 |
+
| 0.2645 | 6.59 | 2400 | 0.3628 | 0.2542 |
|
61 |
+
| 0.2395 | 7.69 | 2800 | 0.3702 | 0.2734 |
|
62 |
+
| 0.2208 | 8.79 | 3200 | 0.3667 | 0.2467 |
|
63 |
+
| 0.1974 | 9.89 | 3600 | 0.3688 | 0.2398 |
|
64 |
+
| 0.1772 | 10.99 | 4000 | 0.3819 | 0.2457 |
|
65 |
+
| 0.1695 | 12.09 | 4400 | 0.3840 | 0.2451 |
|
66 |
+
| 0.319 | 13.19 | 4800 | 0.6531 | 0.4084 |
|
67 |
+
| 0.7305 | 14.29 | 5200 | 0.9883 | 0.6348 |
|
68 |
+
| 0.5787 | 15.38 | 5600 | 0.5260 | 0.3063 |
|
69 |
+
| 0.8558 | 16.48 | 6000 | 1.2870 | 0.7692 |
|
70 |
+
| 1.155 | 17.58 | 6400 | 1.0568 | 0.6353 |
|
71 |
+
| 0.8393 | 18.68 | 6800 | 0.7360 | 0.4486 |
|
72 |
+
| 0.6094 | 19.78 | 7200 | 0.6072 | 0.4108 |
|
73 |
+
| 0.5346 | 20.88 | 7600 | 0.5749 | 0.4095 |
|
74 |
+
| 0.5073 | 21.98 | 8000 | 0.5588 | 0.4056 |
|
75 |
+
| 0.4859 | 23.08 | 8400 | 0.5475 | 0.4015 |
|
76 |
+
| 0.475 | 24.18 | 8800 | 0.5430 | 0.4011 |
|
77 |
+
| 0.4683 | 25.27 | 9200 | 0.5400 | 0.3990 |
|
78 |
+
| 0.4673 | 26.37 | 9600 | 0.5407 | 0.4011 |
|
79 |
+
| 0.4665 | 27.47 | 10000 | 0.5408 | 0.3992 |
|
80 |
+
| 0.4703 | 28.57 | 10400 | 0.5420 | 0.4070 |
|
81 |
+
| 0.4709 | 29.67 | 10800 | 0.5423 | 0.4075 |
|
82 |
+
|
83 |
+
|
84 |
+
### Framework versions
|
85 |
+
|
86 |
+
- Transformers 4.11.3
|
87 |
+
- Pytorch 1.10.0+cu113
|
88 |
+
- Datasets 1.18.3
|
89 |
+
- Tokenizers 0.10.3
|