cria-babylm2-subset-default-3e-4
This model was trained from scratch on the kanishka/babylm2-subset dataset. It achieves the following results on the evaluation set:
- Loss: 2.6186
- Accuracy: 0.5184
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: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 32000
- num_epochs: 10.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.5065 | 1.0 | 14142 | 2.7266 | 0.4912 |
| 2.3323 | 2.0 | 28284 | 2.5642 | 0.5074 |
| 2.2158 | 3.0 | 42426 | 2.4670 | 0.5184 |
| 2.1109 | 4.0 | 56568 | 2.4178 | 0.5249 |
| 2.0194 | 5.0 | 70710 | 2.4001 | 0.5280 |
| 1.938 | 6.0 | 84852 | 2.4067 | 0.5291 |
| 1.8569 | 7.0 | 98994 | 2.4313 | 0.5283 |
| 1.7668 | 8.0 | 113136 | 2.4766 | 0.5260 |
| 1.6733 | 9.0 | 127278 | 2.5417 | 0.5229 |
| 1.579 | 10.0 | 141420 | 2.6186 | 0.5184 |
Framework versions
- Transformers 4.42.4
- Pytorch 2.2.0+cu121
- Datasets 2.16.1
- Tokenizers 0.19.1
- Downloads last month
- 9
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.