beit-base-patch16-224_alpha0.7_temp3.0_t3

This model is a fine-tuned version of c14kevincardenas/ClimBEiT-t3 on the c14kevincardenas/beta_caller_284_person_crop_seq_withlimb dataset. It achieves the following results on the evaluation set:

• Loss: 0.7372
• Accuracy: 0.8221

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: 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
0.5666	1.0	164	1.4207	0.2560
0.5216	2.0	328	1.3016	0.3774
0.3815	3.0	492	0.9241	0.6779
0.324	4.0	656	0.8860	0.6996
0.2839	5.0	820	0.7562	0.7831
0.263	6.0	984	0.7733	0.7874
0.2514	7.0	1148	0.7520	0.7939
0.2459	8.0	1312	0.7618	0.7961
0.2449	9.0	1476	0.7675	0.7983
0.2312	10.0	1640	0.7584	0.8145
0.2297	11.0	1804	0.7757	0.8037
0.2322	12.0	1968	0.7372	0.8221
0.2239	13.0	2132	0.7653	0.8134
0.2282	14.0	2296	0.7397	0.8232
0.2291	15.0	2460	0.7518	0.8048
0.2243	16.0	2624	0.7498	0.8134
0.2285	17.0	2788	0.7431	0.8243
0.2191	18.0	2952	0.7446	0.8243
0.2238	19.0	3116	0.7432	0.8232
0.2187	20.0	3280	0.7468	0.8243

Framework versions

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