hyenadna-small-32k-seqlen-hf_ft_BioS74_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: 0.4605
• F1 Score: 0.8107
• Precision: 0.7689
• Recall: 0.8574
• Accuracy: 0.7904
• Auc: 0.8669
• Prc: 0.8594

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
0.5565	0.1314	500	0.5060	0.7829	0.7556	0.8122	0.7641	0.8275	0.8077
0.4991	0.2629	1000	0.5124	0.7801	0.7916	0.7690	0.7731	0.8364	0.8255
0.494	0.3943	1500	0.4957	0.7822	0.8035	0.7619	0.7778	0.8460	0.8325
0.4752	0.5258	2000	0.5005	0.7964	0.7661	0.8292	0.7781	0.8496	0.8368
0.4546	0.6572	2500	0.4923	0.8041	0.7414	0.8785	0.7760	0.8515	0.8359
0.4858	0.7886	3000	0.4669	0.8017	0.7556	0.8538	0.7789	0.8495	0.8355
0.4677	0.9201	3500	0.4842	0.8019	0.7881	0.8162	0.7889	0.8583	0.8467
0.4695	1.0515	4000	0.4893	0.7893	0.8102	0.7695	0.7849	0.8616	0.8504
0.4576	1.1830	4500	0.4612	0.8078	0.7760	0.8423	0.7902	0.8599	0.8482
0.4574	1.3144	5000	0.4591	0.8122	0.7633	0.8679	0.7899	0.8629	0.8519
0.445	1.4458	5500	0.5035	0.7831	0.8194	0.7499	0.7825	0.8640	0.8589
0.4302	1.5773	6000	0.4984	0.8064	0.7856	0.8282	0.7917	0.8622	0.8502
0.4252	1.7087	6500	0.4651	0.8007	0.7973	0.8041	0.7904	0.8642	0.8551
0.4408	1.8402	7000	0.4837	0.7988	0.8103	0.7875	0.7923	0.8619	0.8567
0.4444	1.9716	7500	0.4605	0.8107	0.7689	0.8574	0.7904	0.8669	0.8594

Framework versions

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