Igor Carron
@igorcarron.bsky.social
LightOn, President
GenAI for enterprise, LLMs
Former Nuclear and Rocket Scientist
Nuit Blanche writer (Compressive Sensing, Matrix Factorizations to DL). Paris Machine Learning meetup cofounder
GenAI for enterprise, LLMs
Former Nuclear and Rocket Scientist
Nuit Blanche writer (Compressive Sensing, Matrix Factorizations to DL). Paris Machine Learning meetup cofounder
Pinned
Igor Carron
@igorcarron.bsky.social
· Jan 22
🚀Super proud of the team @lightonai.bsky.social for this new release of Paradigm, the GenAI platform for the Enterprise.
Luminous Lynx marks a major leap forward, integrating advanced agentic 🤖 behavior and cutting-edge tools 🔨✂️
Wanna more ⬇️
lighton.ai/luminous-lynx
Luminous Lynx marks a major leap forward, integrating advanced agentic 🤖 behavior and cutting-edge tools 🔨✂️
Wanna more ⬇️
lighton.ai/luminous-lynx
Reposted by Igor Carron
✊ Transformers... Assemble!
Introducing ♊ Ettin Suite, a SoTA open recipe to outperform existing Generative & Retrieval Models.
Developed by Johns Hopkins University in collaboration with LightOn, Ettin is the first-ever SoTA suite of paired encoder & decoder models.
Introducing ♊ Ettin Suite, a SoTA open recipe to outperform existing Generative & Retrieval Models.
Developed by Johns Hopkins University in collaboration with LightOn, Ettin is the first-ever SoTA suite of paired encoder & decoder models.
July 16, 2025 at 3:57 PM
✊ Transformers... Assemble!
Introducing ♊ Ettin Suite, a SoTA open recipe to outperform existing Generative & Retrieval Models.
Developed by Johns Hopkins University in collaboration with LightOn, Ettin is the first-ever SoTA suite of paired encoder & decoder models.
Introducing ♊ Ettin Suite, a SoTA open recipe to outperform existing Generative & Retrieval Models.
Developed by Johns Hopkins University in collaboration with LightOn, Ettin is the first-ever SoTA suite of paired encoder & decoder models.
Reposted by Igor Carron
From RAG to riches, LightOn introduces FastPlaid, a new milestone for late-interaction architectures!
👏 Congratulations to @raphaelsty.bsky.social, ML Engineer at LightOn, on the release of FastPlaid, a high-performance engine for multi-vector search.
👏 Congratulations to @raphaelsty.bsky.social, ML Engineer at LightOn, on the release of FastPlaid, a high-performance engine for multi-vector search.
June 5, 2025 at 3:41 PM
From RAG to riches, LightOn introduces FastPlaid, a new milestone for late-interaction architectures!
👏 Congratulations to @raphaelsty.bsky.social, ML Engineer at LightOn, on the release of FastPlaid, a high-performance engine for multi-vector search.
👏 Congratulations to @raphaelsty.bsky.social, ML Engineer at LightOn, on the release of FastPlaid, a high-performance engine for multi-vector search.
Another week, another SOTA. This time in Reasoning at @lightonai.bsky.social !
🌐 From Matching to Reasoning — Retrieval just grew a brain.
LightOn introduces Reason-ModernColBERT, a State-of-the-Art multi-vector model purpose-built for the era of Deep Research — where matching isn’t enough, and true insight demands reasoning.
LightOn introduces Reason-ModernColBERT, a State-of-the-Art multi-vector model purpose-built for the era of Deep Research — where matching isn’t enough, and true insight demands reasoning.
LightOn Unlocks Agentic RAG with new SOTA Model Reason-ModernColBERT - LightOn
After the recent release of GTE-ModernColBERT exhibiting state-of-the-art results on long-context retrieval, LightOn announces a major leap forward in AI-driven knowledge discovery: Reason-ModernColBE...
lighton.ai
May 22, 2025 at 4:37 PM
Another week, another SOTA. This time in Reasoning at @lightonai.bsky.social !
Reposted by Igor Carron
💫 LightOn releases GTE-ModernColBERT, the first State-of-the-Art late-interaction model trained on PyLate!
April 30, 2025 at 3:49 PM
💫 LightOn releases GTE-ModernColBERT, the first State-of-the-Art late-interaction model trained on PyLate!
Reposted by Igor Carron
✨ Your know-how is your edge, AI is made to enhance it on your terms!
👀 Would you let AI access data you wouldn’t share with your team?
