Wholesome Training Datasets Wholesome, unrestrictively-licensed model training datasets. madebyollin/megalith-10m Viewer • Updated Jan 7 • 9.58M • 628 • 102 madebyollin/soa-aesthetic Viewer • Updated Jun 18, 2023 • 91.8k • 86 • 11 madebyollin/soa-full Viewer • Updated Jul 31, 2024 • 2.67M • 514 • 8
Tiny AutoEncoders Tiny distilled autoencoders offering faster encode / decode at slightly-reduced quality. madebyollin/taesd Updated Nov 7, 2024 • 25.8k • 67 madebyollin/taesdxl Updated Nov 7, 2024 • 4.16k • 64 madebyollin/taesd3 Updated Jun 14, 2024 • 1.12k • 38 madebyollin/taef1 Updated Dec 23, 2025 • 2.38k • 52
Wholesome Training Datasets Wholesome, unrestrictively-licensed model training datasets. madebyollin/megalith-10m Viewer • Updated Jan 7 • 9.58M • 628 • 102 madebyollin/soa-aesthetic Viewer • Updated Jun 18, 2023 • 91.8k • 86 • 11 madebyollin/soa-full Viewer • Updated Jul 31, 2024 • 2.67M • 514 • 8
Tiny AutoEncoders Tiny distilled autoencoders offering faster encode / decode at slightly-reduced quality. madebyollin/taesd Updated Nov 7, 2024 • 25.8k • 67 madebyollin/taesdxl Updated Nov 7, 2024 • 4.16k • 64 madebyollin/taesd3 Updated Jun 14, 2024 • 1.12k • 38 madebyollin/taef1 Updated Dec 23, 2025 • 2.38k • 52
