limbxy_hands

This model is a fine-tuned version of c14kevincardenas/beit-large-patch16-384-limb on the c14kevincardenas/beta_caller_284_limbxy_hands dataset. It achieves the following results on the evaluation set:

• Loss: 0.0048
• Rmse: 0.0691

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
0.1563	1.0	51	0.1439	0.3793
0.066	2.0	102	0.0629	0.2507
0.031	3.0	153	0.0240	0.1550
0.0189	4.0	204	0.0222	0.1491
0.0362	5.0	255	0.0700	0.2646
0.0106	6.0	306	0.0118	0.1088
0.0217	7.0	357	0.0419	0.2048
0.0167	8.0	408	0.0170	0.1303
0.0078	9.0	459	0.0111	0.1052
0.0102	10.0	510	0.0158	0.1256
0.009	11.0	561	0.0064	0.0802
0.0067	12.0	612	0.0074	0.0862
0.0047	13.0	663	0.0055	0.0739
0.0036	14.0	714	0.0068	0.0827
0.0072	15.0	765	0.0075	0.0869
0.006	16.0	816	0.0055	0.0739
0.0022	17.0	867	0.0052	0.0721
0.0051	18.0	918	0.0054	0.0738
0.0023	19.0	969	0.0051	0.0714
0.0016	20.0	1020	0.0048	0.0691

Framework versions

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