limbxy_pose_2heads_1layers_64embeddim

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: 0.0148
• Rmse: 0.1218

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
0.2345	1.0	89	0.2326	0.4823
0.2345	2.0	178	0.2326	0.4823
0.2177	3.0	267	0.1564	0.3954
0.2261	4.0	356	0.1935	0.4399
0.1607	5.0	445	0.1622	0.4028
0.1566	6.0	534	0.1466	0.3829
0.1579	7.0	623	0.1461	0.3822
0.1596	8.0	712	0.1493	0.3863
0.1524	9.0	801	0.1456	0.3816
0.1477	10.0	890	0.1447	0.3804
0.1618	11.0	979	0.1677	0.4095
0.0386	12.0	1068	0.0263	0.1620
0.0246	13.0	1157	0.0348	0.1864
0.0254	14.0	1246	0.0181	0.1344
0.0194	15.0	1335	0.0156	0.1250
0.0166	16.0	1424	0.0175	0.1323
0.0169	17.0	1513	0.0152	0.1234
0.0148	18.0	1602	0.0196	0.1399
0.0148	19.0	1691	0.0164	0.1279
0.0141	20.0	1780	0.0148	0.1218

Framework versions

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