gpt2_small_RU_unigram_11703_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.3735
- Accuracy: 0.3991
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: 5e-05
- train_batch_size: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|
| 5.7297 | 1.0 | 1746 | 4.2763 | 0.3123 |
| 4.1033 | 2.0 | 3492 | 3.8519 | 0.3477 |
| 3.7822 | 3.0 | 5238 | 3.6689 | 0.3655 |
| 3.6799 | 4.0 | 6984 | 3.5647 | 0.3769 |
| 3.548 | 5.0 | 8730 | 3.5000 | 0.3841 |
| 3.4667 | 6.0 | 10476 | 3.4499 | 0.3897 |
| 3.4065 | 7.0 | 12222 | 3.4205 | 0.3927 |
| 3.3719 | 8.0 | 13968 | 3.3972 | 0.3959 |
| 3.329 | 9.0 | 15714 | 3.3805 | 0.3984 |
| 3.3022 | 10.0 | 17460 | 3.3735 | 0.3991 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
