limbxy_pose_4heads_1layers_16embeddim

This model is a fine-tuned version of c14kevincardenas/beit-large-patch16-384-limb on the c14kevincardenas/beta_caller_284_limbxy_pose dataset. It achieves the following results on the evaluation set:

• Loss: 0.1449
• Rmse: 0.3806

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: 5e-05
• train_batch_size: 64
• eval_batch_size: 64
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Rmse
0.2128	1.0	89	0.1745	0.4177
0.1574	2.0	178	0.1486	0.3855
0.2045	3.0	267	0.1519	0.3897
0.1697	4.0	356	0.1632	0.4040
0.1818	5.0	445	0.1949	0.4414
0.1624	6.0	534	0.1475	0.3841
0.1645	7.0	623	0.1484	0.3852
0.1655	8.0	712	0.1471	0.3835
0.1594	9.0	801	0.1535	0.3918
0.1513	10.0	890	0.1449	0.3806
0.1488	11.0	979	0.1455	0.3814
0.1507	12.0	1068	0.1536	0.3919
0.1522	13.0	1157	0.1449	0.3807
0.1458	14.0	1246	0.1453	0.3811
0.1506	15.0	1335	0.1455	0.3815
0.1505	16.0	1424	0.1452	0.3810
0.1463	17.0	1513	0.1449	0.3807
0.1463	18.0	1602	0.1449	0.3806
0.1494	19.0	1691	0.1457	0.3816
0.1454	20.0	1780	0.1451	0.3810

Framework versions

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