gena-lm-bert-large-t2t_ft_BioS2_1kbpHG19_DHSs_H3K27AC

This model is a fine-tuned version of AIRI-Institute/gena-lm-bert-large-t2t on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3605
• F1 Score: 0.8722
• Precision: 0.8099
• Recall: 0.9449
• Accuracy: 0.8549
• Auc: 0.9416
• Prc: 0.9432

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Score	Precision	Recall	Accuracy	Auc	Prc
0.5815	0.0840	500	0.5396	0.7416	0.8093	0.6843	0.7500	0.8334	0.8249
0.4864	0.1680	1000	0.4740	0.8142	0.7803	0.8513	0.7964	0.8687	0.8579
0.4701	0.2520	1500	0.5119	0.7828	0.8415	0.7317	0.7872	0.8688	0.8706
0.4654	0.3360	2000	0.4932	0.8274	0.7359	0.9449	0.7934	0.8922	0.8869
0.4376	0.4200	2500	0.4289	0.8365	0.8007	0.8756	0.8206	0.8984	0.8962
0.423	0.5039	3000	0.4881	0.8366	0.7595	0.9311	0.8093	0.9056	0.9034
0.4226	0.5879	3500	0.3795	0.8504	0.8110	0.8939	0.8352	0.9131	0.9133
0.4035	0.6719	4000	0.3801	0.8491	0.7886	0.9196	0.8287	0.9186	0.9217
0.4041	0.7559	4500	0.3687	0.8553	0.8172	0.8971	0.8409	0.9226	0.9244
0.3948	0.8399	5000	0.4052	0.8421	0.7465	0.9657	0.8102	0.9290	0.9314
0.3914	0.9239	5500	0.4325	0.8483	0.7639	0.9538	0.8213	0.9294	0.9319
0.3801	1.0079	6000	0.3344	0.8677	0.8623	0.8731	0.8604	0.9339	0.9368
0.3334	1.0919	6500	0.3710	0.8695	0.8301	0.9128	0.8564	0.9337	0.9368
0.3567	1.1759	7000	0.3799	0.8672	0.8740	0.8606	0.8619	0.9372	0.9398
0.3528	1.2599	7500	0.3344	0.8698	0.8609	0.8788	0.8621	0.9378	0.9404
0.3542	1.3439	8000	0.3936	0.8769	0.8249	0.9359	0.8623	0.9393	0.9405
0.3497	1.4279	8500	0.3461	0.8730	0.8586	0.8878	0.8646	0.9393	0.9403
0.3379	1.5118	9000	0.4219	0.8683	0.7987	0.9513	0.8488	0.9408	0.9417
0.347	1.5958	9500	0.3649	0.8746	0.8491	0.9016	0.8644	0.9391	0.9409
0.3513	1.6798	10000	0.3315	0.8655	0.8944	0.8385	0.8634	0.9414	0.9438
0.3259	1.7638	10500	0.3605	0.8722	0.8099	0.9449	0.8549	0.9416	0.9432

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.18.0
• Tokenizers 0.19.0