trillsson3-ft-keyword-spotting-15
This model is a fine-tuned version of vumichien/nonsemantic-speech-trillsson3 on the superb dataset. It achieves the following results on the evaluation set:
- Loss: 0.3563
- Accuracy: 0.9041
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: 32
- eval_batch_size: 64
- seed: 0
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10.0
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
1.1824 | 1.0 | 798 | 0.6478 | 0.7489 |
0.7448 | 2.0 | 1596 | 0.4274 | 0.8728 |
0.7089 | 3.0 | 2394 | 0.3723 | 0.8950 |
0.6781 | 4.0 | 3192 | 0.3563 | 0.9041 |
0.6386 | 5.0 | 3990 | 0.3441 | 0.8986 |
0.6342 | 6.0 | 4788 | 0.3380 | 0.8994 |
0.6275 | 7.0 | 5586 | 0.3376 | 0.8982 |
0.6349 | 8.0 | 6384 | 0.3333 | 0.9014 |
0.6261 | 9.0 | 7182 | 0.3295 | 0.9025 |
0.6188 | 10.0 | 7980 | 0.3322 | 0.9025 |
Framework versions
- Transformers 4.23.0.dev0
- Pytorch 1.12.1+cu113
- Datasets 2.7.1
- Tokenizers 0.13.2
- Downloads last month
- 10
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.