c14kevincardenas/beit-base-patch16-224_alpha0.7_temp3.0_t2

beit-base-patch16-224_alpha0.7_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: 0.5565
• Accuracy: 0.8142

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
1.1373	1.0	90	1.4268	0.2806
0.902	2.0	180	1.1061	0.5326
0.5748	3.0	270	0.7085	0.7401
0.3815	4.0	360	0.6586	0.7688
0.2747	5.0	450	0.5978	0.7925
0.2098	6.0	540	0.6238	0.7895
0.1954	7.0	630	0.6208	0.7747
0.182	8.0	720	0.5785	0.8162
0.1748	9.0	810	0.5769	0.7994
0.1568	10.0	900	0.5669	0.8083
0.1542	11.0	990	0.5794	0.7994
0.1417	12.0	1080	0.5789	0.7935
0.1335	13.0	1170	0.5565	0.8142
0.126	14.0	1260	0.5735	0.8162
0.1277	15.0	1350	0.5776	0.7964
0.1226	16.0	1440	0.5593	0.8034
0.1238	17.0	1530	0.5681	0.8083
0.1149	18.0	1620	0.5669	0.8063
0.1156	19.0	1710	0.5673	0.8103
0.1127	20.0	1800	0.5651	0.8103

Framework versions

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