Jan Margeta
@jmargetaMaker of Medical Imaging and Computer Vision tools 🤖 Founder @KardioMe 🫀 Alumnus VIBOT II, ViCOROB, CISTIB, Asclepios, MSR #ML #AI #Python #WebDev #3Dprinting
Similar User
@BernhardKainz1
@endo_vis
@cherise_go
@jelmerwolterink
@ezeq_dlrosa
@mathemage
@oscarcamararey
@techevents_eu
@IslesChallenge
@Sky_LeiLi
@hbogunovic
@OpenCampBA
@rmartimarly
We've built a comprehensive cochlear image analysis tool to enable large scale studies, enabling access to cutting-edge #ML from the browser How? We detail the pipeline in our paper that has just been published in the Journal of Clinical Medicine: mdpi.com/2077-0383/11/2…
uv - the best thing happening to Python tooling these days. Thanks for the extra attention to the needs of Torch users 🚀
The latest uv release includes support for conflicting dependencies across optional groups. A subtle but very powerful feature. For example: use the PyTorch CPU build with `uv sync --extra cpu` and the CUDA build with `uv sync --extra gpu`. All powered by a single lockfile.
Could not be more excited than seeing the one and only Paula López Diez @PaulaLopezDiez defending her PhD thesis on deep learning for implantable devices. Starts in a minute: dtudk.zoom.us/j/65102271918?… 🐌🫀 Gogogo Paula, you rock! 💪
💯! Plus nearly instant virtual env creation even when slow or no internet connection is available. Heavy caching and the --offline flag make `uv` a great travel companion for Python devs on the go ✈️
Been using uv for several months. I enjoy the heavy caching and the fact they hard link to it. Of course it makes creating a venv fast. But one of the little talked-about benefits is that if you have the same dep in 100 venv, you have only one copy and it saves a ton of space
Smooth sailing with @rustdesk today. My Linux machine connected to a Windows desktop like a charm! 🚀 This is simply the perfect alternative to the excellent TeamViewer! Open-source, fast, and always improving
Don't miss out on the latest additions to the high fidelity robot models in MuJoCo menagerie such as the humanoid Unitree G1 or the bi-manipulator ALOHA 2 🤖 The slick updated gallery by @kevin_zakka makes the discovery and access much simpler, thanks!🙏 github.com/google-deepmin…
🙏
Sunday rant. For software engineering, my sense is that the phrase “premature optimization is the root of all evil” has massively backfired. Its from a book on data structures and mainly tried to dissuade people from prematurely write things in assembler. But the point was to…
Truly magical work @Snosixtytwo 👏 and what a walk around the extraordinary research site of @inria_sophia! Feels like being there again! Happy to get a glimpse of my former PhD window in Fermat represented by more than a few Gaussians 🤩 Great times! See you soon @InriaEpione?
Happy to announce the results of our latest research, which takes 3D Gaussian Splatting to the next level: "A Hierarchical 3D Gaussian Representation for Real-Time Rendering of Very Large Datasets," which has been accepted at #SIGGRAPH2024!🎉 Find it here: repo-sam.inria.fr/fungraph/hiera…
Brilliant! There are times when tectonic shifts in ML happen.
MLPs are so foundational, but are there alternatives? MLPs place activation functions on neurons, but can we instead place (learnable) activation functions on weights? Yes, we KAN! We propose Kolmogorov-Arnold Networks (KAN), which are more accurate and interpretable than MLPs.🧵
🫀A top research opportunity with @WenjiaBai and an awesome team at Imperial! 🫀 So much can be discovered about the cardiac shapes and motion when powerful 3D+t models like CHeart and the massive UK Biobank dataset combine.
❤️ New Job ❤️ An exciting opportunity to join our team @BioMedIAICL @ImperialDSI @ImperialBrains funded by @TheBHF New Horizons Grant. Develop machine learning and 3D shape analysis techniques for population-level cardiac imaging data. Advert: imperial.ac.uk/jobs/search-jo…
Embedding personal 📚 @zotero collections and 📝 @logseq notes with 🦙 @ollama makes a fantastic research companion ❤️ Thanks to these open-source champions, using my favorite tools together is now simpler than ever 🐣 Let's build a distributed database of knowledge.