🛡️Internal governance remains paramount!
This isn’t just smart permissions, it’s safer intelligence — activated
👀 Would you let AI access data you wouldn’t share with your team?
🛡️Internal governance remains paramount!
This isn’t just smart permissions, it’s safer intelligence — activated
April 23, 2025 at 3:24 PM
✨ Your know-how is your edge, AI is made to enhance it on your terms!
👀 Would you let AI access data you wouldn’t share with your team?
🛡️Internal governance remains paramount!
This isn’t just smart permissions, it’s safer intelligence — activated
👀 Would you let AI access data you wouldn’t share with your team?
🛡️Internal governance remains paramount!
This isn’t just smart permissions, it’s safer intelligence — activated
Reposted by Igor Carron
🎭 Your know-how comes in many forms, scattered across countless formats.
💡 What if the insight you seek - hidden in PDFs, PPTXs, XLSXs, PNGs, and JPEGs - was just one question away?
This isn’t just multimodal search, it’s enterprise intelligence — activated
💡 What if the insight you seek - hidden in PDFs, PPTXs, XLSXs, PNGs, and JPEGs - was just one question away?
This isn’t just multimodal search, it’s enterprise intelligence — activated
April 17, 2025 at 4:52 PM
🎭 Your know-how comes in many forms, scattered across countless formats.
💡 What if the insight you seek - hidden in PDFs, PPTXs, XLSXs, PNGs, and JPEGs - was just one question away?
This isn’t just multimodal search, it’s enterprise intelligence — activated
💡 What if the insight you seek - hidden in PDFs, PPTXs, XLSXs, PNGs, and JPEGs - was just one question away?
This isn’t just multimodal search, it’s enterprise intelligence — activated
Reposted by Igor Carron
Join us at #MWC25 in Barcelona next week!
Come discover LightOn latest innovations on Gcore 📌 Booth 6D44 in Hall 6
To know more 👉 lighton.ai/contactus
Come discover LightOn latest innovations on Gcore 📌 Booth 6D44 in Hall 6
To know more 👉 lighton.ai/contactus
February 28, 2025 at 3:38 PM
Join us at #MWC25 in Barcelona next week!
Come discover LightOn latest innovations on Gcore 📌 Booth 6D44 in Hall 6
To know more 👉 lighton.ai/contactus
Come discover LightOn latest innovations on Gcore 📌 Booth 6D44 in Hall 6
To know more 👉 lighton.ai/contactus
Trying to make sense of visual cues in your company documents so that you can have a richer conversation with them? Introducing visual RAG un Paradigm so that you can make the best of your documents 🚀 in a secure and private way. At @lightonai.bsky.social we're shipping fast and often 🚢
February 27, 2025 at 8:13 PM
Trying to make sense of visual cues in your company documents so that you can have a richer conversation with them? Introducing visual RAG un Paradigm so that you can make the best of your documents 🚀 in a secure and private way. At @lightonai.bsky.social we're shipping fast and often 🚢
Reposted by Igor Carron
✨ Introducing the first ever Visual RAG pipeline!
Paradigm now offers a whole pipeline working directly on your visual documents as you see them, preventing any loss of information and allowing you to query non-textual elements!
Paradigm now offers a whole pipeline working directly on your visual documents as you see them, preventing any loss of information and allowing you to query non-textual elements!
February 26, 2025 at 4:51 PM
✨ Introducing the first ever Visual RAG pipeline!
Paradigm now offers a whole pipeline working directly on your visual documents as you see them, preventing any loss of information and allowing you to query non-textual elements!
Paradigm now offers a whole pipeline working directly on your visual documents as you see them, preventing any loss of information and allowing you to query non-textual elements!
🚀Super proud of the team @lightonai.bsky.social for this new release of Paradigm, the GenAI platform for the Enterprise.
Luminous Lynx marks a major leap forward, integrating advanced agentic 🤖 behavior and cutting-edge tools 🔨✂️
Wanna more ⬇️
lighton.ai/luminous-lynx
Luminous Lynx marks a major leap forward, integrating advanced agentic 🤖 behavior and cutting-edge tools 🔨✂️
Wanna more ⬇️
lighton.ai/luminous-lynx
January 22, 2025 at 5:22 PM
🚀Super proud of the team @lightonai.bsky.social for this new release of Paradigm, the GenAI platform for the Enterprise.
