Stable Diffusion 3
Posted: Thu Jun 13, 2024 7:44 am
Stable Diffusion 3 (SD3) is the latest text-to-image model from Stability AI, launched on June 12th, 2024. It builds upon the success of its predecessor, SDXL, and offers enhanced capabilities for transforming text into images.
Acquiring SD3: To get your hands on SD3, head over to Civitai where you can download everything you need:
SD3 Medium (4.2 GB)
SD3 Medium Incl Clips (5.8 GB)
SD3 Medium Incl Clips T5XXLFP8 (10.6 GB)
Text Encoder Clip L (234 MB)
Text Encoder Clip G (1.3 GB)
Text Encoder t5xxl_fp8_e4m3fn (4.7 GB)
Text Encoder t5xxl_fp16 (9.5 GB)
Important Download Information: ComfyUI supports SD3, so ensure you have the latest version. As of the morning of June 12th, 2024, Automatic1111 WebUI does not yet support SD3, but it’s expected to within a week.
When choosing your SD3 checkpoint, consider:
sd3_medium.safetensors includes MMDiT and VAE weights but not text encoders.
sd3_medium_incl_clips_t5xxlfp8.safetensors contains all necessary weights, including the fp8 version of the T5XXL text encoder.
sd3_medium_incl_clips.safetensors includes all necessary weights except for the T5XXL text encoder.
Creating Images with SD3: Once you have the necessary files, you can start creating images using pre-made ComfyUI workflows.
License and Training: Be sure to review the license considerations before using SD3. If you’re interested in training the model further, resources are available for that purpose.
Additional Information: For more details, the community and Stability AI are available to assist with any questions.
This overview provides a concise and clear understanding of what Stable Diffusion 3 is and how to begin using it for your creative projects.
Acquiring SD3: To get your hands on SD3, head over to Civitai where you can download everything you need:
SD3 Medium (4.2 GB)
SD3 Medium Incl Clips (5.8 GB)
SD3 Medium Incl Clips T5XXLFP8 (10.6 GB)
Text Encoder Clip L (234 MB)
Text Encoder Clip G (1.3 GB)
Text Encoder t5xxl_fp8_e4m3fn (4.7 GB)
Text Encoder t5xxl_fp16 (9.5 GB)
Important Download Information: ComfyUI supports SD3, so ensure you have the latest version. As of the morning of June 12th, 2024, Automatic1111 WebUI does not yet support SD3, but it’s expected to within a week.
When choosing your SD3 checkpoint, consider:
sd3_medium.safetensors includes MMDiT and VAE weights but not text encoders.
sd3_medium_incl_clips_t5xxlfp8.safetensors contains all necessary weights, including the fp8 version of the T5XXL text encoder.
sd3_medium_incl_clips.safetensors includes all necessary weights except for the T5XXL text encoder.
Creating Images with SD3: Once you have the necessary files, you can start creating images using pre-made ComfyUI workflows.
License and Training: Be sure to review the license considerations before using SD3. If you’re interested in training the model further, resources are available for that purpose.
Additional Information: For more details, the community and Stability AI are available to assist with any questions.
This overview provides a concise and clear understanding of what Stable Diffusion 3 is and how to begin using it for your creative projects.