File size: 3,067 Bytes
0e8822d |
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 92 93 94 95 96 97 98 99 100 |
---
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-1b
tags:
- generated_from_trainer
datasets:
- common_voice_13_0
metrics:
- wer
model-index:
- name: LugandaASRwav20Vec1B
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_13_0
type: common_voice_13_0
config: lg
split: validation
args: lg
metrics:
- name: Wer
type: wer
value: 0.23043478260869565
---
<!-- 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. -->
# LugandaASRwav20Vec1B
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_13_0 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1854
- Wer: 0.2304
## 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: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 24
- total_train_batch_size: 96
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 200
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|
| 4.303 | 0.14 | 100 | 2.1141 | 1.0 |
| 0.7155 | 0.27 | 200 | 0.5656 | 0.6752 |
| 0.4493 | 0.41 | 300 | 0.4402 | 0.5607 |
| 0.3964 | 0.54 | 400 | 0.3918 | 0.5114 |
| 0.3646 | 0.68 | 500 | 0.3601 | 0.4592 |
| 0.3294 | 0.81 | 600 | 0.3381 | 0.4467 |
| 0.3339 | 0.95 | 700 | 0.3340 | 0.4266 |
| 0.2893 | 1.08 | 800 | 0.2913 | 0.3670 |
| 0.2743 | 1.22 | 900 | 0.2854 | 0.3600 |
| 0.262 | 1.36 | 1000 | 0.2666 | 0.3318 |
| 0.2545 | 1.49 | 1100 | 0.2601 | 0.3341 |
| 0.2437 | 1.63 | 1200 | 0.2488 | 0.3152 |
| 0.2235 | 1.76 | 1300 | 0.2416 | 0.3015 |
| 0.2188 | 1.9 | 1400 | 0.2330 | 0.2902 |
| 0.2054 | 2.03 | 1500 | 0.2218 | 0.2750 |
| 0.1743 | 2.17 | 1600 | 0.2153 | 0.2672 |
| 0.1722 | 2.3 | 1700 | 0.2098 | 0.2575 |
| 0.1656 | 2.44 | 1800 | 0.2011 | 0.2538 |
| 0.1608 | 2.58 | 1900 | 0.2000 | 0.2475 |
| 0.1574 | 2.71 | 2000 | 0.1937 | 0.2428 |
| 0.1531 | 2.85 | 2100 | 0.1882 | 0.2347 |
| 0.1451 | 2.98 | 2200 | 0.1854 | 0.2304 |
### Framework versions
- Transformers 4.32.0
- Pytorch 2.0.1+cu118
- Datasets 2.13.0
- Tokenizers 0.13.3
|