vit-msn-small-ultralytics_yolo_cropped_lateral_flow_ivalidation

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: 0.1774
• Accuracy: 0.9489

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9231	3	0.8678	0.4277
No log	1.8462	6	0.6171	0.7
No log	2.7692	9	0.4174	0.8723
0.6518	4.0	13	0.5366	0.7106
0.6518	4.9231	16	0.3255	0.8851
0.6518	5.8462	19	0.6159	0.6809
0.4119	6.7692	22	0.3017	0.9191
0.4119	8.0	26	0.5130	0.7128
0.4119	8.9231	29	0.2183	0.9255
0.3387	9.8462	32	0.2523	0.9149
0.3387	10.7692	35	0.1774	0.9489
0.3387	12.0	39	0.2376	0.9255
0.3055	12.9231	42	0.3930	0.8383
0.3055	13.8462	45	0.2308	0.9234
0.3055	14.7692	48	0.1587	0.9468
0.2909	16.0	52	0.6113	0.6830
0.2909	16.9231	55	0.2910	0.8915
0.2909	17.8462	58	0.3612	0.8447
0.2227	18.7692	61	0.3117	0.8787
0.2227	20.0	65	0.2684	0.9170
0.2227	20.9231	68	0.3767	0.8404
0.2129	21.8462	71	0.2527	0.9234
0.2129	22.7692	74	0.3270	0.8745
0.2129	24.0	78	0.4314	0.8064
0.213	24.9231	81	0.2874	0.9
0.213	25.8462	84	0.4797	0.7894
0.213	26.7692	87	0.4896	0.7851
0.1758	28.0	91	0.3144	0.8723
0.1758	28.9231	94	0.5881	0.7213
0.1758	29.8462	97	0.5599	0.7298
0.1766	30.7692	100	0.3413	0.8702
0.1766	32.0	104	0.3453	0.8638
0.1766	32.9231	107	0.3634	0.8596
0.1583	33.8462	110	0.3799	0.8468
0.1583	34.7692	113	0.3840	0.8468
0.1583	36.0	117	0.3890	0.8447
0.1969	36.9231	120	0.3950	0.8426

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1