update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- common_voice
|
7 |
+
model-index:
|
8 |
+
- name: wav2vec2-common_voice-tr-demo-dist
|
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-common_voice-tr-demo-dist
|
16 |
+
|
17 |
+
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the common_voice dataset.
|
18 |
+
It achieves the following results on the evaluation set:
|
19 |
+
- Loss: 0.3893
|
20 |
+
- Wer: 0.3238
|
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: 4
|
41 |
+
- eval_batch_size: 8
|
42 |
+
- seed: 42
|
43 |
+
- distributed_type: multi-GPU
|
44 |
+
- num_devices: 4
|
45 |
+
- total_train_batch_size: 16
|
46 |
+
- total_eval_batch_size: 32
|
47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
48 |
+
- lr_scheduler_type: linear
|
49 |
+
- lr_scheduler_warmup_steps: 500
|
50 |
+
- num_epochs: 15.0
|
51 |
+
- mixed_precision_training: Native AMP
|
52 |
+
|
53 |
+
### Training results
|
54 |
+
|
55 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
56 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
57 |
+
| 3.5279 | 0.46 | 100 | 3.6260 | 1.0 |
|
58 |
+
| 3.1065 | 0.92 | 200 | 3.0854 | 0.9999 |
|
59 |
+
| 1.4111 | 1.38 | 300 | 1.3343 | 0.8839 |
|
60 |
+
| 0.8468 | 1.83 | 400 | 0.6920 | 0.6826 |
|
61 |
+
| 0.6242 | 2.29 | 500 | 0.6001 | 0.5996 |
|
62 |
+
| 0.4181 | 2.75 | 600 | 0.5655 | 0.5680 |
|
63 |
+
| 0.4311 | 3.21 | 700 | 0.4478 | 0.5003 |
|
64 |
+
| 0.3601 | 3.67 | 800 | 0.4548 | 0.5011 |
|
65 |
+
| 0.2756 | 4.13 | 900 | 0.4444 | 0.4682 |
|
66 |
+
| 0.2373 | 4.59 | 1000 | 0.4111 | 0.4432 |
|
67 |
+
| 0.1831 | 5.05 | 1100 | 0.4178 | 0.4447 |
|
68 |
+
| 0.2423 | 5.5 | 1200 | 0.3881 | 0.4277 |
|
69 |
+
| 0.2128 | 5.96 | 1300 | 0.3865 | 0.4018 |
|
70 |
+
| 0.1256 | 6.42 | 1400 | 0.3818 | 0.4137 |
|
71 |
+
| 0.1038 | 6.88 | 1500 | 0.3739 | 0.3942 |
|
72 |
+
| 0.1662 | 7.34 | 1600 | 0.3938 | 0.3929 |
|
73 |
+
| 0.198 | 7.8 | 1700 | 0.3831 | 0.3837 |
|
74 |
+
| 0.0728 | 8.26 | 1800 | 0.3910 | 0.3867 |
|
75 |
+
| 0.123 | 8.72 | 1900 | 0.3722 | 0.3735 |
|
76 |
+
| 0.0776 | 9.17 | 2000 | 0.3938 | 0.3725 |
|
77 |
+
| 0.1597 | 9.63 | 2100 | 0.3786 | 0.3697 |
|
78 |
+
| 0.1124 | 10.09 | 2200 | 0.3947 | 0.3590 |
|
79 |
+
| 0.0965 | 10.55 | 2300 | 0.3952 | 0.3562 |
|
80 |
+
| 0.0612 | 11.01 | 2400 | 0.3810 | 0.3476 |
|
81 |
+
| 0.0764 | 11.47 | 2500 | 0.3734 | 0.3507 |
|
82 |
+
| 0.0973 | 11.93 | 2600 | 0.3935 | 0.3472 |
|
83 |
+
| 0.0649 | 12.39 | 2700 | 0.3672 | 0.3413 |
|
84 |
+
| 0.0542 | 12.84 | 2800 | 0.3732 | 0.3369 |
|
85 |
+
| 0.087 | 13.3 | 2900 | 0.3833 | 0.3458 |
|
86 |
+
| 0.0196 | 13.76 | 3000 | 0.3761 | 0.3303 |
|
87 |
+
| 0.0548 | 14.22 | 3100 | 0.3855 | 0.3274 |
|
88 |
+
| 0.0577 | 14.68 | 3200 | 0.3893 | 0.3238 |
|
89 |
+
|
90 |
+
|
91 |
+
### Framework versions
|
92 |
+
|
93 |
+
- Transformers 4.20.0.dev0
|
94 |
+
- Pytorch 1.11.0+cu102
|
95 |
+
- Datasets 2.2.1
|
96 |
+
- Tokenizers 0.12.1
|