smolm-autoreg-bpe-babylm-aann-counterfactual-anan-1e-3

This model is a fine-tuned version of models/smolm-autoreg-bpe-babylm-aann-counterfactual-anan-1e-3/config.json on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.1635
• Accuracy: 0.4321

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: 128
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.4374	1.0	18353	3.5466	0.3796
3.2099	2.0	36706	3.3491	0.3995
3.0711	3.0	55059	3.2446	0.4134
2.9939	4.0	73412	3.1977	0.4186
2.933	5.0	91765	3.1660	0.4224
2.8889	6.0	110118	3.1475	0.4254
2.8508	7.0	128471	3.1383	0.4273
2.8168	8.0	146824	3.1297	0.4288
2.7896	9.0	165177	3.1278	0.4295
2.7611	10.0	183530	3.1289	0.4303
2.7357	11.0	201883	3.1282	0.4309
2.7211	12.0	220236	3.1289	0.4312
2.695	13.0	238589	3.1367	0.4315
2.6675	14.0	256942	3.1350	0.4323
2.6523	15.0	275295	3.1402	0.4318
2.6323	16.0	293648	3.1416	0.4322
2.6102	17.0	312001	3.1500	0.4322
2.5871	18.0	330354	3.1509	0.4325
2.5708	19.0	348707	3.1577	0.4323
2.549	20.0	367060	3.1635	0.4321

Framework versions

• Transformers 4.34.0
• Pytorch 2.0.1+cu117
• Datasets 2.12.0
• Tokenizers 0.14.1