hyenadna-small-32k-seqlen-hf_ft_BioS2_1kbpHG19_DHSs_H3K27AC

This model is a fine-tuned version of LongSafari/hyenadna-small-32k-seqlen-hf on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4638
• F1 Score: 0.8073
• Precision: 0.8018
• Recall: 0.8128
• Accuracy: 0.7952
• Auc: 0.8728
• Prc: 0.8651

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
0.5671	0.0841	500	0.5418	0.7721	0.7343	0.8141	0.7464	0.7985	0.7787
0.5022	0.1683	1000	0.5186	0.7826	0.7680	0.7978	0.7661	0.8279	0.8137
0.5022	0.2524	1500	0.5044	0.8036	0.7154	0.9168	0.7635	0.8360	0.8138
0.4865	0.3366	2000	0.4858	0.7887	0.7708	0.8074	0.7716	0.8440	0.8350
0.4913	0.4207	2500	0.4686	0.8062	0.7560	0.8635	0.7809	0.8553	0.8451
0.472	0.5049	3000	0.4715	0.8107	0.7494	0.8830	0.7824	0.8580	0.8462
0.4776	0.5890	3500	0.4581	0.8064	0.7682	0.8485	0.7849	0.8630	0.8574
0.4571	0.6732	4000	0.4620	0.8162	0.7506	0.8945	0.7874	0.8674	0.8548
0.4702	0.7573	4500	0.4550	0.8009	0.7995	0.8023	0.7895	0.8683	0.8600
0.4533	0.8415	5000	0.4493	0.8121	0.7891	0.8364	0.7957	0.8729	0.8617
0.4401	0.9256	5500	0.4472	0.8134	0.7823	0.8469	0.7949	0.8743	0.8662
0.4505	1.0098	6000	0.4459	0.8100	0.7892	0.8320	0.7940	0.8732	0.8668
0.3944	1.0939	6500	0.4638	0.8073	0.8018	0.8128	0.7952	0.8728	0.8651

Framework versions

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