finetuned-beit-limb-seq-t3-2heads-1layers-1e-4lr

This model is a fine-tuned version of c14kevincardenas/beit-large-patch16-384-limb-person-crop on the c14kevincardenas/beta_caller_284_person_crop_seq_withlimb_3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3876
• Accuracy: 0.9208

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 32
• seed: 2014
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 250
• num_epochs: 20.0
• mixed_precision_training: Native AMP
• label_smoothing_factor: 0.05

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6956	1.0	164	0.4265	0.9089
0.4121	2.0	328	0.4510	0.9089
0.4086	3.0	492	0.4066	0.9187
0.4257	4.0	656	0.3893	0.9241
0.3966	5.0	820	0.4280	0.9165
0.411	6.0	984	0.3949	0.9241
0.3804	7.0	1148	0.3876	0.9208
0.3808	8.0	1312	0.4051	0.9089
0.3966	9.0	1476	0.3958	0.9165
0.3856	10.0	1640	0.3938	0.9241
0.3471	11.0	1804	0.3890	0.9273
0.3597	12.0	1968	0.3919	0.9165
0.3558	13.0	2132	0.3905	0.9295
0.3726	14.0	2296	0.3947	0.9252
0.3589	15.0	2460	0.3997	0.9230
0.3552	16.0	2624	0.3923	0.9230
0.332	17.0	2788	0.3911	0.9197
0.3374	18.0	2952	0.3916	0.9208
0.3353	19.0	3116	0.3924	0.9252
0.3368	20.0	3280	0.3923	0.9230

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.0+cu124
• Datasets 3.0.1
• Tokenizers 0.20.1