update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
tags:
|
4 |
+
- generated_from_trainer
|
5 |
+
metrics:
|
6 |
+
- wer
|
7 |
+
model-index:
|
8 |
+
- name: wav2vec2-xls-r-300m-th-v2
|
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-th-v2
|
16 |
+
|
17 |
+
This model is a fine-tuned version of [Botnoi/wav2vec2-xls-r-300m-th-v1](https://huggingface.co/Botnoi/wav2vec2-xls-r-300m-th-v1) on the None dataset.
|
18 |
+
It achieves the following results on the evaluation set:
|
19 |
+
- Loss: 0.3630
|
20 |
+
- Wer: 0.3962
|
21 |
+
- Cer: 0.0942
|
22 |
+
- Clean Cer: 0.0767
|
23 |
+
|
24 |
+
## Model description
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Intended uses & limitations
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training and evaluation data
|
33 |
+
|
34 |
+
More information needed
|
35 |
+
|
36 |
+
## Training procedure
|
37 |
+
|
38 |
+
### Training hyperparameters
|
39 |
+
|
40 |
+
The following hyperparameters were used during training:
|
41 |
+
- learning_rate: 4.533e-08
|
42 |
+
- train_batch_size: 16
|
43 |
+
- eval_batch_size: 16
|
44 |
+
- seed: 42
|
45 |
+
- gradient_accumulation_steps: 2
|
46 |
+
- total_train_batch_size: 32
|
47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
48 |
+
- lr_scheduler_type: linear
|
49 |
+
- training_steps: 9000
|
50 |
+
- mixed_precision_training: Native AMP
|
51 |
+
|
52 |
+
### Training results
|
53 |
+
|
54 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | Clean Cer |
|
55 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:---------:|
|
56 |
+
| 0.3323 | 0.68 | 1000 | 0.3635 | 0.3961 | 0.0942 | 0.0767 |
|
57 |
+
| 0.3386 | 1.36 | 2000 | 0.3632 | 0.3962 | 0.0943 | 0.0768 |
|
58 |
+
| 0.3453 | 2.03 | 3000 | 0.3632 | 0.3964 | 0.0943 | 0.0768 |
|
59 |
+
| 0.3392 | 2.71 | 4000 | 0.3632 | 0.3961 | 0.0943 | 0.0767 |
|
60 |
+
| 0.3399 | 3.39 | 5000 | 0.3634 | 0.3961 | 0.0942 | 0.0768 |
|
61 |
+
| 0.347 | 4.07 | 6000 | 0.3632 | 0.3961 | 0.0942 | 0.0767 |
|
62 |
+
| 0.3414 | 4.74 | 7000 | 0.3631 | 0.3962 | 0.0942 | 0.0767 |
|
63 |
+
| 0.3378 | 5.42 | 8000 | 0.3631 | 0.3961 | 0.0942 | 0.0767 |
|
64 |
+
| 0.3421 | 6.1 | 9000 | 0.3630 | 0.3962 | 0.0942 | 0.0767 |
|
65 |
+
|
66 |
+
|
67 |
+
### Framework versions
|
68 |
+
|
69 |
+
- Transformers 4.26.0.dev0
|
70 |
+
- Pytorch 1.13.0+cu116
|
71 |
+
- Datasets 2.8.0
|
72 |
+
- Tokenizers 0.13.2
|