vit-base-patch16-224-ve-U11-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: 0.5456
• Accuracy: 0.8913

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	0.3696
1.3848	2.0	13	1.3692	0.5217
1.3848	2.92	19	1.3197	0.5435
1.3517	4.0	26	1.2264	0.5
1.2334	4.92	32	1.1280	0.4348
1.2334	6.0	39	1.0437	0.5435
1.073	6.92	45	0.9771	0.5870
0.9358	8.0	52	0.9470	0.6739
0.9358	8.92	58	0.8528	0.7826
0.7955	10.0	65	0.7839	0.7609
0.6429	10.92	71	0.7620	0.7391
0.6429	12.0	78	0.6466	0.8043
0.5096	12.92	84	0.7396	0.7174
0.4086	14.0	91	0.7335	0.7174
0.4086	14.92	97	0.6473	0.7391
0.3355	16.0	104	0.6019	0.7391
0.2511	16.92	110	0.5275	0.8261
0.2511	18.0	117	0.6069	0.7826
0.1925	18.92	123	0.6447	0.7826
0.2121	20.0	130	0.5044	0.8261
0.2121	20.92	136	0.4805	0.8478
0.1883	22.0	143	0.6723	0.8043
0.1883	22.92	149	0.7730	0.7391
0.1693	24.0	156	0.6574	0.7609
0.1252	24.92	162	0.8192	0.7391
0.1252	26.0	169	0.5984	0.7826
0.1439	26.92	175	0.7633	0.7826
0.137	28.0	182	0.6566	0.8478
0.137	28.92	188	0.6550	0.8261
0.1316	30.0	195	0.7163	0.7391
0.1101	30.92	201	0.6241	0.7826
0.1101	32.0	208	0.6360	0.8478
0.0947	32.92	214	0.5273	0.8696
0.0885	34.0	221	0.6579	0.8261
0.0885	34.92	227	0.5920	0.8696
0.0967	36.0	234	0.6779	0.8261
0.0812	36.92	240	0.7354	0.8043
0.0812	38.0	247	0.6825	0.8261
0.0752	38.92	253	0.6348	0.8478
0.0757	40.0	260	0.7726	0.8043
0.0757	40.92	266	0.6737	0.8261
0.086	42.0	273	0.6738	0.7826
0.086	42.92	279	0.7295	0.7609
0.0533	44.0	286	0.6897	0.8261
0.0574	44.92	292	0.6427	0.8261
0.0574	46.0	299	0.6471	0.8261
0.0739	46.92	305	0.6645	0.8261
0.0849	48.0	312	0.6858	0.8043
0.0849	48.92	318	0.7475	0.8043
0.0719	50.0	325	0.6735	0.8261
0.0434	50.92	331	0.6892	0.8478
0.0434	52.0	338	0.6820	0.8478
0.0564	52.92	344	0.6677	0.8478
0.0408	54.0	351	0.7379	0.8043
0.0408	54.92	357	0.5456	0.8913
0.0464	56.0	364	0.7951	0.7826
0.0463	56.92	370	0.6356	0.8478
0.0463	58.0	377	0.7529	0.8261
0.0361	58.92	383	0.8017	0.8261
0.0457	60.0	390	0.7877	0.8478
0.0457	60.92	396	0.8019	0.7826
0.0371	62.0	403	0.8015	0.8043
0.0371	62.92	409	0.8487	0.8043
0.0452	64.0	416	0.9401	0.7609
0.0455	64.92	422	0.9647	0.7609
0.0455	66.0	429	0.8958	0.7609
0.0408	66.92	435	0.8531	0.7826
0.0418	68.0	442	0.8206	0.8043
0.0418	68.92	448	0.8045	0.8043
0.0424	70.0	455	0.8090	0.8043
0.038	70.92	461	0.7902	0.8043
0.038	72.0	468	0.8008	0.8261
0.0401	72.92	474	0.8122	0.8043
0.0347	73.85	480	0.8161	0.8043

Framework versions

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