gpt2_small_babyLM_25_bigbatch_20e

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.0814
• Accuracy: 0.4476

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: 0.0001
• train_batch_size: 1024
• eval_batch_size: 1024
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 4000
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
8.1529	1.0	349	5.6417	0.3264
4.8394	2.0	698	4.4517	0.3615
4.0801	3.0	1047	4.0320	0.3825
3.8874	4.0	1396	3.7920	0.3955
3.631	5.0	1745	3.6363	0.4047
3.4693	6.0	2094	3.5230	0.4122
3.346	7.0	2443	3.4352	0.4184
3.2935	8.0	2792	3.3636	0.4227
3.2034	9.0	3141	3.3068	0.4276
3.134	10.0	3490	3.2580	0.4311
3.0741	11.0	3839	3.2183	0.4342
3.0375	12.0	4188	3.1834	0.4370
2.9774	13.0	4537	3.1554	0.4392
2.9339	14.0	4886	3.1321	0.4414
2.8966	15.0	5235	3.1161	0.4432
2.8681	16.0	5584	3.1045	0.4446
2.8363	17.0	5933	3.0933	0.4458
2.8111	18.0	6282	3.0875	0.4467
2.7909	19.0	6631	3.0831	0.4472
2.7743	20.0	6980	3.0814	0.4476

Framework versions

• Transformers 4.30.2
• Pytorch 2.9.1+cu128
• Datasets 4.1.1
• Tokenizers 0.13.3