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

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

• Loss: 3.1671
• Accuracy: 0.4317

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
3.4391	1.0	18353	3.5453	0.3791
3.209	2.0	36706	3.3518	0.3998
3.0697	3.0	55059	3.2533	0.4131
2.9929	4.0	73412	3.2030	0.4182
2.9314	5.0	91765	3.1689	0.4226
2.8882	6.0	110118	3.1527	0.4250
2.8492	7.0	128471	3.1471	0.4265
2.8162	8.0	146824	3.1343	0.4289
2.7884	9.0	165177	3.1328	0.4291
2.76	10.0	183530	3.1282	0.4301
2.7351	11.0	201883	3.1330	0.4305
2.7198	12.0	220236	3.1316	0.4304
2.6935	13.0	238589	3.1416	0.4309
2.6666	14.0	256942	3.1364	0.4316
2.6514	15.0	275295	3.1413	0.4315
2.6313	16.0	293648	3.1467	0.4319
2.6089	17.0	312001	3.1499	0.4320
2.586	18.0	330354	3.1559	0.4318
2.5692	19.0	348707	3.1615	0.4317
2.5476	20.0	367060	3.1671	0.4317

Framework versions

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