smolm-autoreg-bpe-counterfactual-babylm-keys_to_pipps_2913-1e-3

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.3564
• Accuracy: 0.4106

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
3.6052	1.0	18629	3.7008	0.3603
3.3934	2.0	37258	3.4882	0.3819
3.261	3.0	55887	3.4100	0.3927
3.1793	4.0	74516	3.3695	0.3987
3.1256	5.0	93145	3.3170	0.4025
3.0813	6.0	111774	3.3176	0.4054
3.0435	7.0	130403	3.3097	0.4068
3.0127	8.0	149032	3.3222	0.4076
2.9872	9.0	167661	3.2991	0.4085
2.9646	10.0	186290	3.2921	0.4099
2.9347	11.0	204919	3.3166	0.4099
2.9194	12.0	223548	3.3042	0.4112
2.8968	13.0	242177	3.3127	0.4103
2.8738	14.0	260806	3.3203	0.4110
2.8528	15.0	279435	3.3126	0.4112
2.8322	16.0	298064	3.3313	0.4112
2.817	17.0	316693	3.3349	0.4111
2.7973	18.0	335322	3.3472	0.4105
2.7825	19.0	353951	3.3492	0.4107
2.7582	20.0	372580	3.3564	0.4106

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu121
• Datasets 2.12.0
• Tokenizers 0.14.1