LiukG/gut_1024-finetuned-lora-bigbird-base-t2t

test_run

This model is a fine-tuned version of AIRI-Institute/gena-lm-bigbird-base-t2t on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2397
• F1: 0.8195
• Mcc Score: 0.5808
• Accuracy: 0.7933

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Mcc Score	Accuracy
0.6329	1.0	94	0.5532	0.7711	0.4531	0.7335
0.4921	2.0	188	0.4789	0.8359	0.5501	0.7832
0.3981	3.0	282	0.4760	0.8347	0.5789	0.7987
0.3579	4.0	376	0.6767	0.7737	0.5377	0.7587
0.2488	5.0	470	0.5478	0.8327	0.5887	0.8015
0.1889	6.0	564	0.7844	0.8231	0.5846	0.7962
0.1569	7.0	658	0.8773	0.8254	0.5868	0.7978
0.1034	8.0	752	1.4445	0.7499	0.4939	0.7353
0.0832	9.0	846	1.6405	0.7195	0.4955	0.7205
0.1051	10.0	940	1.2397	0.8195	0.5808	0.7933

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2