vit-base-patch16-224-ve-U13-b-80

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

• Loss: 0.5742
• Accuracy: 0.8696

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: 5.5e-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: 0.05
• num_epochs: 80

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.92	6	1.3848	0.3478
1.3848	2.0	13	1.3692	0.5217
1.3848	2.92	19	1.3184	0.5870
1.352	4.0	26	1.2217	0.4565
1.2316	4.92	32	1.1418	0.4783
1.2316	6.0	39	1.0689	0.4783
1.0849	6.92	45	0.9931	0.5870
0.9314	8.0	52	0.9458	0.6957
0.9314	8.92	58	0.8675	0.6957
0.8001	10.0	65	0.8148	0.7174
0.6493	10.92	71	0.7692	0.7609
0.6493	12.0	78	0.6428	0.8043
0.5145	12.92	84	0.6025	0.8261
0.379	14.0	91	0.5621	0.8043
0.379	14.92	97	0.5298	0.8478
0.2942	16.0	104	0.5791	0.8043
0.2096	16.92	110	0.5814	0.7826
0.2096	18.0	117	0.7829	0.7174
0.2113	18.92	123	0.5658	0.8478
0.2143	20.0	130	0.7036	0.7609
0.2143	20.92	136	0.5924	0.7826
0.1752	22.0	143	0.6852	0.7609
0.1752	22.92	149	0.7237	0.7609
0.1238	24.0	156	0.6743	0.8043
0.1401	24.92	162	0.8463	0.6957
0.1401	26.0	169	0.7872	0.7609
0.1544	26.92	175	0.5492	0.8261
0.1163	28.0	182	0.5756	0.8043
0.1163	28.92	188	0.7621	0.7609
0.1121	30.0	195	0.6972	0.7826
0.1065	30.92	201	0.5723	0.8261
0.1065	32.0	208	0.7503	0.8261
0.1021	32.92	214	0.6127	0.8043
0.1048	34.0	221	0.5734	0.8478
0.1048	34.92	227	0.5817	0.8478
0.0848	36.0	234	0.5903	0.8261
0.0769	36.92	240	0.7074	0.8261
0.0769	38.0	247	0.5835	0.8478
0.0825	38.92	253	0.6373	0.8043
0.0676	40.0	260	0.6793	0.8261
0.0676	40.92	266	0.6556	0.8261
0.0703	42.0	273	0.6329	0.8478
0.0703	42.92	279	0.6868	0.8261
0.0574	44.0	286	0.5997	0.8043
0.0523	44.92	292	0.5846	0.8261
0.0523	46.0	299	0.7214	0.8478
0.064	46.92	305	0.5230	0.8478
0.082	48.0	312	0.5850	0.8478
0.082	48.92	318	0.6346	0.8478
0.0694	50.0	325	0.6389	0.8261
0.0462	50.92	331	0.5813	0.8478
0.0462	52.0	338	0.5792	0.8478
0.044	52.92	344	0.5724	0.8261
0.0538	54.0	351	0.6294	0.8261
0.0538	54.92	357	0.5742	0.8696
0.0455	56.0	364	0.6951	0.8043
0.0537	56.92	370	0.6458	0.8043
0.0537	58.0	377	0.6259	0.8478
0.038	58.92	383	0.6748	0.8478
0.039	60.0	390	0.7236	0.8261
0.039	60.92	396	0.7758	0.8261
0.0304	62.0	403	0.7253	0.7609
0.0304	62.92	409	0.7513	0.8261
0.051	64.0	416	0.7547	0.8261
0.0355	64.92	422	0.8115	0.7826
0.0355	66.0	429	0.7768	0.8043
0.0435	66.92	435	0.7829	0.8043
0.0313	68.0	442	0.7787	0.8043
0.0313	68.92	448	0.7721	0.8261
0.0378	70.0	455	0.7672	0.8261
0.0339	70.92	461	0.7634	0.8261
0.0339	72.0	468	0.7615	0.8261
0.0311	72.92	474	0.7605	0.8261
0.0302	73.85	480	0.7603	0.8261

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0