update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,112 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
datasets:
|
6 |
+
- common_voice
|
7 |
+
model-index:
|
8 |
+
- name: wav2vec2-xls-r-300m-gn-cv8-3
|
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-xls-r-300m-gn-cv8-3
|
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 common_voice dataset.
|
18 |
+
It achieves the following results on the evaluation set:
|
19 |
+
- Loss: 0.9517
|
20 |
+
- Wer: 0.8542
|
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: 0.0001
|
40 |
+
- train_batch_size: 8
|
41 |
+
- eval_batch_size: 8
|
42 |
+
- seed: 42
|
43 |
+
- gradient_accumulation_steps: 2
|
44 |
+
- total_train_batch_size: 16
|
45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
46 |
+
- lr_scheduler_type: linear
|
47 |
+
- lr_scheduler_warmup_steps: 100
|
48 |
+
- training_steps: 5000
|
49 |
+
- mixed_precision_training: Native AMP
|
50 |
+
|
51 |
+
### Training results
|
52 |
+
|
53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
54 |
+
|:-------------:|:------:|:----:|:---------------:|:------:|
|
55 |
+
| 19.9125 | 5.54 | 100 | 5.4279 | 1.0 |
|
56 |
+
| 3.8031 | 11.11 | 200 | 3.3070 | 1.0 |
|
57 |
+
| 3.3783 | 16.65 | 300 | 3.2450 | 1.0 |
|
58 |
+
| 3.3472 | 22.22 | 400 | 3.2424 | 1.0 |
|
59 |
+
| 3.2714 | 27.76 | 500 | 3.1100 | 1.0 |
|
60 |
+
| 3.2367 | 33.32 | 600 | 3.1091 | 1.0 |
|
61 |
+
| 3.1968 | 38.86 | 700 | 3.1013 | 1.0 |
|
62 |
+
| 3.2004 | 44.43 | 800 | 3.1173 | 1.0 |
|
63 |
+
| 3.1656 | 49.97 | 900 | 3.0682 | 1.0 |
|
64 |
+
| 3.1563 | 55.54 | 1000 | 3.0457 | 1.0 |
|
65 |
+
| 3.1356 | 61.11 | 1100 | 3.0139 | 1.0 |
|
66 |
+
| 3.086 | 66.65 | 1200 | 2.8108 | 1.0 |
|
67 |
+
| 2.954 | 72.22 | 1300 | 2.3238 | 1.0 |
|
68 |
+
| 2.6125 | 77.76 | 1400 | 1.6461 | 1.0 |
|
69 |
+
| 2.3296 | 83.32 | 1500 | 1.2834 | 0.9744 |
|
70 |
+
| 2.1345 | 88.86 | 1600 | 1.1091 | 0.9693 |
|
71 |
+
| 2.0346 | 94.43 | 1700 | 1.0273 | 0.9233 |
|
72 |
+
| 1.9611 | 99.97 | 1800 | 0.9642 | 0.9182 |
|
73 |
+
| 1.9066 | 105.54 | 1900 | 0.9590 | 0.9105 |
|
74 |
+
| 1.8178 | 111.11 | 2000 | 0.9679 | 0.9028 |
|
75 |
+
| 1.7799 | 116.65 | 2100 | 0.9007 | 0.8619 |
|
76 |
+
| 1.7726 | 122.22 | 2200 | 0.9689 | 0.8951 |
|
77 |
+
| 1.7389 | 127.76 | 2300 | 0.8876 | 0.8593 |
|
78 |
+
| 1.7151 | 133.32 | 2400 | 0.8716 | 0.8542 |
|
79 |
+
| 1.6842 | 138.86 | 2500 | 0.9536 | 0.8772 |
|
80 |
+
| 1.6449 | 144.43 | 2600 | 0.9296 | 0.8542 |
|
81 |
+
| 1.5978 | 149.97 | 2700 | 0.8895 | 0.8440 |
|
82 |
+
| 1.6515 | 155.54 | 2800 | 0.9162 | 0.8568 |
|
83 |
+
| 1.6586 | 161.11 | 2900 | 0.9039 | 0.8568 |
|
84 |
+
| 1.5966 | 166.65 | 3000 | 0.8627 | 0.8542 |
|
85 |
+
| 1.5695 | 172.22 | 3100 | 0.9549 | 0.8824 |
|
86 |
+
| 1.5699 | 177.76 | 3200 | 0.9332 | 0.8517 |
|
87 |
+
| 1.5297 | 183.32 | 3300 | 0.9163 | 0.8338 |
|
88 |
+
| 1.5367 | 188.86 | 3400 | 0.8822 | 0.8312 |
|
89 |
+
| 1.5586 | 194.43 | 3500 | 0.9217 | 0.8363 |
|
90 |
+
| 1.5429 | 199.97 | 3600 | 0.9564 | 0.8568 |
|
91 |
+
| 1.5273 | 205.54 | 3700 | 0.9508 | 0.8542 |
|
92 |
+
| 1.5043 | 211.11 | 3800 | 0.9374 | 0.8542 |
|
93 |
+
| 1.4724 | 216.65 | 3900 | 0.9622 | 0.8619 |
|
94 |
+
| 1.4794 | 222.22 | 4000 | 0.9550 | 0.8363 |
|
95 |
+
| 1.4843 | 227.76 | 4100 | 0.9577 | 0.8465 |
|
96 |
+
| 1.4781 | 233.32 | 4200 | 0.9543 | 0.8440 |
|
97 |
+
| 1.4507 | 238.86 | 4300 | 0.9553 | 0.8491 |
|
98 |
+
| 1.4997 | 244.43 | 4400 | 0.9728 | 0.8491 |
|
99 |
+
| 1.4371 | 249.97 | 4500 | 0.9543 | 0.8670 |
|
100 |
+
| 1.4825 | 255.54 | 4600 | 0.9636 | 0.8619 |
|
101 |
+
| 1.4187 | 261.11 | 4700 | 0.9609 | 0.8440 |
|
102 |
+
| 1.4363 | 266.65 | 4800 | 0.9567 | 0.8593 |
|
103 |
+
| 1.4463 | 272.22 | 4900 | 0.9581 | 0.8542 |
|
104 |
+
| 1.4117 | 277.76 | 5000 | 0.9517 | 0.8542 |
|
105 |
+
|
106 |
+
|
107 |
+
### Framework versions
|
108 |
+
|
109 |
+
- Transformers 4.16.0
|
110 |
+
- Pytorch 1.10.0+cu111
|
111 |
+
- Datasets 1.18.1
|
112 |
+
- Tokenizers 0.11.0
|