update model card README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,82 @@
|
|
| 1 |
-
---
|
| 2 |
-
license:
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: wav2vec2-xlsr-300M-NPSC-OH
|
| 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 |
+
# wav2vec2-xlsr-300M-NPSC-OH
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the None dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 0.1700
|
| 18 |
+
- Wer: 0.1665
|
| 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: 7.5e-05
|
| 38 |
+
- train_batch_size: 16
|
| 39 |
+
- eval_batch_size: 16
|
| 40 |
+
- seed: 13
|
| 41 |
+
- gradient_accumulation_steps: 4
|
| 42 |
+
- total_train_batch_size: 64
|
| 43 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
+
- lr_scheduler_type: linear
|
| 45 |
+
- lr_scheduler_warmup_steps: 2000
|
| 46 |
+
- num_epochs: 15.0
|
| 47 |
+
- mixed_precision_training: Native AMP
|
| 48 |
+
|
| 49 |
+
### Training results
|
| 50 |
+
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 52 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 53 |
+
| 3.1638 | 0.66 | 500 | 3.0686 | 1.0 |
|
| 54 |
+
| 2.9311 | 1.31 | 1000 | 2.9208 | 1.0 |
|
| 55 |
+
| 2.4175 | 1.97 | 1500 | 1.5009 | 0.9049 |
|
| 56 |
+
| 1.4442 | 2.63 | 2000 | 0.4426 | 0.3783 |
|
| 57 |
+
| 1.2624 | 3.28 | 2500 | 0.3193 | 0.2998 |
|
| 58 |
+
| 1.1889 | 3.94 | 3000 | 0.2867 | 0.2630 |
|
| 59 |
+
| 1.1315 | 4.6 | 3500 | 0.2566 | 0.2444 |
|
| 60 |
+
| 1.0864 | 5.26 | 4000 | 0.2368 | 0.2294 |
|
| 61 |
+
| 1.093 | 5.91 | 4500 | 0.2240 | 0.2151 |
|
| 62 |
+
| 1.0368 | 6.57 | 5000 | 0.2117 | 0.2056 |
|
| 63 |
+
| 1.0178 | 7.23 | 5500 | 0.2020 | 0.1954 |
|
| 64 |
+
| 1.0035 | 7.88 | 6000 | 0.2005 | 0.1924 |
|
| 65 |
+
| 0.9759 | 8.54 | 6500 | 0.1971 | 0.1863 |
|
| 66 |
+
| 0.9795 | 9.2 | 7000 | 0.1892 | 0.1812 |
|
| 67 |
+
| 0.9601 | 9.85 | 7500 | 0.1863 | 0.1795 |
|
| 68 |
+
| 0.9673 | 10.51 | 8000 | 0.1809 | 0.1761 |
|
| 69 |
+
| 0.9233 | 11.17 | 8500 | 0.1818 | 0.1755 |
|
| 70 |
+
| 0.9382 | 11.83 | 9000 | 0.1767 | 0.1741 |
|
| 71 |
+
| 0.9242 | 12.48 | 9500 | 0.1743 | 0.1703 |
|
| 72 |
+
| 0.9703 | 13.14 | 10000 | 0.1711 | 0.1711 |
|
| 73 |
+
| 0.9139 | 13.8 | 10500 | 0.1718 | 0.1672 |
|
| 74 |
+
| 0.9073 | 14.45 | 11000 | 0.1700 | 0.1665 |
|
| 75 |
+
|
| 76 |
+
|
| 77 |
+
### Framework versions
|
| 78 |
+
|
| 79 |
+
- Transformers 4.17.0.dev0
|
| 80 |
+
- Pytorch 1.10.1+cu102
|
| 81 |
+
- Datasets 1.18.2.dev0
|
| 82 |
+
- Tokenizers 0.11.0
|