File size: 3,137 Bytes
145b24a ed087b1 145b24a ed087b1 145b24a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 |
---
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-2b
tags:
- automatic-speech-recognition
- DewiBrynJones/banc-trawsgrifiadau-bangor-normalized
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-xls-r-2b-ft-btb-cy
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# wav2vec2-xls-r-2b-ft-btb-cy
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-2b](https://huggingface.co/facebook/wav2vec2-xls-r-2b) on the DEWIBRYNJONES/BANC-TRAWSGRIFIADAU-BANGOR-NORMALIZED - DEFAULT dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3903
- Wer: 0.2957
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 2500
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:------:|:----:|:---------------:|:------:|
| No log | 0.1414 | 100 | 1.2105 | 0.8709 |
| No log | 0.2829 | 200 | 0.9787 | 0.6986 |
| No log | 0.4243 | 300 | 1.1907 | 0.7127 |
| No log | 0.5658 | 400 | 1.0559 | 0.7169 |
| 1.4456 | 0.7072 | 500 | 1.2106 | 0.7944 |
| 1.4456 | 0.8487 | 600 | 1.0232 | 0.7033 |
| 1.4456 | 0.9901 | 700 | 1.0387 | 0.7336 |
| 1.4456 | 1.1315 | 800 | 0.7234 | 0.5223 |
| 1.4456 | 1.2730 | 900 | 0.7242 | 0.5566 |
| 0.9155 | 1.4144 | 1000 | 0.7097 | 0.5259 |
| 0.9155 | 1.5559 | 1100 | 0.6368 | 0.4797 |
| 0.9155 | 1.6973 | 1200 | 0.6065 | 0.4653 |
| 0.9155 | 1.8388 | 1300 | 0.6207 | 0.4717 |
| 0.9155 | 1.9802 | 1400 | 0.5925 | 0.4707 |
| 0.7436 | 2.1216 | 1500 | 0.5382 | 0.4046 |
| 0.7436 | 2.2631 | 1600 | 0.5201 | 0.3996 |
| 0.7436 | 2.4045 | 1700 | 0.4883 | 0.3698 |
| 0.7436 | 2.5460 | 1800 | 0.4704 | 0.3659 |
| 0.7436 | 2.6874 | 1900 | 0.4443 | 0.3521 |
| 0.5645 | 2.8289 | 2000 | 0.4470 | 0.3476 |
| 0.5645 | 2.9703 | 2100 | 0.4192 | 0.3242 |
| 0.5645 | 3.1117 | 2200 | 0.4178 | 0.3161 |
| 0.5645 | 3.2532 | 2300 | 0.4122 | 0.3054 |
| 0.5645 | 3.3946 | 2400 | 0.3960 | 0.2990 |
| 0.4232 | 3.5361 | 2500 | 0.3903 | 0.2957 |
### Framework versions
- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1
|