The latest version workflow and nodes are here <a target="_blank" rel="ugc" href="https://github.com/peterkickasspeter-civit/ComfyUI-ZImageTurboProgressiveLockedUpscale/tree/experiment">https://github.com/peterkickasspeter-civit/ComfyUI-ZImageTurboProgressiveLockedUpscale/tree/experiment</a> <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">===== Custom node: <a target="_blank" rel="ugc" href="https://github.com/peterkickasspeter-civit/ComfyUI-ZImageTurboProgressiveLockedUpscale">https://github.com/peterkickasspeter-civit/ComfyUI-ZImageTurboProgressiveLockedUpscale</a> Workflow is in the showcase images. More details here - <a target="_blank" rel="ugc" href="https://www.reddit.com/r/StableDiffusion/comments/1r1qyj2/zimageturboprogressivelockedupscale_works_with_z/">https://www.reddit.com/r/StableDiffusion/comments/1r1qyj2/zimageturboprogressivelockedupscale_works_with_z/</a> Distill lora for base - <a target="_blank" rel="ugc" href="https://huggingface.co/alibaba-pai/Z-Image-Fun-Lora-Distill/tree/main">https://huggingface.co/alibaba-pai/Z-Image-Fun-Lora-Distill/tree/main</a> *use either Z-Image-Fun-Lora-Distill-8-Steps.safetensors or Z-Image-Fun-Lora-Distill-4-Steps-2602-ComfyUI.safetensors

Amateur Photography

GPT4O Prompt - <a target="_blank" rel="ugc" href="https://pastebin.com/89SnakX7">https://pastebin.com/89SnakX7</a>

amateurphoto-v5-14-15-1-1.safetensors

Amateur Photography [Flux Dev]

ama photo

v5-final

Zbase Nano banana at home

Zbase test

Z Image Base and Turbo

Z-Image Turbo v1.0

Qwen v1.0

v5-beta

v4-soap-testing

v3.0

v2.0

v1.0

No trigger word needed. But you have to use some photography terms to steer Flux (like photo, photograph etc.). You can start with the below. Its how I tagged the training dataset

Full body shot photo of

Medium shot photo of

Selfie photo of

Close-up photo of

Or what ever you want (adjust the steps, weight based on your prompt). Detailed prompts still works best

******* Photography

Instagram Amateur Photo

Amateurs Photography [Flux Dev]

Flux ******* photography 

Mateurs Photography 

Hello Everyone, Please read this before you use the LoraRecommended Settings (v5-final):<ul><li>Distilled CFG Scale: 2.5 to 4</li><li>Sampling method and Schedule type: Heun with BETA or DEIS with DDIM or [Forge] Flux Realistic (Slow) with Beta / DDIM</li><li>Steps: &gt;=20 (Sometimes I use 20 or 30 or 35 or 40 - You should check at how many steps the image converges - So, please experiment yourself)</li><li>Resolution: 896x1152 (I made sure it works at this resolution since some of you did not like it when i said generate at high resolutions) - But you have to use hiresfix. Below are the settings I used in my example images<ul><li>Hires. fix: 4x ultrasharp, 0.4 denoise, 10 steps (I'm not the best upscale guy. If you have other upscale methods, you can use it)</li><li>You can also directly generate at 1344x1728 if you can</li></ul></li><li>Checkpoint: flux1-dev-Q8_0.gguf (I use the Q8 one. I don't test this Lora with FP8 or Q4 or other quants)</li><li>Lora Weight: 0.3-0.5. You have to experiment based on your prompts</li><li>Positive Prompt: No trigger word needed. But you have to use some photography terms to steer Flux (like photo, photograph etc.). You can start with the below. Its how I tagged the training dataset<ul><li>Full body shot photo of</li><li>Medium shot photo of</li><li>Selfie photo of</li><li>Close-up photo of</li><li>Or what ever you want (adjust the steps, weight based on your prompt). Detailed prompts still works best</li></ul></li></ul><ul><li>Issues: If it happens, please reduce weight<ul><li>Hands</li><li>Text</li><li>People in the background</li><li>Skin texture - its not perfect. I don't like this also. I am trying to find ways to improve it</li></ul></li></ul>I would like to thank everyone who tipped Buzz and made this version possible. Honorable mentions:<ul><li>plectrudecatastrophe</li><li>Paper_Cranes</li><li>congo2008</li></ul>

Hello Everyone, Please read this before you use the Lora

Recommended Settings (v5-final):

Lora Weight: 0.3-0.5. You have to experiment based on your prompts

Positive Prompt: No trigger word needed. But you have to use some photography terms to steer Flux (like photo, photograph etc.). You can start with the below. Its how I tagged the training dataset
        Full body shot photo of
        Medium shot photo of
        Selfie photo of
        Close-up photo of
        Or what ever you want (adjust the steps, weight based on your prompt). Detailed prompts still works best

Issues: If it happens, please reduce weight
        Hands
        Text
        People in the background
        Skin texture - its not perfect. I don't like this also. I am trying to find ways to improve it

Distilled CFG Scale: 2.5 to 4

Sampling method and Schedule type: 
          Heun with BETA or DEIS with DDIM or 
         [Forge] Flux Realistic (Slow) with Beta / DDIM
 
Steps: >=20 (Sometimes I use 20 or 30 or 35 or 40 - You should check at how many steps the image converges - So, please experiment yourself)

Resolution: 896x1152 (I made sure it works at this resolution since some of you did not like it when i said generate at high resolutions) - But you have to use hiresfix. Below are the settings I used in my example images

Hires. fix: 4x ultrasharp, 0.4 denoise, 10 steps (I'm not the best upscale guy. If you have other upscale methods, you can use it)
        You can also directly generate at 1344x1728 if you can

Checkpoint: flux1-dev-Q8_0.gguf (I use the Q8 one. I don't test this Lora with FP8 or Q4 or other quants)


I would like to thank everyone who tipped Buzz and made this version possible. Honorable mentions:
    plectrudecatastrophe
    Paper_Cranes
    congo2008



v5_final

Files with SHA256: ebdb69eeb51361dac540c6cefc54a59639f91ba3019d9ee05fd54560f5d3e90a