metadata

tags:
  - generated_from_trainer
datasets:
  - kanishka/counterfactual_babylm_measure_nps_as_singular_new
metrics:
  - accuracy
model-index:
  - name: smolm-autoreg-bpe-counterfactual_babylm_measure_nps_as_singular_new-1e-3
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: kanishka/counterfactual_babylm_measure_nps_as_singular_new
          type: kanishka/counterfactual_babylm_measure_nps_as_singular_new
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.4110677518157195

smolm-autoreg-bpe-counterfactual_babylm_measure_nps_as_singular_new-1e-3

This model was trained from scratch on the kanishka/counterfactual_babylm_measure_nps_as_singular_new dataset. It achieves the following results on the evaluation set:

Loss: 3.4116
Accuracy: 0.4111

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: 42
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
3.6071	1.0	18602	3.7886	0.3573
3.379	2.0	37204	3.5653	0.3798
3.2515	3.0	55806	3.4692	0.3918
3.1729	4.0	74408	3.4193	0.3983
3.1139	5.0	93010	3.3907	0.4026
3.0709	6.0	111612	3.3642	0.4043
3.0297	7.0	130214	3.3545	0.4067
2.9988	8.0	148816	3.3596	0.4080
2.9717	9.0	167418	3.3723	0.4087
2.9432	10.0	186020	3.3579	0.4093
2.9217	11.0	204622	3.3701	0.4098
2.8986	12.0	223224	3.3646	0.4103
2.8745	13.0	241826	3.3676	0.4105
2.8518	14.0	260428	3.3750	0.4110
2.8328	15.0	279030	3.3722	0.4111
2.8089	16.0	297632	3.3797	0.4115
2.7911	17.0	316234	3.3882	0.4109
2.773	18.0	334836	3.3951	0.4115
2.7517	19.0	353438	3.4023	0.4112
2.7342	20.0	372040	3.4116	0.4111

Framework versions

Transformers 4.38.0
Pytorch 2.3.1+cu121
Datasets 2.16.1
Tokenizers 0.15.2