metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- xtreme_s
metrics:
- accuracy
model-index:
- name: xtreme_s_xlsr_300m_fleurs_langid_truncated
results: []
xtreme_s_xlsr_300m_fleurs_langid_truncated
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the xtreme_s dataset. It achieves the following results on the evaluation set:
- Loss: 1.3514
- Accuracy: 0.7236
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: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2000
- num_epochs: 5.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.5296 | 0.26 | 1000 | 2.6633 | 0.4016 |
| 0.4252 | 0.52 | 2000 | 1.8582 | 0.5751 |
| 0.2989 | 0.78 | 3000 | 1.6780 | 0.6332 |
| 0.3563 | 1.04 | 4000 | 1.4479 | 0.6799 |
| 0.1617 | 1.3 | 5000 | 1.5066 | 0.6679 |
| 0.1409 | 1.56 | 6000 | 1.4082 | 0.6992 |
| 0.01 | 1.82 | 7000 | 1.2448 | 0.7071 |
| 0.0018 | 2.08 | 8000 | 1.1996 | 0.7148 |
| 0.0014 | 2.34 | 9000 | 1.6505 | 0.6410 |
| 0.0188 | 2.6 | 10000 | 1.4050 | 0.6840 |
| 0.0007 | 2.86 | 11000 | 1.5831 | 0.6621 |
| 0.1038 | 3.12 | 12000 | 1.5441 | 0.6829 |
| 0.0003 | 3.38 | 13000 | 1.3483 | 0.6900 |
| 0.0004 | 3.64 | 14000 | 1.7070 | 0.6414 |
| 0.0003 | 3.9 | 15000 | 1.3198 | 0.7075 |
| 0.0002 | 4.16 | 16000 | 1.3118 | 0.7105 |
| 0.0001 | 4.42 | 17000 | 1.4099 | 0.7029 |
| 0.0 | 4.68 | 18000 | 1.3658 | 0.7180 |
| 0.0001 | 4.93 | 19000 | 1.3514 | 0.7236 |
Framework versions
- Transformers 4.18.0.dev0
- Pytorch 1.11.0+cu113
- Datasets 1.18.4.dev0
- Tokenizers 0.11.6