metadata
library_name: transformers
license: apache-2.0
base_model: openai/whisper-large-v3
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: psst-model-8e-1s-difflib
results: []
psst-model-8e-1s-difflib
This model is a fine-tuned version of openai/whisper-large-v3 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4445
- Wer: 0.1554
- Iu F1: 0.7255
- Iu Tp: 847
- Iu Fp: 490
- Iu Fn: 151
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: 1e-05
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 664
- training_steps: 9480
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Iu F1 | Iu Tp | Iu Fp | Iu Fn |
|---|---|---|---|---|---|---|---|---|
| 0.9633 | 1.0 | 1185 | 0.4920 | 0.2578 | 0.5748 | 461 | 567 | 115 |
| 0.6300 | 2.0 | 2370 | 0.4784 | 0.2390 | 0.7674 | 442 | 134 | 134 |
| 0.3400 | 3.0 | 3555 | 0.4964 | 0.2216 | 0.7584 | 449 | 159 | 127 |
| 0.1584 | 4.0 | 4740 | 0.5594 | 0.2321 | 0.7517 | 451 | 173 | 125 |
| 0.0592 | 5.0 | 5925 | 0.6092 | 0.2269 | 0.7496 | 455 | 183 | 121 |
| 0.0264 | 6.0 | 7110 | 0.6907 | 0.2202 | 0.7760 | 452 | 137 | 124 |
| 0.0101 | 7.0 | 8295 | 0.7329 | 0.2196 | 0.7654 | 429 | 116 | 147 |
| 0.0033 | 8.0 | 9480 | 0.7761 | 0.2149 | 0.7690 | 446 | 138 | 130 |
Framework versions
- Transformers 5.6.2
- Pytorch 2.6.0+cu124
- Datasets 2.21.0
- Tokenizers 0.22.2