Ollama supports embedding models! Bring your existing documents or other data, and combine it with text prompts to build RAG (retrieval augmented generation) apps! Learn more: ollama.com/blog/embedding…
PyCon Slovakia is happening this coming week. Grab your tickets while they last: 2024.pycon.sk/en/index.html See you soon, friends of Python at @pyconsk ! #pyconsk
Calibrating classifiers confidence scores to better approach the true probabilities is the final chef's touch on nay trained ML model. If you are in Paris, come and see Guillaume's talk, he has always some excellent ML tricks up his sleeves 🧑🍳 github.com/scikit-learn/s…
🎉 We are thrilled to unveil our next speaker for PyData Paris meetup! 🌟 Guillaume Lemaitre (@glemaitre58), Scikit-Learn maintainer and software engineer at @probabl_ai, as he dives into scikit-learn classifiers. Sign up for March 21st at meetup.com/pydata-paris/e…
When adopted by many, efficient tools might meaningfully lower CO2 emissions and save countless dev lives too. Who knows, maybe we'll even build some new wonders of the world with all the spare time! x.com/Andercot/statu… Charlie, you rock! Thanks for all the Rust magic 🫡
HuggingFace is now using uv in CI for huggingface_hub 🤗🤗🤗 Save 30-40s per workflow with this one weird trick...
This week, two hugely inspirational works challenged my views (pun intended) on 3D reconstruction and pose estimation. DUSt3R takes the images & estimates the depths and poses to generate a full 3D recon Without intrinsics and even with 0 overlap 🤯 magical! Bravo & thank you!
An example of how DUSt3R can do "impossible matching": given two images without any shared visual content (my office, obviously never seen at training), it can output an accurate reconstruction (no intrinsics, no poses!) in seconds
The best way to deeply understand the MRI sequences is to code some. 🧲 Pulseq - a cross-vendor sequence programming and simulation framework in Python and the associated course look extra attractive Well done!👏 Now, if only I had a spare MR machine for real world testing😌
Over-representing the cameras with sets of rays and denoising them to estimate the poses is so clever. Almost obvious in hindsight. Beautiful work!! + seeing Plücker (1828) in a modern computer vision work is such a great reminder of all the Giants on whose shoulders we stand 🙏
[1/6] What representation comes to mind when you think of a ‘camera’? Perhaps an extrinsic + intrinsic matrix? In our ICLR (oral) paper, we instead infer a distributed representation where each pixel is associated with a ray, and show SoTA results for few-view pose estimation.
United States Trends
- 1. Thanksgiving 2,4 Mn posts
- 2. Bears 61,9 B posts
- 3. Caleb Williams 5.888 posts
- 4. Thankful 462 B posts
- 5. #IDEGEN N/A
- 6. Gibbs 13,4 B posts
- 7. Eberflus 6.185 posts
- 8. Turkey 300 B posts
- 9. #CHIvsDET 5.283 posts
- 10. Ben Johnson 2.754 posts
- 11. #OnePride 8.858 posts
- 12. Penei Sewell 2.462 posts
- 13. Goff 6.866 posts
- 14. #NationalDogShow N/A
- 15. DJ Moore 2.068 posts
- 16. #Picksgiving 6.644 posts
- 17. Tyrique Stevenson N/A
- 18. Laporta 5.306 posts
- 19. Tony Romo N/A
- 20. Grateful 284 B posts
Who to follow
-
Bernhard Kainz
@BernhardKainz1 -
Endoscopic Vision (EndoVis) Challenge
@endo_vis -
Chen (Cherise) Chen
@cherise_go -
Jelmer Wolterink
@jelmerwolterink -
Ezequiel
@ezeq_dlrosa -
Karel Ha
@mathemage -
Oscar Camara
@oscarcamararey -
Techevents_eu
@techevents_eu -
ISLES-challenge
@IslesChallenge -
Lei Li
@Sky_LeiLi -
Hrvoje Bogunović
@hbogunovic -
Bratislava OpenCamp
@OpenCampBA -
Robert Martí
@rmartimarly
Something went wrong.
Something went wrong.