c14kevincardenas/beit-base-patch16-224_alpha0.5_temp3.0_t2

beit-base-patch16-224_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: 0.5297
• 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
1.2129	1.0	90	1.4057	0.2787
0.9595	2.0	180	1.0680	0.5435
0.6331	3.0	270	0.6700	0.7480
0.4388	4.0	360	0.6401	0.7727
0.312	5.0	450	0.5945	0.7767
0.2499	6.0	540	0.6064	0.7796
0.2294	7.0	630	0.5901	0.7945
0.2119	8.0	720	0.5700	0.8063
0.2002	9.0	810	0.6045	0.7866
0.1878	10.0	900	0.5636	0.7984
0.1851	11.0	990	0.5647	0.7885
0.1782	12.0	1080	0.5441	0.8103
0.1647	13.0	1170	0.5367	0.8123
0.158	14.0	1260	0.5515	0.8043
0.1592	15.0	1350	0.5537	0.7984
0.1565	16.0	1440	0.5411	0.8083
0.1561	17.0	1530	0.5297	0.8142
0.1447	18.0	1620	0.5404	0.8123
0.148	19.0	1710	0.5337	0.8123
0.1442	20.0	1800	0.5359	0.8123

Framework versions

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