vit-mae-base-effusion-classifier

This model is a fine-tuned version of facebook/vit-mae-base on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.4179
• Accuracy: 0.8174

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-06
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6554	1.0	362	0.6692	0.6030
0.569	2.0	725	0.5891	0.7023
0.6098	3.0	1088	0.5421	0.7367
0.4984	4.0	1451	0.5668	0.7043
0.4884	5.0	1813	0.6061	0.6844
0.4351	6.0	2176	0.4481	0.8098
0.4794	7.0	2539	0.4384	0.8084
0.4636	8.0	2902	0.4343	0.8077
0.4816	9.0	3264	0.5363	0.7491
0.5016	10.0	3627	0.4993	0.7677
0.4826	11.0	3990	0.4483	0.8043
0.4707	12.0	4353	0.4249	0.8112
0.4483	13.0	4715	0.4193	0.8160
0.419	14.0	5078	0.4146	0.8215
0.5039	15.0	5441	0.4188	0.8181
0.4111	16.0	5804	0.4459	0.8112
0.3293	17.0	6166	0.4228	0.8181
0.4171	18.0	6529	0.4239	0.8215
0.3375	19.0	6892	0.4162	0.8215
0.32	19.96	7240	0.4179	0.8174

Framework versions

• Transformers 4.39.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2