efficient-vit-CEMEDE-og

This model is a fine-tuned version of timm/efficientvit_b0.r224_in1k on the cemede dataset. It achieves the following results on the evaluation set:

• Loss: 0.7816
• Accuracy: 0.7913
• F1: 0.7468

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: 0.0002
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.3755	0.0840	100	2.1189	0.3766	0.2053
0.8418	0.1679	200	2.2810	0.3067	0.1753
0.7683	0.2519	300	1.8360	0.4570	0.3601
0.6862	0.3359	400	1.1806	0.6685	0.5573
0.4868	0.4198	500	1.0167	0.6855	0.6038
0.3877	0.5038	600	1.0808	0.6763	0.5733
0.4511	0.5877	700	1.2495	0.6708	0.5675
0.2773	0.6717	800	0.9806	0.7260	0.5884
0.3842	0.7557	900	0.9603	0.6961	0.5869
0.5853	0.8396	1000	1.1610	0.7113	0.6741
0.375	0.9236	1100	1.0845	0.6874	0.6517
0.4001	1.0076	1200	0.7816	0.7913	0.7468
0.2953	1.0915	1300	0.9925	0.7398	0.6851
0.1258	1.1755	1400	0.8799	0.7834	0.7269
0.2446	1.2594	1500	0.8710	0.7646	0.7369
0.2414	1.3434	1600	0.9198	0.7664	0.7199
0.1738	1.4274	1700	0.8230	0.7972	0.7353
0.2122	1.5113	1800	0.9513	0.7664	0.7091
0.153	1.5953	1900	0.8115	0.7766	0.7432
0.159	1.6793	2000	0.9063	0.7940	0.7453
0.1874	1.7632	2100	0.9330	0.8014	0.7365
0.1999	1.8472	2200	0.8745	0.7692	0.6915

Framework versions

• Transformers 4.55.4
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.21.4