Saliou カン retweetledi
Saliou カン
467 posts
Saliou カン
@Saliou_K
Nagoya University graduate, cofounder of @FotographerAI @FotographerAIjp (https://t.co/vQZLByJ3FV) and AI enthusiast in general
Nagoya-shi Chikusa, Aichi Katılım Ağustos 2013
262 Takip Edilen167 Takipçiler
Next feature is this
🛋️ Bring Your Own Furniture (BYOF)
Upload any furniture pic and place it anywhere and Interior AI will place it properly in your room!
@levelsio@levelsio
Okay this was too fun to not make it into a feature 😊 [ 📼 Make before/after video ] You can run it on any interior design and it uses your original interior photo and the design Interior AI made and combines them Live now :D
English
Saliou カン retweetledi
Now you can rearrange the layout of the room while preserving its details with our #AIDirector !!
#FotographerAI #AIVisuals #furniture
English
Saliou カン retweetledi
Saliou カン retweetledi
Latest Post on "Keeping up with ZenCtrl Inpaint". You have a TV and are trying to figure out which tv stand or which living room style would fit the best ? Or Vice versa ? it has you covered
English
Saliou カン retweetledi
Saliou カン retweetledi
We will start showing different use cases on How our latest inpainting model from the Zenctrl Series could be used , starting from ready made examples to give you ideas . Still available on @huggingface
English
You can try the beta version for free while it's still on @huggingface : huggingface.co/spaces/fotogra…
English
i might not be the first saying this nor the last , but inpainting might be the answers to every headache we've had. just imagine having the perfect scene , and wondering how you would be able to place your subject without messing up the quality and fidelity of the image?
English
adding to this : try talking to businesses about using Flux or GPT and hear their answers. You might think at this point "That's it" , but details we would neglect are a "no go" for them
Kane@Kanectai
One of the biggest challenges in Visual Content generation AI is the due to the inherent stability (style, colors etc...) of foundational models. See how these models return different styles with the same prompt (duh): "A crisp macro photograph of a single dandelion seed delicately balanced on a dewy blade of grass at dawn. In the softly blurred background, place a tiny red Elmo figurine raising one hand in cheerful surprise. .." Product placement tasks would benefit a lot if we could use all text2img models at the same time!
English
Saliou カン retweetledi
We kinda solved this problem with our new model: ZenCtrl-Inpaint:
- Use any model to generate your backgrounds
- You can use real scenes
- Perfectly inpaint your subject in the scene
Try the interior version here: still a WIP tho
huggingface.co/spaces/fotogra…
English
Saliou カン retweetledi
Also testing texture fidelity improvements and higher resolutions higher details conditioning plus size conditioning (memory is the issue atm haha):
of course the input is AI generated.
Kane@Kanectai
We kinda solved this problem with our new model: ZenCtrl-Inpaint: - Use any model to generate your backgrounds - You can use real scenes - Perfectly inpaint your subject in the scene Try the interior version here: still a WIP tho huggingface.co/spaces/fotogra…
English
RT @FotographerAI: Our CTO @Kanectai said it perfectly ! dealing with the same single problem for a few years now and after many iterations…
English
Saliou カン retweetledi
We've just released a beta version of our inpainting model based on ZenCtrl and it looks like a game changer when fully released.
Early tests (especially on furniture) are showing impressive consistency. It'S just getting started.
Try it FREE on @huggingface playground:
English
Pretty excited to see how the community is using ZenCtrl 🙌
AI速報【AI&ChatGPT最新情報】@AIMIRAI46487
FotographerAIは、画像内のオブジェクトをテキスト(プロンプト)の指示で様々な背景内に登場させたり、人物に試着させたりといったことのできる、オープンソースのAIツール「ZenCtrl」を公開しました。 無料のデモも用意されています。 Tシャツを男性モデルに試着させた例(↓)
English
Saliou カン retweetledi
cosmetics work well , Keep trying more ! will share examples
Mitsu@MFTGAI
Just tried out #FotographerAi’s new #ZenCtrl — generated a background based on a Chanel perfume, and the result was stunning. No text distortion, amazing glass texture, and faithful detail reproduction!! #AIart #generativeart
English
Saliou カン retweetledi
we've just gone ahead and added more samples for try-ons , looking forward to seeing you try them 🙏
AK@_akhaliq
ZenCtrl One framework to Generate multi-view, diverse-scene, and task-specific high-resolution images from a single subject image—without fine-tuning
English
Saliou カン retweetledi
Saliou カン retweetledi