Leonard Volner รีทวีตแล้ว
That's us! 🌍
The Artemis II crew captured beautiful, high-resolution images of our home planet during their journey to the Moon. As @Astro_Christina put it: "You guys look great."
English
Leonard Volner
1.8K posts
Leonard Volner
@LeonardVolner
Entrepreneur. AI-dabbler and enthusiast. Engineer. Co-founder of @northdispense. Drive Thru Guru. 1x exit. 2x dad. SMB searcher. Real estate developer.
St. Louis Area เข้าร่วม Temmuz 2023
1.4K กำลังติดตาม583 ผู้ติดตาม
Leonard Volner รีทวีตแล้ว
mesh-llm: pool compute to run open models. built by @michaelneale at block: docs.anarchai.org
English
Leonard Volner รีทวีตแล้ว
Leonard Volner รีทวีตแล้ว
RF-DETR is the best open-source choice if you work with aerial or satellite images
in this specific project, RF-DETR Medium outperformed YOLO26 fine-tuned on the same dataset by 12 mAP
in object detection, that is a huge gap
link: github.com/roboflow/rf-de…
SkalskiP@skalskip92
i’m working on geo rf-detr this time i fine-tuned rf-detr on neontreeevaluation dataset; it gives you insight into forest health, tree counts, and land changes without manual field surveys what other aerial / drone use cases should i try next?
English
Leonard Volner รีทวีตแล้ว
We are releasing Falcon Perception, an open-vocabulary referring expression segmentation model. Along with it, a 0.3B OCR model that is on par with 3-10x larger competitors.
Current systems solve this with complex pipelines (separate encoders, late fusion, matching algorithms). We developed a novel simpler "bitter" approach: one early-fusion Transformer (image + text from first layer) with a shared parameter space, and let scale + training signal do the work. Please check our work !
📄 Paper: arxiv.org/pdf/2603.27365
💻 Code: github.com/tiiuae/falcon-…
🎮 Playground: vision.falcon.aidrc.tii.ae
🤗 Blogpost: huggingface.co/blog/tiiuae/fa…
English
Leonard Volner รีทวีตแล้ว
We just got this Force Majeure letter today from AirGas, our helium supplier (for our food science lab, where we have multiple mass-spec instruments that use helium).
The letter says that helium supplies are cut off, and if you're lucky, you might be allotted HALF the helium you need. Even then, you will be charged extra for any helium you get. A LOT extra.
So basically, every mass spec lab in America is about to go offline. AirGas is expressly invoking FM and saying they cannot meet their contractual obligations. Not their fault. Trump did this by attacking Iran.
My lab is fine, of course, because I saw this coming and I ordered my lab staff to buy a one-year supply weeks ago. We already have it in place. So we're still up and running with plenty of helium.
But very few lab science people are paying attention to the Strait of Hormuz, so they are getting blindsided by this.
Trump's war is shutting down science labs all across the country right now. Don't dare call this "winning." It's a loss for America. And the world.
English
Leonard Volner รีทวีตแล้ว
LLMが遂にテキストを脱皮し「現実世界」の3D空間をガチで理解し始めた。
SpatialLMは3D点群にLLMの推論力を適用し、壁やドアの構造、空間の繋がりを解釈して構造化データを出力する。
言語モデルを空間知能へ拡張するブレイクスルー。コードも論文もオープンに公開。
日本語
Leonard Volner รีทวีตแล้ว
Leonard Volner รีทวีตแล้ว
@cohere transcribe
Sota open source transcription model running in the browser :)
Weights on @huggingface link below
English
Leonard Volner รีทวีตแล้ว
Leonard Volner รีทวีตแล้ว
NASA is sending humans to fly around the Moon—and you can watch! 🚀
Live coverage of Artemis II begins today. Full schedule: go.nasa.gov/4dekJwq
For the latest updates on Artemis II, follow @NASA and @NASAArtemis.
English
Here's the github repo where I'm working on something similar. It's not thr latest version here yet but will update with the latest version within the next couple days.
github.com/MisterEase13/P…
Dan McAteer@daniel_mac8
This is amazing. Do this.
English
I wish. But at this stage, I'm still pretty amateur vibe coder. IIRC, I have an early version pushed to Github but not the latest. I ran out of tokens before I could tell claude code or antigravity to commit. It's been basically sitting back burner since that day.
I'm out and about right now, but will try to push the latest version tonight or tomorrow.
github.com/MisterEase13/P…
English
@KnowledgeNewfie @daniel_mac8 Example- User says they have no children and no plans for them, the "goal ontology mapper" shuts off all further lines of questioning related to child-rearing goals. E.g. don't ask the user if they plan to save for child's college or house or business fund.
English
English
Make intellectual pursuits sexy again and watch the world change for the better.
Travis Akers 🇺🇸@travisakers
We’re going to the Moon this week for the first time in 53 years and it’s not even a lead story.
English
This is my modus operandi. Even when I think I'm certain of something, I always ask the models to challenge my assumptions and surface the most compelling reasons why I could be wrong.
Just make sure you avoid the analysis paralysis attractor state.
Andrej Karpathy@karpathy
- Drafted a blog post - Used an LLM to meticulously improve the argument over 4 hours. - Wow, feeling great, it’s so convincing! - Fun idea let’s ask it to argue the opposite. - LLM demolishes the entire argument and convinces me that the opposite is in fact true. - lol The LLMs may elicit an opinion when asked but are extremely competent in arguing almost any direction. This is actually super useful as a tool for forming your own opinions, just make sure to ask different directions and be careful with the sycophancy.
English
If I've learned one thing about X, it's that the timeline is legit at least two weeks ahead of the news cycle.
English