finetuned-beit-limb-seq-t2-2heads-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_2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3945
• Accuracy: 0.9150

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
0.6204	1.0	180	0.4733	0.9002
0.4236	2.0	360	0.4528	0.8992
0.4123	3.0	540	0.4366	0.9002
0.4292	4.0	720	0.4044	0.9081
0.3976	5.0	900	0.4197	0.9101
0.4198	6.0	1080	0.4055	0.9081
0.4067	7.0	1260	0.4086	0.9091
0.4176	8.0	1440	0.4073	0.9081
0.3967	9.0	1620	0.4019	0.9130
0.3984	10.0	1800	0.4045	0.9121
0.3955	11.0	1980	0.4085	0.9101
0.3791	12.0	2160	0.4016	0.9111
0.3708	13.0	2340	0.3992	0.9150
0.383	14.0	2520	0.3989	0.9140
0.3347	15.0	2700	0.4061	0.9180
0.3367	16.0	2880	0.3990	0.9101
0.3203	17.0	3060	0.4040	0.9101
0.349	18.0	3240	0.3945	0.9150
0.3384	19.0	3420	0.4006	0.9081
0.3466	20.0	3600	0.3997	0.9101

Framework versions

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