Michael Gygli
@GygliMichaelCo-founder and CTO @ Cerrion (YCombinator S22). Work in CV & ML. Previously: Research Scientist @ Google Zurich, Head of AI https://t.co/vlJDsKaErB, PhD @ ETH Zurich
Similar User
@silviocinguetta
@pesarlin
@dlarlus
@VittoFerrariCV
@Jimantha
@lealtaixe
@Awfidius
@abursuc
@jponttuset
@cvondrick
@syhw
@eric_brachmann
@dimadamen
@DavidDuvenaud
@inthebrownbag
Pretty excited to launch the Airtrain.ai playground this week. Chat with a large selection of OSS and proprietary model at once for free. Llama 2 (all sizes), Phi-2, Mistral (7B, Mixtral, Medium), Gemini Pro+Nano, GPT-4, 3.5, Falcon, and more.
Our self-supervised method CycleCL is currently being presented at #WACV2024 ♻️, if you are interested, stop by our poster to chat with @mdestro_
Excited to share our new paper, accepted to #WACV2024 🎉 & done with the amazing @mdestro_ : CycleCL: Self-supervised Learning for Periodic Videos TLDR: We designed an iterative SSL scheme which works on periodic videos, where previous methods fail. arxiv.org/abs/2311.03402
Excited to share our new paper, accepted to #WACV2024 🎉 & done with the amazing @mdestro_ : CycleCL: Self-supervised Learning for Periodic Videos TLDR: We designed an iterative SSL scheme which works on periodic videos, where previous methods fail. arxiv.org/abs/2311.03402
Oops haven't tweeted too much recently; I'm mostly watching with interest the open source LLM ecosystem experiencing early signs of a cambrian explosion. Roughly speaking the story as of now: 1. Pretraining LLM base models remains very expensive. Think: supercomputer + months.…
if i didn't already have a startup, i'd be looking at adept.ai/blog/fuyu-8b right now and thinking about what to build on top of @AdeptAILabs's super cool model
1 day later, there's a paper coming out doing just that: "DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines" arxiv.org/abs/2310.03714 The field moves fast 😮
Great point! Since LLMs function in a specific way, there are specific way to best talk to them! Automatic prompt engineering or compiling a human prompt into a format structured that it fits the format an LLM likes best will come soon
Prompt engineering comes to images! I wonder what else can be done beyond red circles? 🤔 #ICCV2023
An excellent poor man's visual prompt engineering strategy for CLIP: draw red circles on an object in an image auto-magically focuses its attention on that region, leading to a specific embedding #ICCV2023
This afternoon we will present our ViperGPT🐍 paper at #ICCV2023. If you’re in Paris, come talk to us! Oral presentation in room "Paris Sud" and poster 170 in room "Foyer Sud" Project page: viper.cs.columbia.edu (work with @SachitMenon and @cvondrick)
Great point! Since LLMs function in a specific way, there are specific way to best talk to them! Automatic prompt engineering or compiling a human prompt into a format structured that it fits the format an LLM likes best will come soon
The more powerful the models become, the more programs they store. And the benefits of finding the right program (by figuring out the right prompt) become even greater. Unless LLMs change substantially in the future, prompt engineering is not going anywhere. What's going to…
I'm at #ICCV2023 this week. Let me know if you want to chat about research, GenAI and startups
🚀 Exciting News!! Hard work paid off! 🙌 In #MetaConnect2023, MZ showcased our tech enabling high-res image generation in <5 seconds! 🖼️💥 Try it out on WhatsApp, Instagram, and Facebook Huge thanks to our incredible team! @alitabet @artsiom_s @schoenfeldedgar and J. Kohler.
Our new text-to-image model, DALL·E 3, can translate nuanced requests into extremely detailed and accurate images. Coming soon to ChatGPT Plus & Enterprise, which can help you craft amazing prompts to bring your ideas to life: openai.com/dall-e-3
Interesting overview of LLM research directions: huyenchip.com/2023/08/16/llm…
Today, Generative AI takes a big step and comes to user interface design! @helnzhou and I are excited to announce @Galileo_AI : the first AI product that uses natural language to generate UI designs. It lets you design beyond imagination. Early access: useGalileo.ai
Welcome to YC S22, @karimsalehh_, @GygliMichael, @nkobyshev and team at Cerrion! Cerrion helps manufacturers automatically detect problems on their production lines using Computer Vision, enabling production teams to act on deviations faster. ycombinator.com/launches/GxY-c…
In May 2021, I wrote an @FT oped to catalyse a discussion on rewriting the university spinout playbook. We've since crowdsourced deal terms from spinouts all over the world. Today, we’re sharing the @spinoutfyi database: The largest of its kind spinout.fyi/blog/data-laun…🧵
United States Trends
- 1. #UFCMacau 22,6 B posts
- 2. #ArcaneSeason2 179 B posts
- 3. Wayne 94,8 B posts
- 4. Ricci 2.737 posts
- 5. Good Saturday 25,2 B posts
- 6. Leicester 40,8 B posts
- 7. #saturdaymorning 3.396 posts
- 8. Ekko 73,3 B posts
- 9. Jayce 74,8 B posts
- 10. #Caturday 4.208 posts
- 11. #LEICHE 22,5 B posts
- 12. Nicolas Jackson 8.516 posts
- 13. Ndidi 3.620 posts
- 14. Sanchez 232 B posts
- 15. Gabriella Fernandes N/A
- 16. Wang Cong 1.427 posts
- 17. SEVENTEEN 1,45 Mn posts
- 18. Jinx 204 B posts
- 19. Nico Jackson 1.635 posts
- 20. Surgeon General 57,5 B posts
Who to follow
-
Silvio Savarese
@silviocinguetta -
Paul-Edouard Sarlin
@pesarlin -
Diane Larlus
@dlarlus -
Vittorio Ferrari
@VittoFerrariCV -
Noah Snavely
@Jimantha -
Laura Leal-Taixe
@lealtaixe -
Andrew Fitzgibbon
@Awfidius -
Andrei Bursuc
@abursuc -
Jordi Pont-Tuset
@jponttuset -
Carl Vondrick
@cvondrick -
Gabriel Synnaeve
@syhw -
Eric Brachmann
@eric_brachmann -
Dima Damen
@dimadamen -
David Duvenaud
@DavidDuvenaud -
Karteek Alahari
@inthebrownbag
Something went wrong.
Something went wrong.