babylm-base9m-roberta

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

• Loss: 5.5708
• Accuracy: 0.1684

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: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 182
• training_steps: 18200

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
6.8945	0.1052	200	6.4849	0.1015
6.1415	0.2103	400	6.1729	0.1288
6.0058	0.3155	600	6.0430	0.1356
5.9066	0.4206	800	5.9878	0.1416
5.8336	0.5258	1000	5.9424	0.1441
5.8452	0.6309	1200	5.9035	0.1455
5.7943	0.7361	1400	5.8767	0.1483
5.7741	0.8412	1600	5.8617	0.1487
5.7285	0.9464	1800	5.8425	0.1495
5.7183	1.0515	2000	5.8324	0.1509
5.6107	2.1030	4000	5.7540	0.1565
5.5729	3.1546	6000	5.7043	0.1599
5.5388	4.2061	8000	5.6663	0.1614
5.4963	5.2576	10000	5.6391	0.1642
5.455	6.3091	12000	5.6120	0.1661
5.4238	7.3607	14000	5.5991	0.1665
5.4392	8.4122	16000	5.5841	0.1675
5.4325	9.4637	18000	5.5776	0.1677

Framework versions

• Transformers 4.50.3
• Pytorch 2.7.1+cu126
• Datasets 3.6.0
• Tokenizers 0.21.4