README.md

metadata

library_name: transformers
license: apache-2.0
base_model: google/vit-base-patch16-224-in21k
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: vit-base-patch16-224-in21k-finetuned-inaturalist
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: validation
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9483333333333334

vit-base-patch16-224-in21k-finetuned-inaturalist

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: 1.0056
• Accuracy: 0.9483

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-05
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8	2	3.2007	0.0333
No log	2.0	5	3.1889	0.0444
No log	2.8	7	3.1747	0.0639
3.1888	4.0	10	3.1442	0.1097
3.1888	4.8	12	3.1183	0.1458
3.1888	6.0	15	3.0710	0.2194
3.1888	6.8	17	3.0331	0.3042
3.0673	8.0	20	2.9627	0.4389
3.0673	8.8	22	2.9109	0.4944
3.0673	10.0	25	2.8360	0.5764
3.0673	10.8	27	2.7809	0.6056
2.8151	12.0	30	2.6958	0.6542
2.8151	12.8	32	2.6401	0.6764
2.8151	14.0	35	2.5583	0.6944
2.8151	14.8	37	2.5034	0.7083
2.5143	16.0	40	2.4202	0.7347
2.5143	16.8	42	2.3662	0.7375
2.5143	18.0	45	2.2884	0.7444
2.5143	18.8	47	2.2374	0.7569
2.2236	20.0	50	2.1632	0.7778
2.2236	20.8	52	2.1175	0.7833
2.2236	22.0	55	2.0528	0.7931
2.2236	22.8	57	2.0099	0.7958
1.9677	24.0	60	1.9488	0.8014
1.9677	24.8	62	1.9113	0.8097
1.9677	26.0	65	1.8582	0.8139
1.9677	26.8	67	1.8242	0.8139
1.7467	28.0	70	1.7740	0.8111
1.7467	28.8	72	1.7458	0.8056
1.7467	30.0	75	1.7013	0.8181
1.7467	30.8	77	1.6714	0.8194
1.5765	32.0	80	1.6316	0.8264
1.5765	32.8	82	1.6083	0.8236
1.5765	34.0	85	1.5738	0.8292
1.5765	34.8	87	1.5531	0.8347
1.4431	36.0	90	1.5228	0.8431
1.4431	36.8	92	1.5046	0.8444
1.4431	38.0	95	1.4780	0.8472
1.4431	38.8	97	1.4608	0.8458
1.3049	40.0	100	1.4357	0.8458
1.3049	40.8	102	1.4188	0.85
1.3049	42.0	105	1.3949	0.8528
1.3049	42.8	107	1.3808	0.8528
1.2312	44.0	110	1.3636	0.8458
1.2312	44.8	112	1.3513	0.8486
1.2312	46.0	115	1.3329	0.8528
1.2312	46.8	117	1.3193	0.8528
1.1368	48.0	120	1.3025	0.8528
1.1368	48.8	122	1.2945	0.8542
1.1368	50.0	125	1.2820	0.8528
1.1368	50.8	127	1.2705	0.8569
1.0821	52.0	130	1.2616	0.8583
1.0821	52.8	132	1.2545	0.8556
1.0821	54.0	135	1.2423	0.8542
1.0821	54.8	137	1.2332	0.8597
1.0232	56.0	140	1.2210	0.8639
1.0232	56.8	142	1.2161	0.8625
1.0232	58.0	145	1.2094	0.8569
1.0232	58.8	147	1.2057	0.8542
0.9814	60.0	150	1.1973	0.85
0.9814	60.8	152	1.1919	0.8486
0.9814	62.0	155	1.1825	0.8625
0.9814	62.8	157	1.1799	0.8597
0.9415	64.0	160	1.1716	0.8597
0.9415	64.8	162	1.1665	0.8625
0.9415	66.0	165	1.1611	0.8639
0.9415	66.8	167	1.1600	0.8625
0.9135	68.0	170	1.1577	0.8639
0.9135	68.8	172	1.1547	0.8639
0.9135	70.0	175	1.1493	0.8639
0.9135	70.8	177	1.1464	0.8611
0.8946	72.0	180	1.1423	0.8556
0.8946	72.8	182	1.1402	0.8611
0.8946	74.0	185	1.1375	0.8583
0.8946	74.8	187	1.1360	0.8597
0.8866	76.0	190	1.1344	0.8625
0.8866	76.8	192	1.1334	0.8639
0.8866	78.0	195	1.1324	0.8639
0.8866	78.8	197	1.1320	0.8639
0.8798	80.0	200	1.1319	0.8639

Framework versions

• Transformers 4.45.2
• Pytorch 2.4.1+cu121
• Datasets 2.21.0
• Tokenizers 0.20.1