limbxy_pose_2heads_1layers_8embeddim

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: 0.0077
• Rmse: 0.0876

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
0.174	1.0	89	0.1471	0.3835
0.1559	2.0	178	0.1466	0.3829
0.1485	3.0	267	0.1661	0.4075
0.1624	4.0	356	0.1418	0.3765
0.1457	5.0	445	0.1437	0.3790
0.1635	6.0	534	0.1424	0.3773
0.1428	7.0	623	0.1584	0.3980
0.1481	8.0	712	0.1408	0.3753
0.1494	9.0	801	0.1478	0.3845
0.1417	10.0	890	0.1545	0.3930
0.1421	11.0	979	0.1432	0.3785
0.145	12.0	1068	0.1403	0.3745
0.1466	13.0	1157	0.1443	0.3799
0.0601	14.0	1246	0.0208	0.1443
0.0154	15.0	1335	0.0124	0.1115
0.0102	16.0	1424	0.0128	0.1133
0.0071	17.0	1513	0.0129	0.1137
0.0076	18.0	1602	0.0085	0.0920
0.0057	19.0	1691	0.0079	0.0888
0.0046	20.0	1780	0.0077	0.0876

Framework versions

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