c14kevincardenas/swin-tiny-patch4-window7-224_alpha0.5_temp3.0_t2

swin-tiny-patch4-window7-224_alpha0.5_temp3.0

This model is a fine-tuned version of c14kevincardenas/ClimBEiTv2 on the c14kevincardenas/beta_caller_284_person_crop_seq_withlimb_2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5948
• Accuracy: 0.7925

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: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.2209	1.0	90	1.4077	0.2727
1.0885	2.0	180	1.2163	0.4407
0.8453	3.0	270	0.9650	0.5850
0.6364	4.0	360	0.7619	0.7016
0.4816	5.0	450	0.6795	0.7480
0.3772	6.0	540	0.6776	0.7599
0.3145	7.0	630	0.6579	0.7638
0.2638	8.0	720	0.6325	0.7727
0.2466	9.0	810	0.6574	0.7767
0.2318	10.0	900	0.6499	0.7628
0.2069	11.0	990	0.6260	0.7856
0.1938	12.0	1080	0.6059	0.7935
0.1873	13.0	1170	0.6024	0.7895
0.1824	14.0	1260	0.6062	0.7826
0.1818	15.0	1350	0.5969	0.7955
0.1749	16.0	1440	0.6104	0.7866
0.1726	17.0	1530	0.6117	0.7945
0.1607	18.0	1620	0.5983	0.7915
0.1635	19.0	1710	0.5950	0.7915
0.1607	20.0	1800	0.5948	0.7925

Framework versions

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