siglip-fashion-5ep

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

• Loss: 0.0405

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss
0.0388	0.4643	500	0.0368
0.0259	0.9285	1000	0.0268
0.0221	1.3928	1500	0.0243
0.0209	1.8570	2000	0.0216
0.0162	2.3213	2500	0.0217
0.0156	2.7855	3000	0.0209
0.0139	3.2498	3500	0.0206
0.0128	3.7140	4000	0.0206
0.0129	4.1783	4500	0.0222
0.0119	4.6425	5000	0.0205
0.0096	5.1068	5500	0.0211
0.0099	5.5710	6000	0.0223
0.0101	6.0353	6500	0.0244
0.0083	6.4995	7000	0.0241
0.0092	6.9638	7500	0.0228
0.009	7.4280	8000	0.0256
0.0087	7.8923	8500	0.0266
0.0076	8.3565	9000	0.0283
0.0071	8.8208	9500	0.0259
0.0072	9.2851	10000	0.0289
0.0068	9.7493	10500	0.0297
0.0064	10.2136	11000	0.0340
0.0066	10.6778	11500	0.0327
0.0061	11.1421	12000	0.0367
0.0055	11.6063	12500	0.0338
0.0064	12.0706	13000	0.0358
0.005	12.5348	13500	0.0369
0.0053	12.9991	14000	0.0367
0.0049	13.4633	14500	0.0381
0.0042	13.9276	15000	0.0400
0.0043	14.3918	15500	0.0400
0.0039	14.8561	16000	0.0405

Framework versions

• Transformers 4.57.1
• Pytorch 2.6.0+cu124
• Datasets 4.4.1
• Tokenizers 0.22.1