c14kevincardenas/beit-base-patch16-384_alpha0.5_temp3.0_t2

beit-base-patch16-384_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: 0.4789
• Accuracy: 0.8449

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
1.1985	1.0	90	1.3602	0.3340
0.8935	2.0	180	0.9506	0.5988
0.5967	3.0	270	0.6939	0.7322
0.4388	4.0	360	0.5797	0.7925
0.3085	5.0	450	0.5758	0.7964
0.2767	6.0	540	0.5370	0.8211
0.2428	7.0	630	0.5692	0.8063
0.2204	8.0	720	0.5434	0.8113
0.2009	9.0	810	0.5459	0.8202
0.1867	10.0	900	0.5644	0.8172
0.1866	11.0	990	0.5078	0.8399
0.1795	12.0	1080	0.5015	0.8261
0.1782	13.0	1170	0.5340	0.8152
0.1638	14.0	1260	0.5085	0.8389
0.1633	15.0	1350	0.5042	0.8468
0.1604	16.0	1440	0.4789	0.8449
0.1543	17.0	1530	0.4908	0.8429
0.1507	18.0	1620	0.4911	0.8399
0.1537	19.0	1710	0.4875	0.8449
0.1492	20.0	1800	0.4885	0.8429

Framework versions

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