metadata

license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: vit-base-skin
    results: []

vit-base-skin

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.7902
Accuracy: 0.8446
F1: 0.8443
Precision: 0.8449
Recall: 0.8446

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: 0.0002
train_batch_size: 16
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 6
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.8192	0.16	100	0.7593	0.7098	0.6664	0.7150	0.7098
0.7637	0.32	200	0.6606	0.7513	0.7008	0.7697	0.7513
0.6177	0.48	300	0.5816	0.7927	0.7706	0.7834	0.7927
0.837	0.64	400	0.5884	0.7927	0.7831	0.7980	0.7927
0.6409	0.8	500	0.5473	0.8290	0.8270	0.8419	0.8290
0.5896	0.96	600	0.5079	0.8083	0.7904	0.8440	0.8083
0.4101	1.12	700	0.4893	0.8290	0.8219	0.8343	0.8290
0.6753	1.28	800	0.6570	0.7668	0.7800	0.8298	0.7668
0.3907	1.44	900	0.4257	0.8238	0.8152	0.8117	0.8238
0.447	1.6	1000	0.5717	0.8031	0.8030	0.8447	0.8031
0.4131	1.76	1100	0.4189	0.8342	0.8271	0.8304	0.8342
0.3913	1.92	1200	0.3728	0.8860	0.8877	0.8995	0.8860
0.2686	2.08	1300	0.5161	0.7979	0.7995	0.8199	0.7979
0.2466	2.24	1400	0.4671	0.8601	0.8570	0.8683	0.8601
0.2456	2.4	1500	0.4479	0.8446	0.8371	0.8372	0.8446
0.2218	2.56	1600	0.5276	0.8342	0.8360	0.8483	0.8342
0.2019	2.72	1700	0.4866	0.8290	0.8289	0.8338	0.8290
0.1856	2.88	1800	0.4727	0.8342	0.8371	0.8426	0.8342
0.1614	3.04	1900	0.5576	0.8135	0.8126	0.8164	0.8135
0.0985	3.19	2000	0.5765	0.8394	0.8350	0.8413	0.8394
0.0826	3.35	2100	0.6482	0.8238	0.8062	0.7906	0.8238
0.1288	3.51	2200	0.6919	0.8238	0.8247	0.8415	0.8238
0.0808	3.67	2300	0.7174	0.8135	0.8141	0.8175	0.8135
0.0764	3.83	2400	0.6201	0.8446	0.8445	0.8454	0.8446
0.1346	3.99	2500	0.6639	0.8187	0.8165	0.8177	0.8187
0.0246	4.15	2600	0.7757	0.8135	0.8153	0.8177	0.8135
0.0059	4.31	2700	0.7933	0.8238	0.8195	0.8170	0.8238
0.0187	4.47	2800	0.8311	0.8238	0.8192	0.8153	0.8238
0.023	4.63	2900	0.8048	0.8238	0.8216	0.8217	0.8238
0.0186	4.79	3000	0.8510	0.8342	0.8301	0.8295	0.8342
0.0562	4.95	3100	0.7534	0.8446	0.8434	0.8429	0.8446
0.0046	5.11	3200	0.7817	0.8497	0.8480	0.8509	0.8497
0.0032	5.27	3300	0.7576	0.8446	0.8429	0.8447	0.8446
0.0031	5.43	3400	0.7692	0.8394	0.8374	0.8373	0.8394
0.0399	5.59	3500	0.7673	0.8394	0.8395	0.8415	0.8394
0.0052	5.75	3600	0.7809	0.8394	0.8388	0.8399	0.8394
0.0203	5.91	3700	0.7902	0.8446	0.8443	0.8449	0.8446

Framework versions

Transformers 4.29.2
Pytorch 1.13.1
Datasets 2.14.5
Tokenizers 0.13.3