<!-- 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. -->
pratyush_whisper_small_distil_libri360_enc_6_dec_4_batch_4_epoch_2_try2
This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:
- Loss: 2.7423
- Wer: 9.7882
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: 0.0005
- train_batch_size: 4
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 256
- total_train_batch_size: 1024
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.2
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
2.8038 | 0.25 | 50 | 2.7934 | 11.8801 |
2.8683 | 0.49 | 100 | 2.8112 | 13.4039 |
2.8987 | 0.74 | 150 | 2.8008 | 11.8782 |
2.884 | 0.98 | 200 | 2.7877 | 11.1632 |
2.8539 | 1.23 | 250 | 2.7721 | 10.6549 |
2.8348 | 1.48 | 300 | 2.7557 | 10.3250 |
2.8261 | 1.72 | 350 | 2.7522 | 10.1403 |
2.8161 | 1.97 | 400 | 2.7423 | 9.7882 |
Framework versions
- Transformers 4.29.2
- Pytorch 1.12.1
- Datasets 2.8.0
- Tokenizers 0.13.2