matsschneider/nucleotide-virus-model

nucleotide-transformer-finetuned-virus

This model is a fine-tuned version of InstaDeepAI/nucleotide-transformer-500m-human-ref on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2432
• Accuracy: 0.9922
• F1: 0.9922
• Precision: 0.9925
• Recall: 0.9922

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: 16
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 250
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.7706	1.2195	100	1.7555	0.2109	0.1031	0.1165	0.2109
1.7314	2.4390	200	1.7009	0.5234	0.4263	0.4459	0.5234
1.6533	3.6585	300	1.6000	0.8828	0.8708	0.9316	0.8828
1.5609	4.8780	400	1.5144	0.8828	0.8820	0.9230	0.8828
1.4758	6.0976	500	1.4379	0.8281	0.8186	0.9057	0.8281
1.403	7.3171	600	1.3714	0.9688	0.9686	0.9708	0.9688
1.3421	8.5366	700	1.3186	0.9844	0.9844	0.9851	0.9844
1.2952	9.7561	800	1.2775	0.9766	0.9767	0.9783	0.9766
1.2619	10.9756	900	1.2507	1.0	1.0	1.0	1.0
1.2419	12.1951	1000	1.2432	0.9922	0.9922	0.9925	0.9922

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1