limbxy_seq_t2_heads2_layers2

This model is a fine-tuned version of c14kevincardenas/beit-large-patch16-384-limb on the c14kevincardenas/beta_caller_284_limbxy_seq_2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0051
• Rmse: 0.0715

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: 32
• eval_batch_size: 32
• 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: 50.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rmse
0.017	1.0	150	0.0194	0.1392
0.0166	2.0	300	0.0165	0.1283
0.0114	3.0	450	0.0105	0.1023
0.0052	4.0	600	0.0078	0.0884
0.0056	5.0	750	0.0067	0.0816
0.0038	6.0	900	0.0077	0.0879
0.0027	7.0	1050	0.0072	0.0849
0.0023	8.0	1200	0.0064	0.0798
0.002	9.0	1350	0.0071	0.0843
0.0022	10.0	1500	0.0063	0.0793
0.0019	11.0	1650	0.0058	0.0761
0.0018	12.0	1800	0.0065	0.0804
0.0011	13.0	1950	0.0055	0.0741
0.0011	14.0	2100	0.0057	0.0754
0.001	15.0	2250	0.0059	0.0769
0.001	16.0	2400	0.0062	0.0787
0.0008	17.0	2550	0.0055	0.0740
0.0007	18.0	2700	0.0057	0.0753
0.001	19.0	2850	0.0056	0.0751
0.0006	20.0	3000	0.0055	0.0740
0.0007	21.0	3150	0.0057	0.0752
0.0006	22.0	3300	0.0055	0.0741
0.0005	23.0	3450	0.0055	0.0741
0.0006	24.0	3600	0.0053	0.0726
0.0004	25.0	3750	0.0055	0.0743
0.0004	26.0	3900	0.0054	0.0734
0.0005	27.0	4050	0.0054	0.0737
0.0004	28.0	4200	0.0055	0.0742
0.0004	29.0	4350	0.0053	0.0726
0.0005	30.0	4500	0.0055	0.0745
0.0004	31.0	4650	0.0052	0.0720
0.0003	32.0	4800	0.0054	0.0735
0.0002	33.0	4950	0.0053	0.0727
0.0003	34.0	5100	0.0052	0.0724
0.0003	35.0	5250	0.0052	0.0721
0.0003	36.0	5400	0.0053	0.0727
0.0002	37.0	5550	0.0053	0.0729
0.0002	38.0	5700	0.0052	0.0721
0.0002	39.0	5850	0.0052	0.0724
0.0002	40.0	6000	0.0051	0.0717
0.0002	41.0	6150	0.0052	0.0720
0.0002	42.0	6300	0.0052	0.0720
0.0002	43.0	6450	0.0052	0.0719
0.0001	44.0	6600	0.0052	0.0721
0.0001	45.0	6750	0.0051	0.0717
0.0001	46.0	6900	0.0052	0.0720
0.0001	47.0	7050	0.0052	0.0719
0.0001	48.0	7200	0.0051	0.0716
0.0001	49.0	7350	0.0051	0.0715
0.0001	50.0	7500	0.0051	0.0716

Framework versions

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