kevinLee @kevinlan2008
GPT YOLOV MML Joined March 2023-
Tweets410
-
Followers58
-
Following603
-
Likes112
How does @deepseek_ai Sparse Attention (DSA) work? It has 2 components: the Lightning Indexer and Sparse Multi-Latent Attention (MLA). The indexer keeps a small key cache of 128 per token (vs. 512 for MLA). It scores incoming queries. The top-2048 tokens to pass to Sparse MLA.
How does @deepseek_ai Sparse Attention (DSA) work? It has 2 components: the Lightning Indexer and Sparse Multi-Latent Attention (MLA). The indexer keeps a small key cache of 128 per token (vs. 512 for MLA). It scores incoming queries. The top-2048 tokens to pass to Sparse MLA. https://t.co/QzzPRvAaNa
DeepSeek V3.2 breakdown 1. Sparse attention via lightning indexer + top_k attention 2. Uses V3.1 Terminus + 1T continued pretraining tokens 3. 5 specialized models (coding, math etc) via RL then distillation for final ckpt 4. GRPO. Reward functions for length penalty, language…
给大家带来 Qwen3-Next-80B-A3B 新架构的技术解析 说实话这个架构我乍一看立刻想到了——推测性解码，同样都是先用精简架构进行快速生成，然后用复杂但是精度高的架构提升生成质量。但是二者还是有决定性的不同之处的，来给大家捋一捋： Qwen3-Next 是先用 Gated…
$BABA Banger. Cloud accelerated to +26% (+18% last Q), they’re cooking in AI (+“triple digits”), international +19% (almost at break-even now!), total China ecom. +10%. Food/instant delivery war is the main issue here (hurting margin), but $BABA is winning. 萬丈陽光照耀你我
Nice paper that is basically an empirical survey of all the recent RL algorithm improvements.
Just published my new article, "Marketplace": my first attempt at efficient GPU training without backprop 😄🎉 I've been considering eliminating backprop for a while. I had an idea, experimented for two weeks, and it worked! Here's how it works:
m3-agent：带记忆的智能体 字节开源的智能体，能把多模态数据也转化成记忆，包括文本、音频、图片、视频数据。分别微调了2个模型用于记忆和执行，记忆模型基于Qwen 2.5 omni，执行模型基于Qwen 3 32B。 Github：github.com/ByteDance-Seed…
Google 刚刚发布了 Gemma-3-270M 对，你没看错， 270M，不是270B。这是个多模态模型，能接受文本和图片输入，并且输出文本。输入图片会标准化为 896 x 896 分辨率。 官方并没有做过多的介绍。不过这种大小的模型通常可以放在移动端设备运行。 模型地址：huggingface.co/google/gemma-3…
Gemma 3 270M! Great to see another awesome, small open-weight LLM for local tinkering. Here's a side-by-side comparison with Qwen3. Biggest surprise that it only has 4 attention heads!
Gemma 3 270M! Great to see another awesome, small open-weight LLM for local tinkering. Here's a side-by-side comparison with Qwen3. Biggest surprise that it only has 4 attention heads! https://t.co/Iy7O0DsQGu
Let's compare OpenAI gpt-oss and Qwen-3 on maths & reasoning:
Summary of GPT-OSS architectural innovations: 1. sliding window attention (ref: arxiv.org/abs/1901.02860) 2. mixture of experts (ref: arxiv.org/abs/2101.03961) 3. RoPE w/ Yarn (ref: arxiv.org/abs/2309.00071) 4. attention sinks (ref: streaming llm arxiv.org/abs/2309.17453)
OCRFlux：开源最佳PDF识别模型 效果非常好，相比其他OCR有几大核心亮点： 1）能自动跨页合并段落和表格（首个实现该功能的开源模型），即使页面底部有注释，也能自动跳转注释的位置。如下方图1和图2 2）识别精准度高，96%的准确率，明显高于olmOCR-7B 3）3B大小的模型，3090单卡可运行…
ByteDance presents ProtoReasoning Prototypes as the Foundation for Generalizable Reasoning in LLMs
One of the most effective things the U.S. or any other nation can do to ensure its competitiveness in AI is to welcome high-skilled immigration and international students who have the potential to become high-skilled. For centuries, the U.S. has welcomed immigrants, and this…
Gemini 2.5 Technical Report 🧵
Google I/O 发布的 MedGemma 也太酷了，直接可以识别你的 X 光等报告，然后分析存在的问题，感觉大部分医生都啥用了.... 在线体验 demo 👉 huggingface.co/spaces/google/…
