update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- generated_from_trainer
|
| 4 |
+
model-index:
|
| 5 |
+
- name: librispeech-semi-supervised-without-LM
|
| 6 |
+
results: []
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 10 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 11 |
+
|
| 12 |
+
# librispeech-semi-supervised-without-LM
|
| 13 |
+
|
| 14 |
+
This model was trained from scratch on the None dataset.
|
| 15 |
+
It achieves the following results on the evaluation set:
|
| 16 |
+
- Loss: 0.1837
|
| 17 |
+
- Wer: 0.0580
|
| 18 |
+
|
| 19 |
+
## Model description
|
| 20 |
+
|
| 21 |
+
More information needed
|
| 22 |
+
|
| 23 |
+
## Intended uses & limitations
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Training and evaluation data
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training procedure
|
| 32 |
+
|
| 33 |
+
### Training hyperparameters
|
| 34 |
+
|
| 35 |
+
The following hyperparameters were used during training:
|
| 36 |
+
- learning_rate: 0.0001
|
| 37 |
+
- train_batch_size: 16
|
| 38 |
+
- eval_batch_size: 8
|
| 39 |
+
- seed: 42
|
| 40 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 41 |
+
- lr_scheduler_type: linear
|
| 42 |
+
- lr_scheduler_warmup_steps: 1000
|
| 43 |
+
- num_epochs: 15
|
| 44 |
+
- mixed_precision_training: Native AMP
|
| 45 |
+
|
| 46 |
+
### Training results
|
| 47 |
+
|
| 48 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 49 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 50 |
+
| 0.0565 | 0.56 | 1000 | 0.1354 | 0.0641 |
|
| 51 |
+
| 0.0548 | 1.12 | 2000 | 0.1320 | 0.0628 |
|
| 52 |
+
| 0.0478 | 1.68 | 3000 | 0.1247 | 0.0612 |
|
| 53 |
+
| 0.0451 | 2.24 | 4000 | 0.1256 | 0.0613 |
|
| 54 |
+
| 0.0401 | 2.8 | 5000 | 0.1269 | 0.0606 |
|
| 55 |
+
| 0.035 | 3.36 | 6000 | 0.1370 | 0.0595 |
|
| 56 |
+
| 0.0344 | 3.92 | 7000 | 0.1280 | 0.0589 |
|
| 57 |
+
| 0.031 | 4.48 | 8000 | 0.1350 | 0.0589 |
|
| 58 |
+
| 0.031 | 5.04 | 9000 | 0.1418 | 0.0614 |
|
| 59 |
+
| 0.0278 | 5.61 | 10000 | 0.1382 | 0.0604 |
|
| 60 |
+
| 0.0272 | 6.17 | 11000 | 0.1502 | 0.0615 |
|
| 61 |
+
| 0.0246 | 6.73 | 12000 | 0.1443 | 0.0609 |
|
| 62 |
+
| 0.0233 | 7.29 | 13000 | 0.1548 | 0.0589 |
|
| 63 |
+
| 0.0224 | 7.85 | 14000 | 0.1547 | 0.0599 |
|
| 64 |
+
| 0.0202 | 8.41 | 15000 | 0.1570 | 0.0590 |
|
| 65 |
+
| 0.0199 | 8.97 | 16000 | 0.1564 | 0.0594 |
|
| 66 |
+
| 0.0186 | 9.53 | 17000 | 0.1598 | 0.0595 |
|
| 67 |
+
| 0.0187 | 10.09 | 18000 | 0.1657 | 0.0585 |
|
| 68 |
+
| 0.017 | 10.65 | 19000 | 0.1690 | 0.0584 |
|
| 69 |
+
| 0.016 | 11.21 | 20000 | 0.1689 | 0.0588 |
|
| 70 |
+
| 0.0156 | 11.77 | 21000 | 0.1745 | 0.0585 |
|
| 71 |
+
| 0.0151 | 12.33 | 22000 | 0.1777 | 0.0583 |
|
| 72 |
+
| 0.0144 | 12.89 | 23000 | 0.1778 | 0.0590 |
|
| 73 |
+
| 0.0142 | 13.45 | 24000 | 0.1803 | 0.0585 |
|
| 74 |
+
| 0.0137 | 14.01 | 25000 | 0.1796 | 0.0581 |
|
| 75 |
+
| 0.0132 | 14.57 | 26000 | 0.1837 | 0.0580 |
|
| 76 |
+
|
| 77 |
+
|
| 78 |
+
### Framework versions
|
| 79 |
+
|
| 80 |
+
- Transformers 4.14.1
|
| 81 |
+
- Pytorch 1.10.2
|
| 82 |
+
- Datasets 1.18.2
|
| 83 |
+
- Tokenizers 0.10.3
|