babylm-base2.5m-gpt2

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 3.3906
  • Accuracy: 0.4129

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: 16
  • eval_batch_size: 16
  • seed: 42
  • optimizer: Use OptimizerNames.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: 195
  • training_steps: 19500
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
5.8389 0.1065 200 4.8806 0.3144
4.8812 0.2130 400 4.4667 0.3240
4.707 0.3195 600 4.2884 0.3341
4.479 0.4260 800 4.2022 0.3348
4.2839 0.5325 1000 4.1496 0.3363
4.1679 0.6390 1200 4.1096 0.3369
4.125 0.7455 1400 4.0623 0.3393
3.9613 0.8520 1600 4.0172 0.3394
3.9268 0.9585 1800 4.0097 0.3367
3.8225 1.0650 2000 3.9714 0.3417
2.6933 2.1299 4000 3.6543 0.3824
2.3328 3.1949 6000 3.5196 0.3985
2.1778 4.2599 8000 3.4668 0.4016
2.0886 5.3248 10000 3.4159 0.4073
1.9842 6.3898 12000 3.3836 0.4104
1.9168 7.4547 14000 3.3775 0.4112
1.8811 8.5197 16000 3.3978 0.4121
1.8392 9.5847 18000 3.3874 0.4125

Framework versions

  • Transformers 4.50.3
  • Pytorch 2.7.1+cu126
  • Datasets 3.6.0
  • Tokenizers 0.21.4
Downloads last month
2
Safetensors
Model size
98.4M params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support