update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
name: xlsr-vlbg
|
| 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 |
+
# xlsr-vlbg
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an unknown dataset.
|
| 15 |
+
It achieves the following results on the evaluation set:
|
| 16 |
+
- Loss: 1.0556
|
| 17 |
+
- Wer: 0.3575
|
| 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: 0.0003
|
| 37 |
+
- train_batch_size: 16
|
| 38 |
+
- eval_batch_size: 8
|
| 39 |
+
- seed: 42
|
| 40 |
+
- gradient_accumulation_steps: 2
|
| 41 |
+
- total_train_batch_size: 32
|
| 42 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 43 |
+
- lr_scheduler_type: linear
|
| 44 |
+
- lr_scheduler_warmup_steps: 500
|
| 45 |
+
- num_epochs: 60
|
| 46 |
+
- mixed_precision_training: Native AMP
|
| 47 |
+
|
| 48 |
+
### Training results
|
| 49 |
+
|
| 50 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 51 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 52 |
+
| 10.4898 | 3.56 | 100 | 3.8744 | 1.0 |
|
| 53 |
+
| 3.4178 | 7.14 | 200 | 3.0383 | 1.0 |
|
| 54 |
+
| 2.9644 | 10.7 | 300 | 2.7548 | 1.0 |
|
| 55 |
+
| 2.0967 | 14.28 | 400 | 1.3798 | 0.7385 |
|
| 56 |
+
| 0.9283 | 17.84 | 500 | 0.8976 | 0.5495 |
|
| 57 |
+
| 0.5498 | 21.42 | 600 | 0.9120 | 0.4711 |
|
| 58 |
+
| 0.3561 | 24.98 | 700 | 0.9389 | 0.4505 |
|
| 59 |
+
| 0.2544 | 28.56 | 800 | 0.8814 | 0.4084 |
|
| 60 |
+
| 0.1859 | 32.14 | 900 | 1.0430 | 0.3790 |
|
| 61 |
+
| 0.1517 | 35.7 | 1000 | 0.9847 | 0.3976 |
|
| 62 |
+
| 0.107 | 39.28 | 1100 | 0.9529 | 0.3702 |
|
| 63 |
+
| 0.0863 | 42.84 | 1200 | 1.1043 | 0.3477 |
|
| 64 |
+
| 0.0686 | 46.42 | 1300 | 1.1166 | 0.3732 |
|
| 65 |
+
| 0.0508 | 49.98 | 1400 | 1.0854 | 0.3565 |
|
| 66 |
+
| 0.0425 | 53.56 | 1500 | 1.0830 | 0.3614 |
|
| 67 |
+
| 0.0362 | 57.14 | 1600 | 1.0556 | 0.3575 |
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
### Framework versions
|
| 71 |
+
|
| 72 |
+
- Transformers 4.11.3
|
| 73 |
+
- Pytorch 1.10.0+cu113
|
| 74 |
+
- Datasets 1.18.3
|
| 75 |
+
- Tokenizers 0.10.3
|