<h1 id="flux-compact-ysf9obwba">FLUX COMPACT</h1>I am excited to present a collection of compact Flux models, with Clips and VAE included in a single model, optimized to integrate seamlessly into a variety of existing workflows. These compact versions come with the various combinations of Clips and Models, ensuring that they can be easily loaded directly via nodes such as 'CheckpointLoaderSimple' and others, without the need for additional configuration.As Sampler Node you can use the ones already present for SDXL using CFG 1, or you can use 'SamplersCustomAdvanced' that is dedicated to FLUX.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/23e4e1de-6e42-4daf-bc82-15b5aa67ab86/width=525/23e4e1de-6e42-4daf-bc82-15b5aa67ab86.jpeg" />If you use the dedicated FLUX node above , you can use one of the 'Guiders' as nodes to manage the guidance instead of the 'CFG of the sampler'<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/326164a8-ad28-4e91-b115-699f2425aa42/width=525/326164a8-ad28-4e91-b115-699f2425aa42.jpeg" />if not, you can use the 'FluxGuidance' node placed between the prompt and the sampler to manage the guidance on the other samplers already present for SDXL, keeping their CFG at 1 and modifying that of the new node.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/8366d733-887e-40a3-9452-602a26dcead7/width=525/8366d733-887e-40a3-9452-602a26dcead7.jpeg" />By reducing the final size by about 30%, these models are also accessible on less powerful machines, although I am not sure how this is possible. In my tests, I have not observed any differences in the generated outputs compared to the full versions.Let me know more with your experiments.I claim no credit for the original creation and publication of these models; all rights are reserved to 'Black Forest Labs' and their talented team. My goal is to make their excellent work more accessible to the community.I hope these compact Flux models will be useful to you and contribute to your creative and technical endeavors. Your feedback is greatly appreciated as we continue to explore and refine these tools together.<h2 id="flux.1-dev-compact-ci2ex8lle">Flux.1-Dev Compact</h2>MAIN SAMPLER: Euler + SimpleNORMAL CFG: 1FLUX GUIDANCE: 2+BETTER STEPS: 20+<h2 id="flux.1-schnell-compact-863sf2vq6">Flux.1-Schnell Compact</h2>MAIN SAMPLER: Euler + SimpleNORMAL CFG: 1FLUX GUIDANCE: 2+BETTER STEPS: 4+<h1 id="versions-contents:-u759uqf2k">VERSIONS CONTENTS:</h1><h3 id="flux.1-dev-fp8-3ex1nwi8b">Flux.1-Dev fp8</h3><ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp8_e4m3fn</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul><h3 id="flux.1-dev-fp16-738g2x0nx">Flux.1-Dev fp16</h3><ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp16</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul><h3 id="flux.1-schnell-fp8-s83on4xca">Flux.1-Schnell fp8</h3><ul><li>Unet: Flux.1-Schnell</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp8_e4m3fn</li><li>Clip2: clip_l</li><li>VAE: ae_Schnell</li></ul><h3 id="flux.1-schnell-fp16-pwl6fy8qs">Flux.1-Schnell fp16</h3><ul><li>Unet: Flux.1-Schnell</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp16</li><li>Clip2: clip_l</li><li>VAE: ae_Schnell</li></ul><h3 id="flux.1-dev-fp16-alt-7pspbhojf">Flux.1-Dev fp16 ALT</h3><ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e5m2</li><li>Clip1: t5xxl_fp16</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul><h3 id="flux.1-dev-fp8-alt-178zxacyf">Flux.1-Dev fp8 ALT</h3><ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e5m2</li><li>Clip1: t5xxl_fp8_e4m3fn</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul>

