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

Loss: 3.4361
Accuracy: 0.4089

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
3.6075	1.0	18593	3.7579	0.3590
3.3887	2.0	37186	3.5741	0.3799
3.2563	3.0	55779	3.4841	0.3909
3.1774	4.0	74372	3.4295	0.3974
3.1289	5.0	92965	3.4102	0.4011
3.0832	6.0	111558	3.3941	0.4035
3.0461	7.0	130151	3.4059	0.4050
3.0149	8.0	148744	3.3837	0.4063
2.9868	9.0	167337	3.3869	0.4071
2.9592	10.0	185930	3.3717	0.4084
2.9402	11.0	204523	3.3799	0.4090
2.9148	12.0	223116	3.3744	0.4095
2.8923	13.0	241709	3.3787	0.4090
2.8739	14.0	260302	3.3927	0.4091
2.8561	15.0	278895	3.4054	0.4086
2.8348	16.0	297488	3.4210	0.4086
2.8099	17.0	316081	3.4029	0.4092
2.7966	18.0	334674	3.4200	0.4090
2.7784	19.0	353267	3.4326	0.4089
2.7578	20.0	371860	3.4361	0.4089

Framework versions

Transformers 4.38.0
Pytorch 2.3.1+cu121
Datasets 2.16.1
Tokenizers 0.15.2

kanishka/smolm-autoreg-bpe-counterfactual_babylm_aann_low_variability_adj-seed_211-1e-3

smolm-autoreg-bpe-counterfactual_babylm_aann_low_variability_adj_211-1e-3

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

kanishka/
smolm-autoreg-bpe-counterfactual_babylm_aann_low_variability_adj-seed_211-1e-3