limbxy_pose_2heads_1layers_128embeddim

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: 0.1004
• Rmse: 0.3168

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
0.1305	1.0	89	0.1136	0.3371
0.1225	2.0	178	0.1131	0.3364
0.1121	3.0	267	0.1023	0.3198
0.1148	4.0	356	0.1069	0.3270
0.1199	5.0	445	0.1202	0.3467
0.1065	6.0	534	0.1013	0.3183
0.1051	7.0	623	0.1025	0.3201
0.1023	8.0	712	0.1106	0.3326
0.105	9.0	801	0.1151	0.3392
0.1018	10.0	890	0.1053	0.3245
0.1063	11.0	979	0.1008	0.3176
0.0982	12.0	1068	0.1076	0.3280
0.1007	13.0	1157	0.1014	0.3185
0.0996	14.0	1246	0.1047	0.3236
0.0996	15.0	1335	0.1037	0.3221
0.104	16.0	1424	0.1019	0.3192
0.0985	17.0	1513	0.1007	0.3173
0.0972	18.0	1602	0.1008	0.3174
0.0984	19.0	1691	0.1008	0.3175
0.0973	20.0	1780	0.1004	0.3168

Framework versions

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