Shedinja/vit-pet-classification

resultados

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

• Loss: 0.2450
• Accuracy: 0.9350
• F1: 0.9345
• Precision: 0.9383
• Recall: 0.9350

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: 2e-05
• 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
No log	1.0	185	3.0723	0.3166	0.2902	0.3034	0.3166
No log	2.0	370	1.1438	0.8850	0.8847	0.8924	0.8850
2.2334	3.0	555	0.3604	0.9229	0.9230	0.9266	0.9229
2.2334	4.0	740	0.2413	0.9350	0.9349	0.9383	0.9350
2.2334	5.0	925	0.2297	0.9323	0.9315	0.9349	0.9323
0.1102	6.0	1110	0.2163	0.9391	0.9390	0.9417	0.9391
0.1102	7.0	1295	0.2179	0.9364	0.9359	0.9395	0.9364
0.1102	8.0	1480	0.2164	0.9378	0.9373	0.9404	0.9378
0.0078	9.0	1665	0.2169	0.9364	0.9359	0.9392	0.9364
0.0078	10.0	1850	0.2214	0.9350	0.9345	0.9382	0.9350
0.003	11.0	2035	0.2225	0.9364	0.9359	0.9395	0.9364
0.003	12.0	2220	0.2252	0.9364	0.9359	0.9394	0.9364
0.003	13.0	2405	0.2265	0.9364	0.9359	0.9394	0.9364
0.0018	14.0	2590	0.2276	0.9350	0.9345	0.9383	0.9350
0.0018	15.0	2775	0.2308	0.9350	0.9345	0.9383	0.9350
0.0018	16.0	2960	0.2320	0.9364	0.9359	0.9394	0.9364
0.0012	17.0	3145	0.2327	0.9350	0.9345	0.9383	0.9350
0.0012	18.0	3330	0.2351	0.9350	0.9345	0.9383	0.9350
0.0009	19.0	3515	0.2364	0.9350	0.9345	0.9383	0.9350
0.0009	20.0	3700	0.2374	0.9350	0.9345	0.9383	0.9350
0.0009	21.0	3885	0.2386	0.9350	0.9345	0.9383	0.9350
0.0007	22.0	4070	0.2396	0.9350	0.9345	0.9383	0.9350
0.0007	23.0	4255	0.2412	0.9350	0.9345	0.9383	0.9350
0.0007	24.0	4440	0.2422	0.9350	0.9345	0.9383	0.9350
0.0006	25.0	4625	0.2429	0.9350	0.9345	0.9383	0.9350
0.0006	26.0	4810	0.2433	0.9350	0.9345	0.9383	0.9350
0.0006	27.0	4995	0.2440	0.9350	0.9345	0.9383	0.9350
0.0005	28.0	5180	0.2446	0.9350	0.9345	0.9383	0.9350
0.0005	29.0	5365	0.2449	0.9350	0.9345	0.9383	0.9350
0.0005	30.0	5550	0.2450	0.9350	0.9345	0.9383	0.9350

Framework versions

• Transformers 4.52.3
• Pytorch 2.6.0+cu124
• Datasets 2.14.4
• Tokenizers 0.21.1