gena-lm-bigbird-base-t2t_ft_BioS2_1kbpHG19_DHSs_H3K27AC

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

• Loss: 0.4979
• F1 Score: 0.8766
• Precision: 0.8781
• Recall: 0.8750
• Accuracy: 0.8683
• Auc: 0.9406
• Prc: 0.9418

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
0.5349	0.0841	500	0.4552	0.8265	0.7778	0.8816	0.8022	0.8727	0.8639
0.4552	0.1682	1000	0.4734	0.8272	0.7263	0.9607	0.7856	0.8927	0.8877
0.4577	0.2523	1500	0.4191	0.8381	0.7512	0.9477	0.8044	0.9022	0.9005
0.4282	0.3364	2000	0.4104	0.8528	0.7777	0.9440	0.8259	0.9128	0.8970
0.4127	0.4205	2500	0.3636	0.8611	0.8367	0.8870	0.8471	0.9213	0.9216
0.4226	0.5045	3000	0.3621	0.8623	0.8096	0.9223	0.8426	0.9248	0.9255
0.4231	0.5886	3500	0.3553	0.8629	0.7931	0.9462	0.8394	0.9317	0.9329
0.3945	0.6727	4000	0.3843	0.8631	0.7856	0.9575	0.8377	0.9345	0.9341
0.3911	0.7568	4500	0.4173	0.8681	0.8571	0.8794	0.8572	0.9315	0.9330
0.4233	0.8409	5000	0.3419	0.8741	0.8249	0.9295	0.8569	0.9355	0.9376
0.3787	0.9250	5500	0.3880	0.8650	0.7891	0.9572	0.8404	0.9346	0.9357
0.3849	1.0091	6000	0.3629	0.8766	0.8512	0.9037	0.8641	0.9353	0.9359
0.3522	1.0932	6500	0.3683	0.8803	0.8558	0.9062	0.8683	0.9381	0.9381
0.3376	1.1773	7000	0.4292	0.8640	0.7824	0.9644	0.8377	0.9392	0.9373
0.365	1.2614	7500	0.4852	0.8667	0.7858	0.9663	0.8412	0.9403	0.9371
0.3569	1.3454	8000	0.5700	0.8720	0.8112	0.9427	0.8522	0.9352	0.9287
0.3822	1.4295	8500	0.3894	0.8817	0.8720	0.8917	0.8722	0.9406	0.9418
0.3391	1.5136	9000	0.4167	0.8696	0.8863	0.8536	0.8633	0.9413	0.9434
0.3591	1.5977	9500	0.3554	0.8853	0.8631	0.9087	0.8742	0.9432	0.9436
0.3699	1.6818	10000	0.4540	0.8812	0.8868	0.8757	0.8739	0.9440	0.9441
0.3777	1.7659	10500	0.4137	0.8849	0.8583	0.9131	0.8730	0.9421	0.9423
0.3602	1.8500	11000	0.3798	0.8736	0.8835	0.8640	0.8665	0.9414	0.9444
0.3583	1.9341	11500	0.4461	0.8840	0.8405	0.9323	0.8693	0.9438	0.9458
0.3573	2.0182	12000	0.4979	0.8766	0.8781	0.8750	0.8683	0.9406	0.9418

Framework versions

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