Commit
•
594a567
1
Parent(s):
1220b51
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- common_voice
|
| 7 |
+
model-index:
|
| 8 |
+
- name: ''
|
| 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 |
+
#
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-1b](https://huggingface.co/facebook/wav2vec2-xls-r-1b) on the common_voice dataset.
|
| 18 |
+
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.6773
|
| 20 |
+
- Wer: 0.3668
|
| 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: 5e-05
|
| 40 |
+
- train_batch_size: 8
|
| 41 |
+
- eval_batch_size: 16
|
| 42 |
+
- seed: 42
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 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: 1500
|
| 48 |
+
- num_epochs: 50.0
|
| 49 |
+
- mixed_precision_training: Native AMP
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 54 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 55 |
+
| 2.514 | 2.07 | 400 | 1.4589 | 0.8531 |
|
| 56 |
+
| 1.4289 | 4.15 | 800 | 0.8940 | 0.6475 |
|
| 57 |
+
| 1.276 | 6.22 | 1200 | 0.7743 | 0.6089 |
|
| 58 |
+
| 1.2213 | 8.29 | 1600 | 0.6919 | 0.4973 |
|
| 59 |
+
| 1.1522 | 10.36 | 2000 | 0.6635 | 0.4588 |
|
| 60 |
+
| 1.0914 | 12.44 | 2400 | 0.6839 | 0.4586 |
|
| 61 |
+
| 1.0499 | 14.51 | 2800 | 0.7151 | 0.4467 |
|
| 62 |
+
| 1.0238 | 16.58 | 3200 | 0.6824 | 0.4436 |
|
| 63 |
+
| 0.9963 | 18.65 | 3600 | 0.6872 | 0.4437 |
|
| 64 |
+
| 0.9728 | 20.73 | 4000 | 0.7047 | 0.4244 |
|
| 65 |
+
| 0.9373 | 22.8 | 4400 | 0.6569 | 0.4189 |
|
| 66 |
+
| 0.9028 | 24.87 | 4800 | 0.6623 | 0.4094 |
|
| 67 |
+
| 0.8759 | 26.94 | 5200 | 0.6723 | 0.4152 |
|
| 68 |
+
| 0.8824 | 29.02 | 5600 | 0.6467 | 0.4017 |
|
| 69 |
+
| 0.8371 | 31.09 | 6000 | 0.6911 | 0.4080 |
|
| 70 |
+
| 0.8205 | 33.16 | 6400 | 0.7145 | 0.4063 |
|
| 71 |
+
| 0.7837 | 35.23 | 6800 | 0.7037 | 0.3930 |
|
| 72 |
+
| 0.7708 | 37.31 | 7200 | 0.6925 | 0.3840 |
|
| 73 |
+
| 0.7359 | 39.38 | 7600 | 0.7034 | 0.3829 |
|
| 74 |
+
| 0.7153 | 41.45 | 8000 | 0.7030 | 0.3794 |
|
| 75 |
+
| 0.7127 | 43.52 | 8400 | 0.6823 | 0.3761 |
|
| 76 |
+
| 0.6884 | 45.6 | 8800 | 0.6854 | 0.3711 |
|
| 77 |
+
| 0.6835 | 47.67 | 9200 | 0.6723 | 0.3665 |
|
| 78 |
+
| 0.6703 | 49.74 | 9600 | 0.6773 | 0.3668 |
|
| 79 |
+
|
| 80 |
+
|
| 81 |
+
### Framework versions
|
| 82 |
+
|
| 83 |
+
- Transformers 4.17.0.dev0
|
| 84 |
+
- Pytorch 1.10.2+cu102
|
| 85 |
+
- Datasets 1.18.2.dev0
|
| 86 |
+
- Tokenizers 0.11.0
|