vit-base-kidney-stone-4-Jonathan_El-Beze_-w256_1k_v1-_SEC

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: 0.2940
• Accuracy: 0.9242
• Precision: 0.9321
• Recall: 0.9242
• F1: 0.9251

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: 16
• 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: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.1207	0.3333	100	0.5525	0.8333	0.8760	0.8333	0.8303
0.0178	0.6667	200	0.3368	0.8883	0.9298	0.8883	0.8927
0.0396	1.0	300	0.3187	0.9108	0.9213	0.9108	0.9104
0.0074	1.3333	400	1.1846	0.7583	0.8167	0.7583	0.7339
0.0125	1.6667	500	0.2940	0.9242	0.9321	0.9242	0.9251
0.0029	2.0	600	0.5031	0.8958	0.9051	0.8958	0.8929
0.0021	2.3333	700	0.5150	0.9008	0.9114	0.9008	0.8977
0.0016	2.6667	800	0.4894	0.9092	0.9191	0.9092	0.9069
0.0013	3.0	900	0.5048	0.9092	0.9194	0.9092	0.9067
0.0011	3.3333	1000	0.5066	0.9092	0.9187	0.9092	0.9070
0.001	3.6667	1100	0.5179	0.9092	0.9189	0.9092	0.9070
0.0008	4.0	1200	0.5369	0.9092	0.9198	0.9092	0.9069
0.0007	4.3333	1300	0.5459	0.9092	0.9198	0.9092	0.9069
0.0006	4.6667	1400	0.5508	0.9092	0.9198	0.9092	0.9069
0.0006	5.0	1500	0.5557	0.91	0.9203	0.91	0.9079
0.0005	5.3333	1600	0.5605	0.9108	0.9210	0.9108	0.9088
0.0004	5.6667	1700	0.5647	0.9108	0.9210	0.9108	0.9088
0.0004	6.0	1800	0.5735	0.9108	0.9210	0.9108	0.9088
0.0004	6.3333	1900	0.5797	0.9108	0.9210	0.9108	0.9088
0.0003	6.6667	2000	0.5840	0.9108	0.9210	0.9108	0.9088
0.0003	7.0	2100	0.5877	0.9108	0.9210	0.9108	0.9088
0.0003	7.3333	2200	0.5942	0.9108	0.9210	0.9108	0.9088
0.0003	7.6667	2300	0.6003	0.9117	0.9222	0.9117	0.9096
0.0003	8.0	2400	0.5999	0.9108	0.9210	0.9108	0.9088
0.0002	8.3333	2500	0.6042	0.91	0.9203	0.91	0.9080
0.0002	8.6667	2600	0.6076	0.9108	0.9215	0.9108	0.9088
0.0002	9.0	2700	0.6098	0.9108	0.9210	0.9108	0.9088
0.0002	9.3333	2800	0.6135	0.9108	0.9215	0.9108	0.9088
0.0002	9.6667	2900	0.6157	0.9108	0.9215	0.9108	0.9088
0.0002	10.0	3000	0.6191	0.9108	0.9215	0.9108	0.9088
0.0002	10.3333	3100	0.6216	0.9108	0.9215	0.9108	0.9088
0.0002	10.6667	3200	0.6241	0.9108	0.9215	0.9108	0.9088
0.0002	11.0	3300	0.6265	0.9108	0.9215	0.9108	0.9088
0.0002	11.3333	3400	0.6291	0.9108	0.9215	0.9108	0.9088
0.0001	11.6667	3500	0.6308	0.9108	0.9215	0.9108	0.9088
0.0001	12.0	3600	0.6325	0.9108	0.9215	0.9108	0.9088
0.0001	12.3333	3700	0.6339	0.9108	0.9215	0.9108	0.9088
0.0001	12.6667	3800	0.6351	0.9108	0.9215	0.9108	0.9088
0.0001	13.0	3900	0.6371	0.9108	0.9215	0.9108	0.9088
0.0001	13.3333	4000	0.6376	0.9108	0.9215	0.9108	0.9088
0.0001	13.6667	4100	0.6393	0.9108	0.9215	0.9108	0.9088
0.0001	14.0	4200	0.6403	0.9108	0.9215	0.9108	0.9088
0.0001	14.3333	4300	0.6410	0.9108	0.9215	0.9108	0.9088
0.0001	14.6667	4400	0.6413	0.9108	0.9215	0.9108	0.9088
0.0001	15.0	4500	0.6414	0.9108	0.9215	0.9108	0.9088

Framework versions

• Transformers 4.48.2
• Pytorch 2.6.0+cu126
• Datasets 3.1.0
• Tokenizers 0.21.0