limbxy_pose_8heads_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: 0.1402
• Rmse: 0.3745

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
0.3414	1.0	89	0.3311	0.5754
0.1834	2.0	178	0.2026	0.4501
0.1645	3.0	267	0.1491	0.3862
0.1729	4.0	356	0.2542	0.5042
0.1612	5.0	445	0.1539	0.3923
0.1578	6.0	534	0.1464	0.3827
0.1594	7.0	623	0.1860	0.4313
0.1546	8.0	712	0.1433	0.3785
0.1517	9.0	801	0.1416	0.3763
0.1461	10.0	890	0.1576	0.3969
0.1519	11.0	979	0.1623	0.4029
0.1491	12.0	1068	0.1411	0.3756
0.1489	13.0	1157	0.1416	0.3763
0.1425	14.0	1246	0.1426	0.3776
0.145	15.0	1335	0.1407	0.3751
0.1418	16.0	1424	0.1443	0.3799
0.1411	17.0	1513	0.1403	0.3745
0.1398	18.0	1602	0.1403	0.3746
0.143	19.0	1691	0.1405	0.3749
0.1395	20.0	1780	0.1402	0.3745

Framework versions

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