kanishka's picture
End of training
a9ddcc2 verified
metadata
tags:
  - generated_from_trainer
datasets:
  - kanishka/counterfactual-babylm-pipps-random_removal
metrics:
  - accuracy
model-index:
  - name: smolm-autoreg-bpe-counterfactual-babylm-pipps-random_removal-seed_211-1e-3
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: kanishka/counterfactual-babylm-pipps-random_removal
          type: kanishka/counterfactual-babylm-pipps-random_removal
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.40988659662430754

smolm-autoreg-bpe-counterfactual-babylm-pipps-random_removal-seed_211-1e-3

This model was trained from scratch on the kanishka/counterfactual-babylm-pipps-random_removal dataset. It achieves the following results on the evaluation set:

  • Loss: 3.4120
  • Accuracy: 0.4099

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: 0.001
  • train_batch_size: 32
  • eval_batch_size: 64
  • seed: 211
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 32000
  • num_epochs: 20.0
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
3.6072 1.0 18592 3.7972 0.3577
3.3863 2.0 37184 3.5804 0.3804
3.2556 3.0 55776 3.4730 0.3910
3.1829 4.0 74368 3.4019 0.3992
3.1264 5.0 92960 3.3828 0.4020
3.0827 6.0 111552 3.3849 0.4031
3.0461 7.0 130144 3.3728 0.4050
3.0111 8.0 148736 3.3609 0.4069
2.9857 9.0 167328 3.3496 0.4082
2.9608 10.0 185920 3.3683 0.4075
2.9402 11.0 204512 3.3728 0.4086
2.9154 12.0 223104 3.3845 0.4083
2.891 13.0 241696 3.3741 0.4098
2.8754 14.0 260288 3.3674 0.4106
2.8555 15.0 278880 3.3868 0.4095
2.8368 16.0 297472 3.3892 0.4098
2.8185 17.0 316064 3.3865 0.4106
2.7969 18.0 334656 3.4006 0.4099
2.7805 19.0 353248 3.3997 0.4104
2.7623 20.0 371840 3.4120 0.4099

Framework versions

  • Transformers 4.37.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.1