The Loss is a field with holes 🕳️⛳️
🛞 Batch Size is your wheel size.
🏎️ Learning Rate is your speed.
21
Pierre Colle retweeted
✨ Lights on, AI activated ✨
Upload ANY photo with lamp(s).
Watch this app instantly illuminate it.
No Photoshop. No editing skills needed.
Just pure AI magic.🔥
Try Finegrain Light Switcher now.
Pierre Colle retweeted
- Synthetic part of the dataset by nfinite.ai - huggingface.co/datasets/Nfin…
Trained by @Piercus =)
Pierre Colle retweeted
Amazing Background Removal Tool, Absolutely Free!
The detail this tool provides is insane!
buff.ly/4gpsq1J
Pierre Colle retweeted
Super COOL - You no longer need any paid app to do this.
Finegrain Box Segmenter Model - High Definition (1024x1024) object-cutout model, making no assumption on what is background and what is foreground: users are fully in control.
Published under the MIT license
---
Existing off-the-shelf solutions, has some drawback:
- Different people will have different definitions for background and foreground in a given image - a way to prompt these models is missing.
- On the other hand, new promptable approaches like SAM or SAM2 don't meet the quality bar for the use cases we are pursuing: they are generating internally a 256x256 low resolution mask - with built-in upscaling mechanisms that create artefacts and struggle with complex masks (a la Eiffel Tower).
The Finegrain Box Segmenter avoids these pitfalls by training MVANet
Pierre Colle retweeted
finegrain-object-cutter、
気になったので3サービスで比較してみた
・Photoshop
・Canva
・finegrain-object-cutter
結果は画像のとおりですが、
普通に凄すぎて感動してます。
これ背景切り抜きで最強なのでは...?
(Photoshop初心者なので、私が使い方下手なだけの可能性もあります)
📝画像編集に苦戦している人は全員「テキスト入力で任意オブジェクト切り抜きサービス」を使ってみてほしい
- 数秒で正確な切り抜きが完了
- テキスト プロンプトまたは境界ボックスを使用して、切り抜き
- 無料でオンライン利用可能
今すぐ試して、画像編集の効率UP!
teapawt.org/i/status/1833520979479…
Pierre Colle retweeted
Object Cutter
Create high-quality HD background removal for ANY object in your image with a text prompt or bounding boxes!
Pierre Colle retweeted
X : You talk about conversational programming ...
Me : ... yes? Is there a question?
X : Like github copilot?
Me : Like Friedman / Negroponte - graphical conversation theory.
X : I don't understand.
Me : Ok, the building of something can be considered as ...
Pierre Colle retweeted
How do LLMs gain the ability to perform complex reasoning using chain-of-thought?
@Francis_YAO_ argues it's a consequence of training on *code* - the structure of procedural coding and OOP teaches it step-by-step thinking and abstraction.
Great article!
yaofu.notion.site/A-Closer-L…
Pierre Colle retweeted
Fully-funded PhD in my lab with @UMontrealDIRO & @Mila_Quebec: "Neuro-inspired cognitive architectures for artificial metacognition and social learning" #PleaseRT #JobAlert
Cette année, 2 coups de coeur : Intrication quantique : un étrange Aspect de la matière. franceculture.fr/emissions/l… et Automatisation : y aura-t-il un pilote dans l'avion ? franceculture.fr/emissions/l….
🎄 Pour notre dernier jour de calendrier, on a mis les petits plats dans les grands 🍽️
🥳 car on vous propose de venir rencontrer l'équipe de @lamethodeFC et assister à une émission en direct !!
Vous avez jusqu'au 25 à minuit pour participer 😍
Suivez bien les règles du jeu ⤵️
Pierre Colle retweeted
We're living in a cyberpunk future:
“Fooling automated surveillance cameras: adversarial patches to attack person detection” arxiv.org/abs/1904.08653
Very glad to be invited by #AWS to share 🍕 and 🍻 at StationF on 24th April 🚀 #FrenchTech #STATIONF aws-stationf-klap.splashthat… youtu.be/lUvEfsuz4Hw
Pierre Colle retweeted
La joie du vainqueur du challenge Startup Architecture of the Year, @Piercus de @getklap ! #AWSSummit Paris
Paris #AWSSummit today !
I will talk around 1pm at Startup Central to present @getklap's Serverless Video Editing Architecture 👷 ☁️🌐
Pierre Colle retweeted
Learning deep neural networks in blind deblurring framework. arxiv.org/abs/1903.02731
👶First release of npmjs.com/package/hasard 🍾
"Random variables and random nested objects manipulation in javascript"
🙏github.com/Mattasher/probabi… @MattAsher and github.com/aleju/imgaug @alexjungaalto for the inspiration
🧐Feedback appreciated 🙏😀
Great 😀 So now we can work on THE ultimate question ... Is 4-4-2 better than 4-3-3 ?
Today we are excited to release the MuJoCo Soccer environment and accompanying paper, Emergent Coordination through Competition: arxiv.org/abs/1902.07151
Environment: git.io/dm_soccer
Sample gameplay: sites.google.com/view/emerge…
Pierre Colle retweeted
💪🏻 Yes, I made the cover of Cortex !!! 🧠
🎨 Artwork: Face-to-phases (2013)
📝 Paper associated: authors.elsevier.com/a/1YToC…
👉 More info on my #ArtScience activity: slideshare.net/gdumas/brains…


































