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

• Loss: 0.3913
• Accuracy: 0.9234

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
1.0589	1.0	148	0.5868	0.8754
0.429	2.0	296	0.4257	0.8970
0.4438	3.0	444	0.4207	0.9078
0.4218	4.0	592	0.4250	0.9006
0.4232	5.0	740	0.3950	0.9198
0.4047	6.0	888	0.3974	0.9198
0.3966	7.0	1036	0.3983	0.9174
0.394	8.0	1184	0.4040	0.9162
0.4226	9.0	1332	0.3954	0.9186
0.3786	10.0	1480	0.3975	0.9174
0.3887	11.0	1628	0.4018	0.9150
0.3966	12.0	1776	0.3930	0.9174
0.382	13.0	1924	0.4001	0.9162
0.3805	14.0	2072	0.4116	0.9138
0.3781	15.0	2220	0.4034	0.9174
0.3796	16.0	2368	0.3913	0.9234
0.3522	17.0	2516	0.3978	0.9210
0.3799	18.0	2664	0.3995	0.9198
0.3741	19.0	2812	0.3941	0.9186
0.3208	20.0	2960	0.3964	0.9222

Framework versions

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