deit-small-distilled-patch16-224_alpha0.7_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: 0.8069
• Accuracy: 0.7581

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
0.5682	1.0	164	1.4102	0.2473
0.532	2.0	328	1.3764	0.2766
0.4029	3.0	492	0.9531	0.6540
0.3205	4.0	656	0.8554	0.7289
0.2769	5.0	820	0.8604	0.7419
0.2493	6.0	984	0.8398	0.7332
0.2375	7.0	1148	0.8179	0.7516
0.2319	8.0	1312	0.8069	0.7581
0.2311	9.0	1476	0.8200	0.7473
0.2183	10.0	1640	0.8133	0.7549
0.2185	11.0	1804	0.8225	0.7462
0.2225	12.0	1968	0.8251	0.7505
0.2146	13.0	2132	0.8196	0.7549
0.2189	14.0	2296	0.8094	0.7560
0.2217	15.0	2460	0.8287	0.7484
0.2187	16.0	2624	0.8123	0.7636
0.2216	17.0	2788	0.8116	0.7614
0.2129	18.0	2952	0.8123	0.7646
0.219	19.0	3116	0.8127	0.7636
0.2125	20.0	3280	0.8113	0.7614

Framework versions

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