Luminous Lynx marks a major leap forward, integrating advanced agentic 🤖 behavior and cutting-edge tools 🔨✂️
Wanna more ⬇️
lighton.ai/luminous-lynx
Luminous Lynx marks a major leap forward, integrating advanced agentic 🤖 behavior and cutting-edge tools 🔨✂️
Wanna more ⬇️
lighton.ai/luminous-lynx
Reposted by Igor Carron
ModernBERT-embed-large is released under Apache 2.0 and is available on Hugging Face:
huggingface.co/lightonai/mo...
Congrats to @nohtow.bsky.social for this great work!
huggingface.co/lightonai/mo...
Congrats to @nohtow.bsky.social for this great work!
lightonai/modernbert-embed-large · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
January 14, 2025 at 4:42 PM
ModernBERT-embed-large is released under Apache 2.0 and is available on Hugging Face:
huggingface.co/lightonai/mo...
Congrats to @nohtow.bsky.social for this great work!
huggingface.co/lightonai/mo...
Congrats to @nohtow.bsky.social for this great work!
Reposted by Igor Carron
LightOn releases ModernBERT-embed-large model, a larger and more proficient iteration of ModernBERT-embed-base!
This competitive fine-tuned variant aims at allowing everyone to play with ModernBERT, becoming an out-of-the-box tool and a starting point for fine-tuning!
huggingface.co/lightonai/mo...
This competitive fine-tuned variant aims at allowing everyone to play with ModernBERT, becoming an out-of-the-box tool and a starting point for fine-tuning!
huggingface.co/lightonai/mo...
lightonai/modernbert-embed-large · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
January 14, 2025 at 4:42 PM
LightOn releases ModernBERT-embed-large model, a larger and more proficient iteration of ModernBERT-embed-base!
This competitive fine-tuned variant aims at allowing everyone to play with ModernBERT, becoming an out-of-the-box tool and a starting point for fine-tuning!
huggingface.co/lightonai/mo...
This competitive fine-tuned variant aims at allowing everyone to play with ModernBERT, becoming an out-of-the-box tool and a starting point for fine-tuning!
huggingface.co/lightonai/mo...
Reposted by Igor Carron
Blog post: huggingface.co/blog/modernb...
Paper link: arxiv.org/abs/2412.13663
With @lightonai.bsky.social folks, we collaborated with @answerai.bsky.social (and friends!) to leverage the advances from recent years of work on LLMs (architecture and training) and apply them to a BERT-style model
Paper link: arxiv.org/abs/2412.13663
With @lightonai.bsky.social folks, we collaborated with @answerai.bsky.social (and friends!) to leverage the advances from recent years of work on LLMs (architecture and training) and apply them to a BERT-style model
Finally, a Replacement for BERT: Introducing ModernBERT
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
December 19, 2024 at 5:36 PM
Blog post: huggingface.co/blog/modernb...
Paper link: arxiv.org/abs/2412.13663
With @lightonai.bsky.social folks, we collaborated with @answerai.bsky.social (and friends!) to leverage the advances from recent years of work on LLMs (architecture and training) and apply them to a BERT-style model
Paper link: arxiv.org/abs/2412.13663
With @lightonai.bsky.social folks, we collaborated with @answerai.bsky.social (and friends!) to leverage the advances from recent years of work on LLMs (architecture and training) and apply them to a BERT-style model
Reposted by Igor Carron
I am thrilled to announce the release of ModernBERT, the long-awaited BERT replacement!
There might be a few LLM releases per week, but there is only one drop-in replacement that brings Pareto improvements over the 6 years old BERT while going at lightspeed
There might be a few LLM releases per week, but there is only one drop-in replacement that brings Pareto improvements over the 6 years old BERT while going at lightspeed
December 19, 2024 at 5:36 PM
I am thrilled to announce the release of ModernBERT, the long-awaited BERT replacement!
There might be a few LLM releases per week, but there is only one drop-in replacement that brings Pareto improvements over the 6 years old BERT while going at lightspeed
There might be a few LLM releases per week, but there is only one drop-in replacement that brings Pareto improvements over the 6 years old BERT while going at lightspeed
Wow!
OK. Practical question times. How are you adjusting your research given progress in reasoning style models? Also how are you adjusting the way you work?
December 22, 2024 at 11:53 AM
Wow!
Thanks Jeremy @howard.fm . Our username on Bluesky is @lightonai.bsky.social
6 years after BERT, we have a replacement: ModernBERT!
@answerdotai, @LightOnIO (et al) took dozens of advances from recent years of work on LLMs, and applied them to a BERT-style model, including updates to the architecture and the training process, eg alternating attention.
