deit-tiny-distilled-patch16-224_alpha0.5_temp5.0_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: 0.7113
• Accuracy: 0.7874

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
0.8044	1.0	164	1.3844	0.2701
0.7648	2.0	328	1.3690	0.2798
0.6383	3.0	492	1.0005	0.6399
0.4941	4.0	656	0.8374	0.6996
0.4194	5.0	820	0.7788	0.7462
0.3512	6.0	984	0.7797	0.7462
0.326	7.0	1148	0.7322	0.7636
0.309	8.0	1312	0.7521	0.7538
0.3064	9.0	1476	0.7324	0.7701
0.2878	10.0	1640	0.7404	0.7549
0.2859	11.0	1804	0.7296	0.7711
0.2909	12.0	1968	0.7288	0.7766
0.2794	13.0	2132	0.7159	0.7820
0.2835	14.0	2296	0.7187	0.7766
0.2845	15.0	2460	0.7233	0.7842
0.2821	16.0	2624	0.7119	0.7863
0.2849	17.0	2788	0.7158	0.7787
0.2744	18.0	2952	0.7113	0.7874
0.282	19.0	3116	0.7140	0.7831
0.2743	20.0	3280	0.7148	0.7820

Framework versions

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