danieleV9H
commited on
Commit
•
070bf05
1
Parent(s):
85c09bf
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 |
+
model-index:
|
6 |
+
- name: hubert-base-timit-demo-google-colab-ft30ep_v5
|
7 |
+
results: []
|
8 |
+
---
|
9 |
+
|
10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
11 |
+
should probably proofread and complete it, then remove this comment. -->
|
12 |
+
|
13 |
+
# hubert-base-timit-demo-google-colab-ft30ep_v5
|
14 |
+
|
15 |
+
This model is a fine-tuned version of [facebook/hubert-base-ls960](https://huggingface.co/facebook/hubert-base-ls960) on the None dataset.
|
16 |
+
It achieves the following results on the evaluation set:
|
17 |
+
- Loss: 0.4763
|
18 |
+
- Wer: 0.3322
|
19 |
+
|
20 |
+
## Model description
|
21 |
+
|
22 |
+
More information needed
|
23 |
+
|
24 |
+
## Intended uses & limitations
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Training and evaluation data
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training procedure
|
33 |
+
|
34 |
+
### Training hyperparameters
|
35 |
+
|
36 |
+
The following hyperparameters were used during training:
|
37 |
+
- learning_rate: 0.0001
|
38 |
+
- train_batch_size: 8
|
39 |
+
- eval_batch_size: 8
|
40 |
+
- seed: 42
|
41 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
42 |
+
- lr_scheduler_type: linear
|
43 |
+
- lr_scheduler_warmup_steps: 1000
|
44 |
+
- num_epochs: 30
|
45 |
+
- mixed_precision_training: Native AMP
|
46 |
+
|
47 |
+
### Training results
|
48 |
+
|
49 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
50 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
51 |
+
| 3.9596 | 0.87 | 500 | 3.1237 | 1.0 |
|
52 |
+
| 2.5388 | 1.73 | 1000 | 1.1689 | 0.9184 |
|
53 |
+
| 1.0448 | 2.6 | 1500 | 0.6106 | 0.5878 |
|
54 |
+
| 0.6793 | 3.46 | 2000 | 0.4912 | 0.5200 |
|
55 |
+
| 0.5234 | 4.33 | 2500 | 0.4529 | 0.4798 |
|
56 |
+
| 0.4368 | 5.19 | 3000 | 0.4239 | 0.4543 |
|
57 |
+
| 0.3839 | 6.06 | 3500 | 0.4326 | 0.4339 |
|
58 |
+
| 0.3315 | 6.92 | 4000 | 0.4265 | 0.4173 |
|
59 |
+
| 0.2878 | 7.79 | 4500 | 0.4304 | 0.4068 |
|
60 |
+
| 0.25 | 8.65 | 5000 | 0.4130 | 0.3940 |
|
61 |
+
| 0.242 | 9.52 | 5500 | 0.4310 | 0.3938 |
|
62 |
+
| 0.2182 | 10.38 | 6000 | 0.4204 | 0.3843 |
|
63 |
+
| 0.2063 | 11.25 | 6500 | 0.4449 | 0.3816 |
|
64 |
+
| 0.2099 | 12.11 | 7000 | 0.4016 | 0.3681 |
|
65 |
+
| 0.1795 | 12.98 | 7500 | 0.4027 | 0.3647 |
|
66 |
+
| 0.1604 | 13.84 | 8000 | 0.4294 | 0.3664 |
|
67 |
+
| 0.1683 | 14.71 | 8500 | 0.4412 | 0.3661 |
|
68 |
+
| 0.1452 | 15.57 | 9000 | 0.4484 | 0.3588 |
|
69 |
+
| 0.1491 | 16.44 | 9500 | 0.4508 | 0.3515 |
|
70 |
+
| 0.1388 | 17.3 | 10000 | 0.4240 | 0.3518 |
|
71 |
+
| 0.1399 | 18.17 | 10500 | 0.4605 | 0.3513 |
|
72 |
+
| 0.1265 | 19.03 | 11000 | 0.4412 | 0.3485 |
|
73 |
+
| 0.1137 | 19.9 | 11500 | 0.4520 | 0.3467 |
|
74 |
+
| 0.106 | 20.76 | 12000 | 0.4873 | 0.3426 |
|
75 |
+
| 0.1243 | 21.63 | 12500 | 0.4456 | 0.3396 |
|
76 |
+
| 0.1055 | 22.49 | 13000 | 0.4819 | 0.3406 |
|
77 |
+
| 0.1124 | 23.36 | 13500 | 0.4613 | 0.3391 |
|
78 |
+
| 0.1064 | 24.22 | 14000 | 0.4842 | 0.3430 |
|
79 |
+
| 0.0875 | 25.09 | 14500 | 0.4661 | 0.3348 |
|
80 |
+
| 0.086 | 25.95 | 15000 | 0.4724 | 0.3371 |
|
81 |
+
| 0.0842 | 26.82 | 15500 | 0.4982 | 0.3381 |
|
82 |
+
| 0.0834 | 27.68 | 16000 | 0.4856 | 0.3337 |
|
83 |
+
| 0.0918 | 28.55 | 16500 | 0.4783 | 0.3344 |
|
84 |
+
| 0.0773 | 29.41 | 17000 | 0.4763 | 0.3322 |
|
85 |
+
|
86 |
+
|
87 |
+
### Framework versions
|
88 |
+
|
89 |
+
- Transformers 4.17.0
|
90 |
+
- Pytorch 1.11.0+cu113
|
91 |
+
- Datasets 1.18.3
|
92 |
+
- Tokenizers 0.12.1
|