finetuned-beit-limb-seq-t3-8heads-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: 0.3853
• Accuracy: 0.9230

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
0.6888	1.0	164	0.4465	0.9035
0.423	2.0	328	0.4432	0.9197
0.433	3.0	492	0.4078	0.9143
0.4242	4.0	656	0.3976	0.9143
0.3916	5.0	820	0.4038	0.9208
0.4067	6.0	984	0.3938	0.9219
0.3767	7.0	1148	0.3853	0.9230
0.3796	8.0	1312	0.4035	0.9121
0.3977	9.0	1476	0.4074	0.9143
0.3875	10.0	1640	0.3893	0.9219
0.3504	11.0	1804	0.3970	0.9197
0.3608	12.0	1968	0.3970	0.9176
0.3585	13.0	2132	0.3972	0.9241
0.3663	14.0	2296	0.3944	0.9208
0.3699	15.0	2460	0.3938	0.9208
0.3586	16.0	2624	0.3915	0.9241
0.3401	17.0	2788	0.3873	0.9219
0.3397	18.0	2952	0.3860	0.9230
0.3459	19.0	3116	0.3874	0.9241
0.3419	20.0	3280	0.3870	0.9273

Framework versions

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