Swin-Tiny Image Classification Model - Open source and free, fine-tuned on specific datasets for precise image classification

Swin Tiny Patch4 Window7 224 Finetuned Woody LeftGR 130epochs

Developed by Alex-VisTas

Image classification model based on Swin Transformer Tiny architecture, fine-tuned for 130 epochs on a specific image dataset

Image Classification

Transformers

Open Source License:Apache-2.0 #High-precision Image Classification #SwinTransformer Fine-tuning #Small Window Attention

Downloads 12

Release Time : 10/27/2022

Model Overview

This is a vision Transformer model based on Microsoft's Swin Transformer Tiny architecture, fine-tuned for 130 epochs, specifically designed for image classification tasks. Achieves 90.47% accuracy on the evaluation set.

Model Features

High Accuracy

Achieves 90.47% classification accuracy on the evaluation set

Swin Transformer Architecture

Efficient vision Transformer architecture using hierarchical window attention mechanism

Extended Training

Fine-tuned through 130 epochs of meticulous training

Model Capabilities

Image Classification

Visual Feature Extraction

Use Cases

Image Recognition

Object Classification

Classifies and identifies main objects in images

90.47% evaluation accuracy

🚀 swin-tiny-patch4-window7-224-finetuned-woody_LeftGR_130epochs

This model is a fine - tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 0.3377
Accuracy: 0.9047

🚀 Quick Start

This model is a fine - tuned image classification model. It can be used for various image - related tasks after proper adaptation.

📚 Documentation

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: 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: 0.1
num_epochs: 130

