GUE_EMP_H3K14ac-seqsight_32768_512_43M-L1_f

This model is a fine-tuned version of mahdibaghbanzadeh/seqsight_32768_512_43M on the mahdibaghbanzadeh/GUE_EMP_H3K14ac dataset. It achieves the following results on the evaluation set:

• Loss: 0.4907
• F1 Score: 0.7713
• Accuracy: 0.7703

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: 0.0005
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 10000

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Score	Accuracy
0.6078	0.97	200	0.5696	0.7214	0.7198
0.5576	1.93	400	0.5322	0.7501	0.7486
0.5381	2.9	600	0.5385	0.7543	0.7528
0.5289	3.86	800	0.5084	0.7646	0.7643
0.5195	4.83	1000	0.5251	0.7586	0.7570
0.5138	5.8	1200	0.5170	0.7626	0.7610
0.5131	6.76	1400	0.5057	0.7662	0.7646
0.5086	7.73	1600	0.5034	0.7698	0.7682
0.5062	8.7	1800	0.5035	0.7668	0.7652
0.5012	9.66	2000	0.5088	0.7659	0.7643
0.5059	10.63	2200	0.5152	0.7624	0.7610
0.4987	11.59	2400	0.4991	0.7686	0.7670
0.5029	12.56	2600	0.5098	0.7674	0.7658
0.4966	13.53	2800	0.5062	0.7658	0.7643
0.4979	14.49	3000	0.5158	0.7632	0.7619
0.4895	15.46	3200	0.4918	0.7751	0.7737
0.4949	16.43	3400	0.5080	0.7645	0.7631
0.4919	17.39	3600	0.4903	0.7742	0.7728
0.4882	18.36	3800	0.4883	0.7733	0.7722
0.4895	19.32	4000	0.4909	0.7752	0.7737
0.4871	20.29	4200	0.4916	0.7761	0.7746
0.487	21.26	4400	0.4970	0.7722	0.7707
0.4855	22.22	4600	0.5079	0.7702	0.7688
0.4866	23.19	4800	0.4903	0.7770	0.7755
0.4869	24.15	5000	0.4891	0.7731	0.7716
0.4828	25.12	5200	0.5005	0.7713	0.7697
0.4815	26.09	5400	0.4942	0.7740	0.7725
0.4814	27.05	5600	0.5042	0.7690	0.7676
0.4829	28.02	5800	0.4832	0.7760	0.7746
0.4815	28.99	6000	0.4999	0.7733	0.7719
0.4804	29.95	6200	0.4979	0.7743	0.7728
0.4816	30.92	6400	0.4819	0.7778	0.7764
0.4798	31.88	6600	0.4874	0.7749	0.7734
0.4784	32.85	6800	0.4942	0.7752	0.7737
0.483	33.82	7000	0.4982	0.7731	0.7716
0.4786	34.78	7200	0.4936	0.7731	0.7716
0.4794	35.75	7400	0.4892	0.7770	0.7755
0.4748	36.71	7600	0.4904	0.7731	0.7716
0.4772	37.68	7800	0.4898	0.7758	0.7743
0.4771	38.65	8000	0.4837	0.7770	0.7755
0.4826	39.61	8200	0.4880	0.7749	0.7734
0.4715	40.58	8400	0.4948	0.7725	0.7710
0.4742	41.55	8600	0.4891	0.7734	0.7719
0.4721	42.51	8800	0.4891	0.7737	0.7722
0.475	43.48	9000	0.4985	0.7743	0.7728
0.4741	44.44	9200	0.4925	0.7740	0.7725
0.4757	45.41	9400	0.4892	0.7731	0.7716
0.469	46.38	9600	0.4934	0.7740	0.7725
0.4794	47.34	9800	0.4906	0.7740	0.7725
0.474	48.31	10000	0.4891	0.7740	0.7725

Framework versions

• PEFT 0.9.0
• Transformers 4.38.2
• Pytorch 2.2.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2