Commit
•
de36f49
1
Parent(s):
e6ef0cc
update model card README.md
Browse files
README.md
ADDED
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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-300m-tamil-colab-final
|
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-300m-tamil-colab-final
|
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: 0.7539
|
20 |
+
- Wer: 0.6135
|
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 |
+
| 11.1466 | 1.0 | 118 | 4.3444 | 1.0 |
|
56 |
+
| 3.4188 | 2.0 | 236 | 3.2496 | 1.0 |
|
57 |
+
| 2.8617 | 3.0 | 354 | 1.6165 | 1.0003 |
|
58 |
+
| 0.958 | 4.0 | 472 | 0.7984 | 0.8720 |
|
59 |
+
| 0.5929 | 5.0 | 590 | 0.6733 | 0.7831 |
|
60 |
+
| 0.4628 | 6.0 | 708 | 0.6536 | 0.7621 |
|
61 |
+
| 0.3834 | 7.0 | 826 | 0.6037 | 0.7155 |
|
62 |
+
| 0.3242 | 8.0 | 944 | 0.6376 | 0.7184 |
|
63 |
+
| 0.2736 | 9.0 | 1062 | 0.6214 | 0.7070 |
|
64 |
+
| 0.2433 | 10.0 | 1180 | 0.6158 | 0.6944 |
|
65 |
+
| 0.2217 | 11.0 | 1298 | 0.6548 | 0.6830 |
|
66 |
+
| 0.1992 | 12.0 | 1416 | 0.6331 | 0.6775 |
|
67 |
+
| 0.1804 | 13.0 | 1534 | 0.6644 | 0.6874 |
|
68 |
+
| 0.1639 | 14.0 | 1652 | 0.6629 | 0.6649 |
|
69 |
+
| 0.143 | 15.0 | 1770 | 0.6927 | 0.6836 |
|
70 |
+
| 0.1394 | 16.0 | 1888 | 0.6933 | 0.6888 |
|
71 |
+
| 0.1296 | 17.0 | 2006 | 0.7039 | 0.6860 |
|
72 |
+
| 0.1212 | 18.0 | 2124 | 0.7042 | 0.6628 |
|
73 |
+
| 0.1121 | 19.0 | 2242 | 0.7132 | 0.6475 |
|
74 |
+
| 0.1069 | 20.0 | 2360 | 0.7423 | 0.6438 |
|
75 |
+
| 0.1063 | 21.0 | 2478 | 0.7171 | 0.6484 |
|
76 |
+
| 0.1025 | 22.0 | 2596 | 0.7396 | 0.6451 |
|
77 |
+
| 0.0946 | 23.0 | 2714 | 0.7400 | 0.6432 |
|
78 |
+
| 0.0902 | 24.0 | 2832 | 0.7385 | 0.6286 |
|
79 |
+
| 0.0828 | 25.0 | 2950 | 0.7368 | 0.6286 |
|
80 |
+
| 0.079 | 26.0 | 3068 | 0.7471 | 0.6306 |
|
81 |
+
| 0.0747 | 27.0 | 3186 | 0.7524 | 0.6201 |
|
82 |
+
| 0.0661 | 28.0 | 3304 | 0.7576 | 0.6201 |
|
83 |
+
| 0.0659 | 29.0 | 3422 | 0.7579 | 0.6130 |
|
84 |
+
| 0.0661 | 30.0 | 3540 | 0.7539 | 0.6135 |
|
85 |
+
|
86 |
+
|
87 |
+
### Framework versions
|
88 |
+
|
89 |
+
- Transformers 4.11.3
|
90 |
+
- Pytorch 1.10.0+cu111
|
91 |
+
- Datasets 1.13.3
|
92 |
+
- Tokenizers 0.10.3
|