vit-bmr-tuned

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.3340
• Accuracy: 0.8889
• Precision: 0.8905
• Recall: 0.8889
• F1: 0.8885

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	1.0	160	0.3340	0.8889	0.8905	0.8889	0.8885
No log	2.0	320	0.4336	0.8765	0.8860	0.8765	0.8764
No log	3.0	480	0.4304	0.9136	0.9194	0.9136	0.9129
0.1675	4.0	640	0.6488	0.8765	0.8907	0.8765	0.8745
0.1675	5.0	800	0.5233	0.9012	0.9046	0.9012	0.9007
0.1675	6.0	960	0.5633	0.9012	0.9046	0.9012	0.9007
0.0141	7.0	1120	0.5923	0.9012	0.9046	0.9012	0.9007
0.0141	8.0	1280	0.6144	0.9012	0.9046	0.9012	0.9007
0.0141	9.0	1440	0.6319	0.9012	0.9046	0.9012	0.9007
0.0016	10.0	1600	0.6455	0.9012	0.9046	0.9012	0.9007
0.0016	11.0	1760	0.6563	0.9012	0.9046	0.9012	0.9007
0.0016	12.0	1920	0.6655	0.9012	0.9046	0.9012	0.9007
0.0011	13.0	2080	0.6731	0.9012	0.9046	0.9012	0.9007
0.0011	14.0	2240	0.6786	0.9012	0.9046	0.9012	0.9007
0.0011	15.0	2400	0.6830	0.9012	0.9046	0.9012	0.9007
0.0009	16.0	2560	0.6863	0.9012	0.9046	0.9012	0.9007
0.0009	17.0	2720	0.6883	0.9012	0.9046	0.9012	0.9007
0.0009	18.0	2880	0.6895	0.9012	0.9046	0.9012	0.9007
0.0008	19.0	3040	0.6899	0.9012	0.9046	0.9012	0.9007
0.0008	20.0	3200	0.6900	0.9012	0.9046	0.9012	0.9007

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1