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6614	1.0	61	0.6404	0.6521
0.5982	2.0	122	0.5548	0.7107
0.579	3.0	183	0.5390	0.7141
0.5621	4.0	244	0.4920	0.7623
0.5567	5.0	305	0.5375	0.7313
0.5271	6.0	366	0.5542	0.7405
0.5312	7.0	427	0.4573	0.7876
0.5477	8.0	488	0.4540	0.7784
0.5554	9.0	549	0.4932	0.7635
0.5247	10.0	610	0.4407	0.7968
0.5239	11.0	671	0.4479	0.7842
0.5294	12.0	732	0.4509	0.7910
0.531	13.0	793	0.4419	0.7933
0.5493	14.0	854	0.4646	0.7784
0.4934	15.0	915	0.4310	0.7968
0.4965	16.0	976	0.4449	0.7876
0.4946	17.0	1037	0.4342	0.8129
0.4716	18.0	1098	0.4129	0.8140
0.4679	19.0	1159	0.4290	0.8002
0.4799	20.0	1220	0.4356	0.7842
0.4744	21.0	1281	0.4042	0.8094
0.4512	22.0	1342	0.3953	0.8117
0.4633	23.0	1403	0.4157	0.7956
0.4528	24.0	1464	0.3920	0.8094
0.4427	25.0	1525	0.3930	0.8220
0.4238	26.0	1586	0.3891	0.8140
0.4257	27.0	1647	0.3700	0.8255
0.4102	28.0	1708	0.4122	0.7968
0.4505	29.0	1769	0.4210	0.7945
0.3973	30.0	1830	0.3923	0.8197
0.3824	31.0	1891	0.3908	0.8473
0.3887	32.0	1952	0.3897	0.8312
0.3723	33.0	2013	0.3747	0.8381
0.3608	34.0	2074	0.3706	0.8301
0.3718	35.0	2135	0.3937	0.8255
0.3692	36.0	2196	0.3984	0.8037
0.3533	37.0	2257	0.3792	0.8335
0.3625	38.0	2318	0.4070	0.8163
0.3633	39.0	2379	0.4130	0.8232
0.3602	40.0	2440	0.3996	0.8186
0.3557	41.0	2501	0.3756	0.8335
0.3373	42.0	2562	0.3914	0.8220
0.3102	43.0	2623	0.4165	0.8507
0.3135	44.0	2684	0.3852	0.8278
0.3286	45.0	2745	0.4164	0.8450
0.316	46.0	2806	0.3498	0.8496
0.2802	47.0	2867	0.3887	0.8462
0.3184	48.0	2928	0.3829	0.8576
0.2785	49.0	2989	0.3627	0.8485
0.2988	50.0	3050	0.3679	0.8370
0.267	51.0	3111	0.3528	0.8645
0.2907	52.0	3172	0.3538	0.8519
0.2857	53.0	3233	0.3593	0.8530
0.2651	54.0	3294	0.3732	0.8439
0.2447	55.0	3355	0.3441	0.8542
0.2542	56.0	3416	0.3897	0.8576
0.2634	57.0	3477	0.4082	0.8657
0.2505	58.0	3538	0.3416	0.8657
0.2555	59.0	3599	0.3725	0.8576
0.2466	60.0	3660	0.3496	0.8680
0.2585	61.0	3721	0.3214	0.8783
0.235	62.0	3782	0.3584	0.8737
0.215	63.0	3843	0.3467	0.8657
0.236	64.0	3904	0.3471	0.8829
0.2211	65.0	3965	0.3318	0.8863
0.1989	66.0	4026	0.3645	0.8852
0.2133	67.0	4087	0.3456	0.8898
0.2169	68.0	4148	0.3287	0.8852
0.223	69.0	4209	0.3182	0.8921
0.2379	70.0	4270	0.3260	0.8840
0.2149	71.0	4331	0.3230	0.8886
0.2007	72.0	4392	0.3926	0.8760
0.2091	73.0	4453	0.4133	0.8783
0.2229	74.0	4514	0.3867	0.8772
0.1903	75.0	4575	0.3594	0.8840
0.2124	76.0	4636	0.3388	0.8875
0.1999	77.0	4697	0.3305	0.8875
0.2053	78.0	4758	0.4670	0.8840
0.1958	79.0	4819	0.3468	0.8909
0.1839	80.0	4880	0.3902	0.8886
0.1715	81.0	4941	0.3830	0.8875
0.1803	82.0	5002	0.3134	0.8967
0.1803	83.0	5063	0.3935	0.8909
0.1865	84.0	5124	0.3882	0.8863
0.1884	85.0	5185	0.3485	0.8990
0.1663	86.0	5246	0.3667	0.8944
0.1665	87.0	5307	0.3545	0.8932
0.1556	88.0	5368	0.3882	0.8944
0.18	89.0	5429	0.3751	0.8898
0.1974	90.0	5490	0.3979	0.8863
0.1622	91.0	5551	0.3623	0.8967
0.1657	92.0	5612	0.3855	0.8978
0.1672	93.0	5673	0.3722	0.8944
0.1807	94.0	5734	0.3994	0.8932
0.1419	95.0	5795	0.4017	0.8863
0.178	96.0	5856	0.4168	0.8886
0.1402	97.0	5917	0.3727	0.8944
0.1427	98.0	5978	0.3919	0.8967
0.1318	99.0	6039	0.3843	0.8955
0.1417	100.0	6100	0.4017	0.8898
0.1536	101.0	6161	0.3613	0.8955
0.1631	102.0	6222	0.3377	0.9047
0.1459	103.0	6283	0.3724	0.8967
0.1499	104.0	6344	0.3934	0.8955
0.1572	105.0	6405	0.3368	0.8967
0.1308	106.0	6466	0.3782	0.8990
0.1535	107.0	6527	0.3306	0.9024
0.125	108.0	6588	0.4076	0.8898
0.1339	109.0	6649	0.3628	0.8990
0.148	110.0	6710	0.3672	0.9013
0.1725	111.0	6771	0.4006	0.8909
0.1326	112.0	6832	0.4117	0.8921
0.1438	113.0	6893	0.3927	0.8978
0.1205	114.0	6954	0.3612	0.8990
0.1531	115.0	7015	0.3594	0.8932
0.1473	116.0	7076	0.4490	0.8875
0.1388	117.0	7137	0.3952	0.8921
0.136	118.0	7198	0.4098	0.8921
0.1579	119.0	7259	0.3595	0.9013
0.1359	120.0	7320	0.3970	0.8944
0.1314	121.0	7381	0.4092	0.8932
0.1337	122.0	7442	0.4192	0.8909
0.1538	123.0	7503	0.4154	0.8898
0.119	124.0	7564	0.4120	0.8909
0.1353	125.0	7625	0.4060	0.8921
0.1489	126.0	7686	0.4162	0.8909
0.1554	127.0	7747	0.4148	0.8944
0.1558	128.0	7808	0.4169	0.8944
0.1268	129.0	7869	0.4110	0.8955
0.1236	130.0	7930	0.4197	0.8944

Framework versions

Transformers 4.23.1
Pytorch 1.12.1+cu113
Datasets 2.6.1
Tokenizers 0.13.1

📄 License

This model is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご