<h1 id="goddess-project-jsh0rzo3c">Goddess Project</h1>--Formerly Uncensored Females--<h3 id="standalone-checkpoint-goddess-works-in-forge-only-qqxyao9qg">Standalone Checkpoint - Goddess works in FORGE only</h3>DO NOT LOAD a separate VAE, TE, or CLIP unless using GGUFThis version is a mixed precision Flux Dev model, with limited UNET changes to allow for feminine anatomy.<ul><li>Run this model in automatic FP16 Lora mode NOT NF4</li><li>This model is full precision in BF16 UNET with mixed precision (NF4) on the TE blocks</li></ul><ul><li>This model fits in a 24GB card and could be run in GPU only mode as such</li><li>High speed BF16 with a slightly lower prompt accuracy compared to the 33GB full model</li></ul><h2 id="links-(for-gguf-only)-crdhvs5ak">Links (For GGUF ONLY)</h2><ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/comfyanonymous/flux_text_encoders/tree/main">Text Encoders</a></li><li><a target="_blank" rel="ugc" href="https://huggingface.co/zer0int/CLIP-GmP-ViT-L-14/blob/main/ViT-L-14-TEXT-detail-improved-hiT-GmP-TE-only-HF.safetensors">Updated CLIP</a> - <a target="_blank" rel="ugc" href="https://huggingface.co/comfyanonymous/flux_text_encoders/tree/main">Standard CLIP-L </a>- <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/892116?type=Model&amp;format=SafeTensor&amp;size=full&amp;fp=fp8">FP8 CLIP-L</a> -- ** <a target="_blank" rel="ugc" href="https://civarchive.com/articles/7668/new-clip-l-fp8-and-flux-comparison">Version Comparison</a> **</li><li><a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.1-dev/tree/main">VAE (AE.safetensors)</a></li></ul>Per the <a target="_blank" rel="ugc" href="https://www.apache.org/licenses/LICENSE-2.0">Apache 2.0 license</a> FLAN is attributed to Google<hr />This model is a training using many individuals with known ages and 2257 forms, it has also been merged to try and ensure that no known individuals can be reproduced. However FLUX seems to like to learn faces even with less then 10% data rather then merge them into a new face.

Godess Project (FLUX)

godessProjectFLUX_fluxDevUFV7FP16.safetensors

FLUX_Dev_UFV7-FP16

Goddess_Dev_Checkpoint

CLIP-L-FP32

FLUX_Schnell_V7-NF4

FP8-V4

GGUF_Dev_V9

GGUF_Dev_V7-Q8

CLIP-L (FP8)

GGUF_Schnell_V7_Q8

NF4_Schnell_Blockwise_V9

Uncensored Females (FLUX) Update

Uncensored Females (FLUX)Standalone Checkpoint DO NOT LOAD a separate VAE, TE, or CLIP unless using GGUFVersion 7Google FLAN T5xxl quantized from Full FP32Reduced weights of some trainings to allow for more face variationNipple, Anus and Vulva details could use finetuning (Message me if you want to use this model as a base)NF4 Schnell has better performance and quality in my opinion then FP8 (Forge)However FP8 version is viable for COMFY UILinks (For GGUF)Note Google FLAN model has some difficulty interfacing with GGUF and is only available in NF4/FP8 currentlyText EncodersUpdated CLIP - Standard CLIP-L - FP8 CLIP-L -- ** Version Comparison **VAE (AE.safetensors)ProsAges 18-40yo trained on models with 2257 forms.Tested with Adult rated Character models and works well with the ones listed.NF4 vs FP8 speeds listed below in my testing (8GB VRAM)ConsFlux tends to draw the same face this model is no exceptionX rated Males still not trained into V5 (Without Loras)XXX has no support (Without Loras)Per the <a target="_blank" rel="ugc" href="https://www.apache.org/licenses/LICENSE-2.0">Apache 2.0 license</a> FLAN is attributed to GoogleMy Speed on a 3050 8GBNF4 (Tested in Forge)832x1216 @ 4.5 Seconds per IT1024x1024 @ 5 Seconds per IT1024x1536 @ 7.5 Seconds per IT2048x2048 @ 110 Seconds per IT (Comfy wins hands down at 2048 with FP8)FP8832x1216 @ 5.5 Seconds per IT1024x1024 @ 6 Seconds per IT1024x1536 @ 15 Seconds per IT2048x2048 @ 28 Seconds per ITFP16/BF16832x1216 @ 8 seconds per IT1024x1024 @ 10 seconds per IT1024x1536 @ 11 seconds per IT (?) Why is it faster then FP8 maybe the cfg?This model is a training using many individuals with known ages and 2257 forms, it has also been merged to try and ensure that no known individuals can be reproduced. However FLUX seems to like to learn faces even with less then 10% data rather then merge them into a new face.

uncensoredFemalesFLUX_fluxDevUFV7FP16.safetensors

Files with SHA256: 9c0c6f8ccb380897632c4b436f419d7034df171305b29be5f110b2918debd693