vit-msn-small-lipid-invalidation-nobg

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

• Loss: 0.1866
• Accuracy: 0.9235

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
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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
No log	1.0	6	0.5868	0.7059
0.583	2.0	12	0.2850	0.8706
0.583	3.0	18	0.3167	0.8647
0.4996	4.0	24	0.3738	0.8294
0.3097	5.0	30	0.2105	0.9176
0.3097	6.0	36	0.2695	0.8824
0.2893	7.0	42	0.2304	0.8941
0.2893	8.0	48	0.2978	0.8529
0.3305	9.0	54	0.2952	0.8882
0.2775	10.0	60	0.2661	0.8882
0.2775	11.0	66	0.2533	0.8824
0.223	12.0	72	0.2586	0.8882
0.223	13.0	78	0.2034	0.9176
0.216	14.0	84	0.1866	0.9235
0.1858	15.0	90	0.2631	0.8882
0.1858	16.0	96	0.2440	0.8882
0.1866	17.0	102	0.2480	0.8941
0.1866	18.0	108	0.3897	0.8647
0.1701	19.0	114	0.3999	0.8647
0.1717	20.0	120	0.3307	0.8706

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.3.2
• Tokenizers 0.19.1