bert-base-cased-ChemTok-ZN250K-V1-finetuned-Tyrosinase-IC50s-V

This model is a fine-tuned version of cafierom/bert-base-cased-ChemTok-ZN250K-V1 on the cafierom/Tyrosinase1407_classes dataset.

It achieves the following results on the evaluation set:

• Loss: 1.662
• Accuracy: 0.7170
• F1: 0.7185

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: 2e-05
• train_batch_size: 128
• eval_batch_size: 128
• 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: constant
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.0669	1.0	10	1.0239	0.4858	0.4655
0.976	2.0	20	0.9348	0.5613	0.5533
0.8835	3.0	30	0.8840	0.5660	0.5618
0.7942	4.0	40	0.8649	0.5802	0.5811
0.7227	5.0	50	0.8239	0.5943	0.5968
0.6424	6.0	60	0.8165	0.6368	0.6303
0.5837	7.0	70	0.8034	0.6604	0.6623
0.5359	8.0	80	0.8236	0.6745	0.6765
0.4812	9.0	90	0.8230	0.6698	0.6683
0.4226	10.0	100	0.8526	0.6604	0.6623
0.3943	11.0	110	0.9106	0.6698	0.6703
0.3506	12.0	120	0.9329	0.6698	0.6721
0.3315	13.0	130	0.9385	0.6981	0.7002
0.3217	14.0	140	0.9566	0.6981	0.7006
0.2702	15.0	150	1.0311	0.6604	0.6646
0.2518	16.0	160	1.0195	0.6792	0.6812
0.2448	17.0	170	1.0773	0.6509	0.6485
0.2532	18.0	180	1.0814	0.6792	0.6808
0.2337	19.0	190	1.1534	0.6887	0.6918
0.2101	20.0	200	1.1669	0.7075	0.7083
0.1754	21.0	210	1.1488	0.6934	0.6942
0.1831	22.0	220	1.1709	0.6981	0.6989
0.1694	23.0	230	1.2001	0.7170	0.7160
0.1429	24.0	240	1.1662	0.7170	0.7185
0.1405	25.0	250	1.2133	0.7075	0.7076
0.1392	26.0	260	1.2233	0.6840	0.6834
0.1631	27.0	270	1.3062	0.6840	0.6860
0.1123	28.0	280	1.2992	0.6745	0.6771
0.0997	29.0	290	1.4074	0.6887	0.6895
0.1132	30.0	300	1.3494	0.7123	0.7144

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1