@answerdotai, @LightOnIO (et al) took dozens of advances from recent years of work on LLMs, and applied them to a BERT-style model, including updates to the architecture and the training process, eg alternating attention.
December 19, 2024 at 8:23 PM
Thanks Jeremy @howard.fm . Our username on Bluesky is @lightonai.bsky.social
Reposted by Igor Carron
IA générative en entreprise : entre acquérir une solution clé en main et construire en interne, faut-il choisir ?
Laurent Daudet, co-fondateur et CEO de LightOn, explore ce dilemme et ses implications : www.latribune.fr/opinions/tri...
#IA #Innovation #Stratégie #Technologie
Laurent Daudet, co-fondateur et CEO de LightOn, explore ce dilemme et ses implications : www.latribune.fr/opinions/tri...
#IA #Innovation #Stratégie #Technologie
Acheter ou construire sa plateforme d'IA, le dilemme des entreprises
OPINION. Avec l’explosion de l’IA générative, les entreprises se retrouvent face à un dilemme stratégique majeur : doivent-elles construire en interne leur propre plateforme d’IA (build) ou acheter un...
www.latribune.fr
December 16, 2024 at 4:30 PM
IA générative en entreprise : entre acquérir une solution clé en main et construire en interne, faut-il choisir ?
Laurent Daudet, co-fondateur et CEO de LightOn, explore ce dilemme et ses implications : www.latribune.fr/opinions/tri...
#IA #Innovation #Stratégie #Technologie
Laurent Daudet, co-fondateur et CEO de LightOn, explore ce dilemme et ses implications : www.latribune.fr/opinions/tri...
#IA #Innovation #Stratégie #Technologie
Reposted by Igor Carron
🚀 LightOn: A publicly traded company on Euronext Growth
Two weeks ago, LightOn became the first publicly traded Generative AI company in Europe (ALTAI on Euronext Growth Paris).
👉 Learn more: www.lighton.ai/lighton-blog...
Two weeks ago, LightOn became the first publicly traded Generative AI company in Europe (ALTAI on Euronext Growth Paris).
👉 Learn more: www.lighton.ai/lighton-blog...
LightOn Goes Public: A Key Step for European AI - LightOn
LightOn has become the first publicly traded GenAI company with a successful listing on Euronext Growth Paris. The IPO marks a significant step for LightOn to enable Enterprise Generative AI in Europ...
www.lighton.ai
December 10, 2024 at 8:01 AM
🚀 LightOn: A publicly traded company on Euronext Growth
Two weeks ago, LightOn became the first publicly traded Generative AI company in Europe (ALTAI on Euronext Growth Paris).
👉 Learn more: www.lighton.ai/lighton-blog...
Two weeks ago, LightOn became the first publicly traded Generative AI company in Europe (ALTAI on Euronext Growth Paris).
👉 Learn more: www.lighton.ai/lighton-blog...
Reposted by Igor Carron
An added advantage is that, since ArabicWeb is not based on Common Crawl, it can be used together with FineWeb-2 to increase data quantity and diversity
Read more on ArabicWeb24: www.lighton.ai/lighton-blog...
Read more on ArabicWeb24: www.lighton.ai/lighton-blog...
ArabicWeb24: Creating a high quality Arabic Web-only pre-training dataset - LightOn
www.lighton.ai
December 9, 2024 at 3:00 PM
An added advantage is that, since ArabicWeb is not based on Common Crawl, it can be used together with FineWeb-2 to increase data quantity and diversity
Read more on ArabicWeb24: www.lighton.ai/lighton-blog...
Read more on ArabicWeb24: www.lighton.ai/lighton-blog...
Reposted by Igor Carron
The FineWeb-2 study evaluated the quality of various pre-training datasets for the Arabic language.
It's great to see that our work on ArabicWeb is competitive with the state of the art.
It's great to see that our work on ArabicWeb is competitive with the state of the art.
December 9, 2024 at 3:00 PM
The FineWeb-2 study evaluated the quality of various pre-training datasets for the Arabic language.
It's great to see that our work on ArabicWeb is competitive with the state of the art.
It's great to see that our work on ArabicWeb is competitive with the state of the art.
Music was always about nostalgia.
I haven't been a fan of streaming apps so far. I am listening now, for the first time in 30 years, to the album Atom Heart Mother by #PinkFloyd. My heart skipped a beat at the exact moment where my record used to skip, and I had to lift the needle to move on.
Where was this useless memory stored?
Where was this useless memory stored?
December 9, 2024 at 8:46 AM
Music was always about nostalgia.