Flux.1 Compact | CLIP and VAE included

flux1CompactCLIPAnd_Flux1DevFp16.safetensors

🟡 Flux.1-Dev fp16

🔴 Flux.1-Dev fp8 ALT

🔴 Flux.1-Dev fp16 ALT

🔵 Flux.1-Schnell fp16

🟦 Flux.1-Schnell fp8

🟨 Flux.1-Dev fp8

A checkpoint I created using Flux1-dev as my base. I needed a smaller model to create some LoRas for Civit's Flux creation competition, so I used this after creating it in ComfyUI.If I can find links to the resources used, I'll edit this post and be sure to include them (give credit where it's due!).For the Flux creation entry, I believe there was something about including any LoRas used and other training information, so I figure what better way to share that information that a workflow!The workflow included below is the workflow I used to generate new images for the datasets used in CivitAI's Flux contest - I'll zip up the full workflow json files and restrict who can download them, if possible. Otherwise, just ask me for them and I'll send 'em your way!<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/edcf5fed-74f0-4570-a877-3b479ee447c6/width=525/edcf5fed-74f0-4570-a877-3b479ee447c6.jpeg" />

Flux_Fantasy_Arbiter

I needed a model that loaded faster than Flux1-Dev, so I created this checkpoint after having run the Flux model through my LoRas and some creative masking and noise generation shenanigans.Fantasy art based, does included nudity, but it's still a Flux based checkpoint, so it takes a bit of coercion.

fluxFantasyArbiter_v10.safetensors

fluxFantasyArbiter_v10_trainingData.zip

v1.0

FLUX COMPACTI am excited to present a collection of compact Flux models, with Clips and VAE included in a single model, optimized to integrate seamlessly into a variety of existing workflows. These compact versions come with the various combinations of Clips and Models, ensuring that they can be easily loaded directly via nodes such as 'CheckpointLoaderSimple' and others, without the need for additional configuration.As Sampler Node you can use the ones already present for SDXL using CFG 1, or you can use 'SamplersCustomAdvanced' that is dedicated to FLUX.If you use the dedicated FLUX node above , you can use one of the 'Guiders' as nodes to manage the guidance instead of the 'CFG of the sampler'if not, you can use the 'FluxGuidance' node placed between the prompt and the sampler to manage the guidance on the other samplers already present for SDXL, keeping their CFG at 1 and modifying that of the new node.By reducing the final size by about 30%, these models are also accessible on less powerful machines, although I am not sure how this is possible. In my tests, I have not observed any differences in the generated outputs compared to the full versions.Let me know more with your experiments.I claim no credit for the original creation and publication of these models; all rights are reserved to 'Black Forest Labs' and their talented team. My goal is to make their excellent work more accessible to the community.I hope these compact Flux models will be useful to you and contribute to your creative and technical endeavors. Your feedback is greatly appreciated as we continue to explore and refine these tools together.Flux.1-Dev CompactMAIN SAMPLER: Euler + SimpleNORMAL CFG: 1FLUX GUIDANCE: 2+BETTER STEPS: 20+Flux.1-Schnell CompactMAIN SAMPLER: Euler + SimpleNORMAL CFG: 1FLUX GUIDANCE: 2+BETTER STEPS: 4+VERSIONS CONTENTS:Flux.1-Dev fp8<ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp8_e4m3fn</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul>Flux.1-Dev fp16<ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp16</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul>Flux.1-Schnell fp8<ul><li>Unet: Flux.1-Schnell</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp8_e4m3fn</li><li>Clip2: clip_l</li><li>VAE: ae_Schnell</li></ul>Flux.1-Schnell fp16<ul><li>Unet: Flux.1-Schnell</li><li>weight_dtype: fp8_e4m3fn</li><li>Clip1: t5xxl_fp16</li><li>Clip2: clip_l</li><li>VAE: ae_Schnell</li></ul>Flux.1-Dev fp16 ALT<ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e5m2</li><li>Clip1: t5xxl_fp16</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul>Flux.1-Dev fp8 ALT<ul><li>Unet: Flux.1-Dev</li><li>weight_dtype: fp8_e5m2</li><li>Clip1: t5xxl_fp8_e4m3fn</li><li>Clip2: clip_l</li><li>VAE: ae_Dev</li></ul>

Files with SHA256: e0186741521463dc38d3ed6f5119d95146fb847176015758493d60e4650ca3d0