finetuned-beit-limb-seq-t4-4heads-2layers-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_4 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3914
• 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
0.7344	1.0	148	0.4724	0.9006
0.437	2.0	296	0.4584	0.8910
0.5132	3.0	444	0.4280	0.9042
0.4165	4.0	592	0.4292	0.9102
0.4309	5.0	740	0.3967	0.9210
0.4253	6.0	888	0.4209	0.9078
0.4175	7.0	1036	0.4076	0.9102
0.403	8.0	1184	0.4002	0.9174
0.4628	9.0	1332	0.3954	0.9162
0.3802	10.0	1480	0.3914	0.9150
0.3973	11.0	1628	0.3952	0.9222
0.3902	12.0	1776	0.4013	0.9138
0.3726	13.0	1924	0.4178	0.9114
0.3913	14.0	2072	0.4156	0.9102
0.3621	15.0	2220	0.4171	0.9102
0.3642	16.0	2368	0.3965	0.9210
0.3344	17.0	2516	0.4092	0.9138
0.3723	18.0	2664	0.4121	0.9126
0.3767	19.0	2812	0.4041	0.9174
0.3037	20.0	2960	0.4081	0.9174

Framework versions

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