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-large-xls-r-53h-turkish-colab
|
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-large-xls-r-53h-turkish-colab
|
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: 0.4135
|
20 |
+
- Wer: 0.3247
|
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: 32
|
49 |
+
- mixed_precision_training: Native AMP
|
50 |
+
|
51 |
+
### Training results
|
52 |
+
|
53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
54 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
55 |
+
| 9.4875 | 0.92 | 100 | 3.5328 | 1.0 |
|
56 |
+
| 3.1866 | 1.83 | 200 | 3.0955 | 1.0 |
|
57 |
+
| 2.027 | 2.75 | 300 | 0.9002 | 0.7685 |
|
58 |
+
| 0.7285 | 3.67 | 400 | 0.6279 | 0.6693 |
|
59 |
+
| 0.4693 | 4.59 | 500 | 0.5672 | 0.5643 |
|
60 |
+
| 0.3615 | 5.5 | 600 | 0.4995 | 0.5094 |
|
61 |
+
| 0.2846 | 6.42 | 700 | 0.4561 | 0.4797 |
|
62 |
+
| 0.2253 | 7.34 | 800 | 0.4742 | 0.4675 |
|
63 |
+
| 0.2004 | 8.26 | 900 | 0.4462 | 0.4345 |
|
64 |
+
| 0.173 | 9.17 | 1000 | 0.4688 | 0.4333 |
|
65 |
+
| 0.1547 | 10.09 | 1100 | 0.4429 | 0.4206 |
|
66 |
+
| 0.1444 | 11.01 | 1200 | 0.4662 | 0.4144 |
|
67 |
+
| 0.1274 | 11.93 | 1300 | 0.4675 | 0.4213 |
|
68 |
+
| 0.1164 | 12.84 | 1400 | 0.4947 | 0.4073 |
|
69 |
+
| 0.1081 | 13.76 | 1500 | 0.4223 | 0.3915 |
|
70 |
+
| 0.1025 | 14.68 | 1600 | 0.4493 | 0.3912 |
|
71 |
+
| 0.0944 | 15.6 | 1700 | 0.4527 | 0.3848 |
|
72 |
+
| 0.0943 | 16.51 | 1800 | 0.4288 | 0.3810 |
|
73 |
+
| 0.0885 | 17.43 | 1900 | 0.4313 | 0.3670 |
|
74 |
+
| 0.0781 | 18.35 | 2000 | 0.4729 | 0.3790 |
|
75 |
+
| 0.0828 | 19.27 | 2100 | 0.4560 | 0.3651 |
|
76 |
+
| 0.0753 | 20.18 | 2200 | 0.4478 | 0.3599 |
|
77 |
+
| 0.0702 | 21.1 | 2300 | 0.4518 | 0.3595 |
|
78 |
+
| 0.0666 | 22.02 | 2400 | 0.4080 | 0.3489 |
|
79 |
+
| 0.0661 | 22.94 | 2500 | 0.4414 | 0.3507 |
|
80 |
+
| 0.0607 | 23.85 | 2600 | 0.4209 | 0.3538 |
|
81 |
+
| 0.058 | 24.77 | 2700 | 0.4302 | 0.3382 |
|
82 |
+
| 0.0596 | 25.69 | 2800 | 0.3939 | 0.3328 |
|
83 |
+
| 0.052 | 26.61 | 2900 | 0.4374 | 0.3311 |
|
84 |
+
| 0.0473 | 27.52 | 3000 | 0.4406 | 0.3363 |
|
85 |
+
| 0.0483 | 28.44 | 3100 | 0.4272 | 0.3286 |
|
86 |
+
| 0.049 | 29.36 | 3200 | 0.4189 | 0.3257 |
|
87 |
+
| 0.0433 | 30.28 | 3300 | 0.4242 | 0.3229 |
|
88 |
+
| 0.0438 | 31.19 | 3400 | 0.4135 | 0.3247 |
|
89 |
+
|
90 |
+
|
91 |
+
### Framework versions
|
92 |
+
|
93 |
+
- Transformers 4.11.3
|
94 |
+
- Pytorch 1.10.0+cu113
|
95 |
+
- Datasets 1.18.3
|
96 |
+
- Tokenizers 0.10.3
|