deit-small-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: 0.6661
• Accuracy: 0.7961

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
0.8116	1.0	164	1.4057	0.2516
0.7514	2.0	328	1.3140	0.3416
0.582	3.0	492	0.8813	0.6985
0.4397	4.0	656	0.7596	0.7570
0.3581	5.0	820	0.7805	0.7451
0.3144	6.0	984	0.7793	0.7646
0.3032	7.0	1148	0.7034	0.7744
0.2971	8.0	1312	0.6883	0.7777
0.2927	9.0	1476	0.6818	0.7831
0.277	10.0	1640	0.6825	0.7852
0.2782	11.0	1804	0.6809	0.7993
0.2824	12.0	1968	0.6774	0.8026
0.2723	13.0	2132	0.6776	0.8004
0.2768	14.0	2296	0.6661	0.7961
0.2802	15.0	2460	0.6797	0.7993
0.2767	16.0	2624	0.6695	0.7972
0.2803	17.0	2788	0.6697	0.7961
0.2701	18.0	2952	0.6696	0.8037
0.2774	19.0	3116	0.6698	0.8048
0.2699	20.0	3280	0.6688	0.7993

Framework versions

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