finetuned-beit-limb-seq-t3-2heads-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_3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3835
• Accuracy: 0.9262

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
0.4747	1.0	164	0.4114	0.9197
0.4009	2.0	328	0.4908	0.8980
0.4301	3.0	492	0.4505	0.9078
0.4405	4.0	656	0.4064	0.9187
0.4124	5.0	820	0.4294	0.9046
0.4141	6.0	984	0.3915	0.9154
0.3795	7.0	1148	0.3918	0.9230
0.4089	8.0	1312	0.4023	0.9165
0.414	9.0	1476	0.4032	0.9219
0.4079	10.0	1640	0.3835	0.9262
0.343	11.0	1804	0.4039	0.9197
0.3617	12.0	1968	0.4090	0.9089
0.361	13.0	2132	0.4000	0.9219
0.3534	14.0	2296	0.4115	0.9176
0.3575	15.0	2460	0.4194	0.9187
0.3462	16.0	2624	0.4013	0.9197
0.3397	17.0	2788	0.3938	0.9219
0.339	18.0	2952	0.3948	0.9241
0.3236	19.0	3116	0.3971	0.9197
0.3178	20.0	3280	0.3969	0.9219

Framework versions

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