bert-base-cased-DA-ChemTok-ZN1540K-V1-finetuned-MAOB-IC50s-V1

This model is a fine-tuned version of cafierom/bert-base-cased-DA-ChemTok-ZN1540K-V1. It is trained on the cafierom/MAOB_4828 dataset and achieves the following results on the evaluation set:

• Loss: 0.7933
• Accuracy: 0.7476
• F1: 0.7457

Model description

More information needed

Intended uses & limitations

Can classify MAOB IC50 values as < 50 nM, < 500 nM, and > 500 nM. See Confusion matrix below.

Training and evaluation data

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• 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
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.843	1.0	129	0.7707	0.6676	0.6566
0.6987	2.0	258	0.7429	0.6828	0.6623
0.6372	3.0	387	0.7077	0.7062	0.6988
0.5815	4.0	516	0.6765	0.7145	0.7152
0.5365	5.0	645	0.6745	0.7393	0.7326
0.4964	6.0	774	0.6776	0.7490	0.7443
0.4625	7.0	903	0.6934	0.7131	0.7195
0.4244	8.0	1032	0.7132	0.7434	0.7456
0.405	9.0	1161	0.6963	0.7490	0.7457
0.3699	10.0	1290	0.7159	0.7421	0.7407
0.3588	11.0	1419	0.7799	0.7448	0.7373
0.3311	12.0	1548	0.7471	0.7462	0.7491
0.3032	13.0	1677	0.7999	0.7352	0.7388
0.2913	14.0	1806	0.7933	0.7476	0.7457
0.2821	15.0	1935	0.8282	0.7324	0.7367
0.2677	16.0	2064	0.8217	0.7462	0.7470
0.2565	17.0	2193	0.8619	0.7393	0.7400
0.2412	18.0	2322	0.8673	0.7407	0.7439
0.2248	19.0	2451	0.8645	0.7407	0.7438
0.2252	20.0	2580	0.8656	0.7379	0.7410

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.1
• Tokenizers 0.21.0