SD3.5 Large PiD Direct 4K Image Generation Workflow

Watch the full video first if you want to understand how this SD3.5 Large + PiD workflow works in practice. The video shows how a 1024 base image can be generated with SD3.5 Large, how the PiD enhancement stage pushes it toward 4K output, and how to launch the workflow online without building a local ComfyUI environment.

This ComfyUI workflow is designed for SD3.5 Large direct 4K image generation using PiD as the high-resolution enhancement stage. Its main purpose is to create a clean SD3.5 Large base image first, capture the correct latent state during sampling, and then send that latent into a PiD 2K-to-4K refinement pipeline. The result is a more detailed and more production-oriented output than a simple native image generation pass.

The workflow is built around sd3.5_large_fp8_scaled.safetensors as the main checkpoint. It uses a 1024×1024 EmptySD3LatentImage as the base canvas, CLIPTextEncode for positive and negative conditioning, PiDTextPrompt for unified prompt and caption handling, PiDKSamplerCapture for SD3 latent generation and PiD latent capture, PiDPrepare for preparing the captured latent, PiDSample for the PiD refinement pass, PiDFinalize for final image conversion, VAEDecode for native baseline output, and SaveImage for both comparison outputs.

The first lane is the native SD3.5 Large generation route. The prompt is written into PiDTextPrompt, then sent into the positive prompt encoder and also used as the caption for PiD preparation. The negative prompt suppresses common image problems such as deformation, blur, low resolution, bad anatomy, extra fingers, JPEG artifacts, and watermarks. This makes the workflow suitable for both artistic generation and cleaner comparison testing.

The key generation node is PiDKSamplerCapture. Unlike a normal KSampler, this node produces the final native latent while also capturing a PiD-ready latent at a selected capture step. In this workflow, the sampling setup uses 25 steps, CFG 4.4, Euler sampler, sgm_uniform scheduler, denoise 1.0, and capture_step 23. The captured PiD sigma is automatically sent into PiDPrepare, which helps the PiD stage receive the correct sampling-state information instead of relying on manual sigma guessing.

After capture, PiDPrepare is configured with the SD3 backbone, 2kto4k PiD checkpoint type, scale 4, auto download enabled, and cleanup after prepare enabled. PiDSample then performs the enhancement pass with 4 PiD steps, CFG scale 1.0, fixed seed, aggressive cleanup, and sequential block offload. PiDFinalize converts the sampled PiD output into the final enhanced image.

Compared with ordinary SD3.5 Large generation, this workflow is not just an external upscale. It captures an intermediate latent from the original generation process and uses PiD to rebuild high-resolution detail in a more integrated way. Compared with a standard upscaler, it is more closely connected to the model’s own latent structure and prompt caption, making it better for typography posters, detailed concept art, premium illustrations, product visuals, and high-resolution creative publishing.

Main features:

• SD3.5 Large PiD direct 4K workflow

• SD3.5 Large fp8 scaled checkpoint route

• 1024×1024 base latent generation

• Positive and negative prompt conditioning

• PiDTextPrompt unified prompt and caption input

• PiDKSamplerCapture intermediate latent capture

• 25-step Euler / sgm_uniform sampling setup

• Capture step 23 for PiD preparation

• Automatic PiD sigma transfer

• PiDPrepare with SD3 backbone

• PiD 2K-to-4K enhancement route

• PiDSample with 4 PiD steps

• Sequential block offload and aggressive cleanup

• Native baseline and enhanced output comparison

Suggested workflow:

Start with a clear visual prompt that defines the subject, typography if needed, composition, lighting, color palette, texture, and final use case. Keep the base resolution at 1024×1024 for the first test. Run the native SD3.5 Large lane first and check whether the layout, text, and major structure are correct. If the native image is weak, fix the prompt before entering PiD. Once the base image is stable, let PiDKSamplerCapture send the captured latent and sigma into PiDPrepare. Then run PiDSample and compare the native baseline with the PiD enhanced output. If the final output becomes too heavy or unstable, keep PiD steps at 4 and rely on cleanup and offload settings rather than forcing more aggressive enhancement.

⚙️ RunningHub Workflow

Try the workflow online right now — no installation required.
👉 Workflow: https://www.runninghub.ai/post/2059961785033256961?inviteCode=rh-v1111

If the results meet your expectations, you can later deploy it locally for customization.

🎁 Fan Benefits: Register to get 1000 points + daily login 100 points — enjoy 4090 performance and 48 GB super power!

📺 Bilibili Updates (Mainland China & Asia-Pacific)

If you’re in the Asia-Pacific region, you can watch the video below to see the workflow demonstration and creative breakdown.
📺 Bilibili Video: https://www.bilibili.com/video/BV1tsVH6kEBJ/

☕ Support Me on Ko-fi

If you find my content helpful and want to support future creations, you can buy me a coffee ☕.
Every bit of support helps me keep creating — just like a spark that can ignite a blazing flame.
👉 Ko-fi: https://ko-fi.com/aiksk

💼 Business Contact

For collaboration or inquiries, please contact aiksk95 on WeChat.

⚙️打开下方链接即可在线体验，无需安装。
👉 工作流： https://www.runninghub.ai/post/2059961785033256961?inviteCode=rh-v1111
如果觉得效果理想，你也可以在本地进行自定义部署。

🎁 粉丝福利： 注册即送 1000 积分，每日登录 100 积分，畅玩 4090 体验 48 G 超级性能！

📺 Bilibili 更新（中国大陆及南亚太地区）

如果你在中国大陆或南亚太地区，可以通过下方视频查看该工作流的实测效果与构思讲解。
📺 B站视频： https://www.bilibili.com/video/BV1tsVH6kEBJ/

我会在 夸克网盘 持续更新模型资源：
👉 https://pan.quark.cn/s/20c6f6f8d87b
这些资源主要面向本地用户，方便进行创作与学习。