metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- common_voice
model-index:
- name: wav2vec2-large-xls-r-300m-hindi
results: []
wav2vec2-large-xls-r-300m-hindi
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset. It achieves the following results on the evaluation set:
- Loss: 2.6718
- Wer: 0.7103
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
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.5682 | 2.72 | 400 | 2.1019 | 0.9188 |
| 0.6506 | 5.44 | 800 | 1.9496 | 0.8048 |
| 0.3249 | 8.16 | 1200 | 1.8901 | 0.7515 |
| 0.222 | 10.88 | 1600 | 1.7736 | 0.7115 |
| 0.171 | 13.6 | 2000 | 2.1061 | 0.7507 |
| 0.1428 | 16.33 | 2400 | 2.2476 | 0.7412 |
| 0.1235 | 19.05 | 2800 | 2.3527 | 0.7554 |
| 0.1076 | 21.77 | 3200 | 2.2145 | 0.7404 |
| 0.0982 | 24.49 | 3600 | 2.3603 | 0.7327 |
| 0.0842 | 27.21 | 4000 | 2.4086 | 0.7465 |
| 0.0732 | 29.93 | 4400 | 2.4182 | 0.7259 |
| 0.0672 | 32.65 | 4800 | 2.5249 | 0.7315 |
| 0.0601 | 35.37 | 5200 | 2.5355 | 0.7207 |
| 0.0534 | 38.09 | 5600 | 2.5170 | 0.7191 |
| 0.0477 | 40.81 | 6000 | 2.6001 | 0.7064 |
| 0.0435 | 43.54 | 6400 | 2.7135 | 0.7142 |
| 0.0374 | 46.26 | 6800 | 2.6552 | 0.7127 |
| 0.0348 | 48.98 | 7200 | 2.6718 | 0.7103 |
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.0+cu113
- Datasets 1.17.1.dev0
- Tokenizers 0.10.3