CivArchive
    Preview 139706510
    Preview 139706677
    Preview 139706690
    Preview 139706710
    Preview 139706851
    Preview 139706852
    Preview 139706855
    Preview 139706999
    Preview 139707000
    Preview 139707005
    Preview 139707045
    Preview 139707044
    Preview 139707049
    Preview 139707112
    Preview 139707109
    Preview 139707117
    Preview 139707137
    Preview 139707138
    Preview 139707139
    Preview 139707174

    Ray Artshoot Krea2 NSFW V4:

    v4 comes in both BF16 and INT8 precision builds. I can’t run the BF16 version on my local hardware, but it works flawlessly on online platforms. Feel free to test it out if you’re looking for the highest possible image quality.I’ve fixed several known issues from V3, readjusted the bypass weights and NSFW generation performance, and integrated the Ray Pantyhose fix Lora.

    Thanks for all your support, and enjoy creating!

    Also, I’ve created a ComfyUI node extension for Krea2-style switching; if you're interested, you can download it here:

    https://civarchive.com/models/2856809/ray-style-switching-extension

    v4提供BF16和INT8版本,我本地不能运行BF16但是可以在线使用,对画质有要求的bro可以多测试。修复了一些V3以知的问题,重新调整旁路和NSFW内容,以合并Ray Pantyhose fix Lora。

    感谢大家的支持,祝你们玩的开心。

    还有,我做了一个krea2风格切换的comfyui节点扩展,有需要的请到这里下载:

    https://civarchive.com/models/2856809/ray-style-switching-extension


    Ray Artshoot Krea2 NSFW V3:

    V3 resolves the overly glossy, wet-texture rendering issue, refines NSFW generation with improved convergence and stability, and boasts far better prompt comprehension. It follows your prompts much more faithfully and tones down the hyper-photorealistic look of earlier versions.

    This release ships as an INT8 quantized checkpoint, delivering a 30% speed increase for local inference. With so many LoRA models already available in the ecosystem, we set out to build a base checkpoint that stays closer to the original model’s native behavior, offers more reliable prompt adherence, and produces cleaner, more consistent outputs.

    Enjoy creating!

    v3解决很湿的问题,收敛nsfw性能,更加懂你的提示词,不再那么写实。

    而且v3是int8版本,在我本地运行速度上升30%。现在已经有很多lora了,我们只需要更接近原版和更听话更干净的检查点。

    祝你们玩的开心。


    Ray Artshoot Krea2 NSFW V2:

    After being refined using 5k Magnum Photos-style images, the AI-like feel has been reduced, resulting in better light and shadow effects and a deeper understanding of photography. This version has fully unlocked NSFW functionality in a comprehensive manner. Its goal is not to focus on adult content generation, but rather to fully unleash creative potential and expand the possibilities of creation.

    经过5k张马格南图片社风格的洗礼,减少AI感,更好的光影表现,更懂摄影。此版本已完整解锁 NSFW 相关权限,但其核心定位并非打造成人内容模型,而是为了最大限度释放创作的边界与可能性。


    Ray Artshoot Krea2 NSFW V1:

    Krea2 delivers excellent performance and fits my workflow perfectly. It supports multimodal capabilities, covers a full range of artistic styles, and offers strong expressive quality. This version has fully unlocked NSFW functionality in a comprehensive manner. Its goal is not to focus on adult content generation, but rather to fully unleash creative potential and expand the possibilities of creation.

    Krea2 表现出色,高度契合我的使用习惯。它支持多模态能力,覆盖全面风格,艺术表现力突出。此版本已完整解锁 NSFW 相关权限,但其核心定位并非打造成人内容模型,而是为了最大限度释放创作的边界与可能性。


    Ray Artshoot ZIB SFW V1:

    This is a fully SFW checkpoint! No NSFW content!

    It features significant aesthetic upgrades with a unique artistic style of its own, and it works universally across all art styles and functional generation scenarios.

    It feels like I’m the only one in the community who still hasn’t given up on Z Image Base.

    这是一个SFW检查点!仅有原版Base级别的NSFW内容!

    美学方面提升很大,同时又有自己的风格。能影响到所有的风格和功能。

    感觉现在整个社区只有我还在没有放弃Z image base。

    If you find this checkpoint useful and wish to support my work, or if you only need the standalone LoRA version, please consider donating or purchasing the LoRA on TensorArt.

    All donations will be reinvested into developing higher-quality checkpoints for everyone. I’m also eagerly looking forward to the open-source release of the Z Image Edit version.

    所以,如果觉得好用,或者只想使用lora版本请到Tensorart打赏我或购买lora。我会利用这份捐助给大家带来更好的检查点的,我现在非常期待Z image edit版本开源。

    TensorArt review process is becoming increasingly strict, Therefore, I will gradually move my new models to civitai and seeart. Reproduction is permitted, but please indicate the source. Thank you for your attention.

    Tensorart 的审核日益严格,所以我会把我的新模型陆续搬到civitai和seeart。可以转载,转载请注明出处。感谢关注。

    For online use, please visit TensorArt | Seaart

    如需在线使用可以到 TenserArt | Seaart

    Description

    Recommended settings推荐设置: re_sde + simple, 8-16 steps, 1 cfg

    V3 resolves the overly glossy, wet-texture rendering issue, refines NSFW generation with improved convergence and stability, and boasts far better prompt comprehension. It follows your prompts much more faithfully and tones down the hyper-photorealistic look of earlier versions.

    This release ships as an INT8 quantized checkpoint, delivering a 30% speed increase for local inference. With so many LoRA models already available in the ecosystem, we set out to build a base checkpoint that stays closer to the original model’s native behavior, offers more reliable prompt adherence, and produces cleaner, more consistent outputs.

    Enjoy creating!

    v3解决很湿的问题,收敛nsfw性能,更加懂你的提示词,不再那么写实。

    而且v3是int8版本,在我本地运行速度上升30%。现在已经有很多lora了,我们只需要更接近原版和更听话更干净的检查点。

    祝你们玩的开心。

    FAQ

    Checkpoint
    Krea 2

    Details

    Downloads
    947
    Platform
    CivitAI
    Platform Status
    Available
    Created
    8/14/2026
    Updated
    8/21/2026
    Deleted
    -