update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: wsj0-5percent-supervised
|
| 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 |
+
# wsj0-5percent-supervised
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [facebook/wav2vec2-large-lv60](https://huggingface.co/facebook/wav2vec2-large-lv60) on the None dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 0.3883
|
| 18 |
+
- Wer: 0.1555
|
| 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: 12
|
| 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: 300
|
| 45 |
+
- mixed_precision_training: Native AMP
|
| 46 |
+
|
| 47 |
+
### Training results
|
| 48 |
+
|
| 49 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 50 |
+
|:-------------:|:------:|:----:|:---------------:|:------:|
|
| 51 |
+
| 6.0248 | 16.67 | 500 | 2.9406 | 1.0 |
|
| 52 |
+
| 2.0466 | 33.33 | 1000 | 0.3935 | 0.3300 |
|
| 53 |
+
| 0.1486 | 50.0 | 1500 | 0.3091 | 0.1931 |
|
| 54 |
+
| 0.052 | 66.67 | 2000 | 0.3562 | 0.2052 |
|
| 55 |
+
| 0.0309 | 83.33 | 2500 | 0.3252 | 0.1773 |
|
| 56 |
+
| 0.0228 | 100.0 | 3000 | 0.3360 | 0.1652 |
|
| 57 |
+
| 0.0177 | 116.67 | 3500 | 0.3423 | 0.1603 |
|
| 58 |
+
| 0.0142 | 133.33 | 4000 | 0.3416 | 0.1611 |
|
| 59 |
+
| 0.0119 | 150.0 | 4500 | 0.3663 | 0.1583 |
|
| 60 |
+
| 0.0094 | 166.67 | 5000 | 0.3617 | 0.1567 |
|
| 61 |
+
| 0.0093 | 183.33 | 5500 | 0.3738 | 0.1668 |
|
| 62 |
+
| 0.0079 | 200.0 | 6000 | 0.3881 | 0.1652 |
|
| 63 |
+
| 0.0065 | 216.67 | 6500 | 0.3752 | 0.1611 |
|
| 64 |
+
| 0.0056 | 233.33 | 7000 | 0.3798 | 0.1603 |
|
| 65 |
+
| 0.0057 | 250.0 | 7500 | 0.3944 | 0.1624 |
|
| 66 |
+
| 0.0047 | 266.67 | 8000 | 0.4038 | 0.1583 |
|
| 67 |
+
| 0.0041 | 283.33 | 8500 | 0.3928 | 0.1547 |
|
| 68 |
+
| 0.0036 | 300.0 | 9000 | 0.3883 | 0.1555 |
|
| 69 |
+
|
| 70 |
+
|
| 71 |
+
### Framework versions
|
| 72 |
+
|
| 73 |
+
- Transformers 4.14.1
|
| 74 |
+
- Pytorch 1.10.2
|
| 75 |
+
- Datasets 1.18.2
|
| 76 |
+
- Tokenizers 0.10.3
|