GUE_EMP_H3K9ac-seqsight_16384_512_56M-L1_f

This model is a fine-tuned version of mahdibaghbanzadeh/seqsight_16384_512_56M on the mahdibaghbanzadeh/GUE_EMP_H3K9ac dataset. It achieves the following results on the evaluation set:

• Loss: 0.4756
• F1 Score: 0.7781
• Accuracy: 0.7776

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
0.5856	1.15	200	0.5516	0.7251	0.7272
0.5437	2.3	400	0.5871	0.6926	0.6977
0.5193	3.45	600	0.5526	0.7251	0.7262
0.5138	4.6	800	0.5540	0.7213	0.7229
0.5066	5.75	1000	0.5250	0.7486	0.7481
0.4989	6.9	1200	0.5327	0.7470	0.7467
0.4933	8.05	1400	0.5393	0.7398	0.7398
0.4915	9.2	1600	0.5634	0.7221	0.7258
0.4827	10.34	1800	0.5314	0.7425	0.7424
0.4864	11.49	2000	0.5129	0.7586	0.7582
0.476	12.64	2200	0.5526	0.7235	0.7265
0.4789	13.79	2400	0.5209	0.7424	0.7427
0.4746	14.94	2600	0.5149	0.7450	0.7452
0.4726	16.09	2800	0.5087	0.7532	0.7531
0.4701	17.24	3000	0.5291	0.7429	0.7434
0.4653	18.39	3200	0.5184	0.7429	0.7438
0.4663	19.54	3400	0.5100	0.7472	0.7481
0.463	20.69	3600	0.5015	0.7633	0.7629
0.4636	21.84	3800	0.5214	0.7373	0.7391
0.46	22.99	4000	0.5220	0.7375	0.7395
0.4624	24.14	4200	0.4973	0.7628	0.7625
0.4533	25.29	4400	0.5217	0.7512	0.7517
0.461	26.44	4600	0.5081	0.7574	0.7575
0.4552	27.59	4800	0.5101	0.7526	0.7531
0.4525	28.74	5000	0.5097	0.7493	0.7503
0.4569	29.89	5200	0.5063	0.7617	0.7618
0.4494	31.03	5400	0.5174	0.7449	0.7463
0.4531	32.18	5600	0.4900	0.7669	0.7665
0.4438	33.33	5800	0.5002	0.7671	0.7668
0.4539	34.48	6000	0.5053	0.7548	0.7553
0.4429	35.63	6200	0.4950	0.7679	0.7675
0.4503	36.78	6400	0.4991	0.7635	0.7636
0.4449	37.93	6600	0.5143	0.7543	0.7549
0.4454	39.08	6800	0.4985	0.7660	0.7661
0.4446	40.23	7000	0.5068	0.7601	0.7607
0.4443	41.38	7200	0.5043	0.7607	0.7611
0.4445	42.53	7400	0.5047	0.7618	0.7621
0.4415	43.68	7600	0.5023	0.7626	0.7629
0.4388	44.83	7800	0.5066	0.7587	0.7593
0.4428	45.98	8000	0.4992	0.7662	0.7661
0.4446	47.13	8200	0.5115	0.7582	0.7589
0.4398	48.28	8400	0.5004	0.7646	0.7647
0.4361	49.43	8600	0.5021	0.7613	0.7614
0.4399	50.57	8800	0.5032	0.7650	0.7650
0.4384	51.72	9000	0.5080	0.7614	0.7618
0.4373	52.87	9200	0.5105	0.7562	0.7567
0.44	54.02	9400	0.5097	0.7608	0.7614
0.4384	55.17	9600	0.5087	0.7598	0.7603
0.4352	56.32	9800	0.5021	0.7627	0.7629
0.4381	57.47	10000	0.5048	0.7622	0.7625

Framework versions

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