metadata
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
- generated_from_trainer
datasets:
- common_voice_16_1
metrics:
- wer
model-index:
- name: wav2vec2-large-xls-r-300m-amharic-demo-colab
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: common_voice_16_1
type: common_voice_16_1
config: am
split: test
args: am
metrics:
- type: wer
value: 1.0006671114076051
name: Wer
wav2vec2-large-xls-r-300m-amharic-demo-colab
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_16_1 dataset. It achieves the following results on the evaluation set:
- Loss: 3.9728
- Wer: 1.0007
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: 100
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 12.5906 | 5.0 | 100 | 4.1542 | 1.0 |
| 4.1313 | 10.0 | 200 | 4.0748 | 1.0 |
| 4.025 | 15.0 | 300 | 3.9942 | 1.0 |
| 3.9704 | 20.0 | 400 | 3.9728 | 1.0007 |
Framework versions
- Transformers 4.42.0
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1