finetuned-beit-limb-seq-t3-4heads-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: 0.3860
• Accuracy: 0.9252

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
0.6998	1.0	164	0.4404	0.9078
0.4207	2.0	328	0.4476	0.9187
0.439	3.0	492	0.4073	0.9132
0.4255	4.0	656	0.3950	0.9187
0.3913	5.0	820	0.3998	0.9197
0.4077	6.0	984	0.3928	0.9230
0.3771	7.0	1148	0.3860	0.9252
0.3788	8.0	1312	0.4121	0.9143
0.3956	9.0	1476	0.4081	0.9154
0.387	10.0	1640	0.3888	0.9262
0.3517	11.0	1804	0.3973	0.9197
0.3631	12.0	1968	0.3963	0.9176
0.3563	13.0	2132	0.3998	0.9252
0.3644	14.0	2296	0.3945	0.9208
0.3712	15.0	2460	0.3941	0.9230
0.3593	16.0	2624	0.3909	0.9241
0.3397	17.0	2788	0.3887	0.9241
0.3412	18.0	2952	0.3872	0.9252
0.3423	19.0	3116	0.3879	0.9230
0.3399	20.0	3280	0.3879	0.9295

Framework versions

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