I am pleased to announce a new version of my RL tutorial. Major update to the LLM chapter (eg DPO, GRPO, thinking), minor updates to the MARL and MBRL chapters and various sections (eg offline RL, DPG, etc). Enjoy! arxiv.org/abs/2412.05265
苹果发布 FastVLM 能在 iPhone 上直接运行的极速视觉语言模型 首 token 输出快 85 倍 FastVLM：先把图像看懂（图像 → token），再根据 token 生成回答或描述（token → 语言） 模型体非常积小，可以很轻松部署在 iPhone、iPad、Mac 上 FastVLM-0.5B：相较于 LLaVA-OneVision-0.5B， 首 token…
上海AI实验室开源Intern VL3系列模型 原生多模态，支持图片和文本输入、文本输出，包括1B、2B、8B、9B、14B、38B、78B七个参数大小，评分比QWen 2.5 VL高。 模型：huggingface.co/collections/Op…
PatriciaTom @MkG1v457xiWm5
1 Followers 1K Following
_ Kristjan Kullamägi... @Qullammggie
60 Followers 3K Following Trading the U.S. markets from Sweden since 2011. Happy to help! PLEASE READ THIS FIRST: https://t.co/IVkRec7H6d Streaming daily at https://t.co/W0khUriaJA qullamaggie
Emanuel B0eminghaus. @E_Boemingheus
68 Followers 1K Following Seit 21 Jahren IT-Unternehmer – Daten, Märkte, Realitäten - IZW Experte - Unternehmen https://t.co/3IeQ9pNOrD
Salome @Janno2592352
48 Followers 2K Following In order to be irreplaceable one must always be different. — Coco Chanel
Sruiiepaw @Sruiiepaw54015
85 Followers 2K Following
Yjawlu @Yjawlu348112
62 Followers 2K Following
石井惠子 @Uledee701954
5 Followers 173 Following
Charlene @BrianaPfan9444
67 Followers 3K Following
Wasteland Capitals @ecommercesharez
166 Followers 2K Following Escaped the Vampire Squid, surviving in the wasteland. Investing in bottlecaps, US UK EU & global assets. Also, jokes. Liking is not endorsing. Do your own DD.
Wasteland Capital @_ecommerceshare
786 Followers 4K Following Escaped the Vampire Squid, surviving in the wasteland. Investing in bottlecaps, US UK EU & global assets. Also,jokes. Liking is not endorsing. Do your own DD.
Toppis @ToppisM02kXN0
124 Followers 7K Following
Createtiv @createtivtttt
75 Followers 2K Following
Verdie Osinski @VerdieO59006
41 Followers 3K Following
Stephania Wolff @StephaniaW80190
29 Followers 3K Following
Peter John @PeterJohn886003
208 Followers 7K Following
Lessharth @Lessharth8VAOb
36 Followers 1K Following
Thirthee @Thirthee0Av5
17 Followers 231 Following
Laura Bopha Phillips @hucHgk65ywLvOm
6 Followers 199 Following Angel Investors Community has a 1-month free trial for new members. Welcome to join and verify our trading strategy https://t.co/EmErb3hl7J
Virginia @virginia_forema
288 Followers 3K Following
Joan @j_sanders26
256 Followers 3K Following
Cassandra @clabaugh_cassan
288 Followers 3K Following
Julie @rainey77julie
496 Followers 3K Following
Rhonda @s_rhonda85
310 Followers 3K Following
Edna @hall_edna1
266 Followers 3K Following
Bertha @bertha_ferrara2
340 Followers 3K Following
Vicki @vicki_cabiness_
297 Followers 3K Following
Kim @kimwyant34
255 Followers 3K Following
Dale @martin74dale
578 Followers 3K Following
Haihao Shen @HaihaoShen
4K Followers 3K Following Creator of #intel Neural Compressor and AutoRound; HF Optimum-Intel Maintainer; OPEA & COIA TSC; Opinions are my own
Ina @gray2ina
261 Followers 3K Following
June @dewaldjune48
295 Followers 3K Following
Vivian @v_sampere89
219 Followers 3K Following
Marilyn @m_morales65
267 Followers 3K Following
Sayndey @sayndey65207
15 Followers 386 Following
Rosa @rosawilcoxon49
269 Followers 3K Following
Jqiuss☆✨ @Jqiuss
16K Followers 3K Following 学习学习再学习📑。副业：投资交易，吃喝玩乐，..♩`♪♫`♬~~☘️🍀🌟 WeChat: PM/私信。https://t.co/6n9ktXQJtn
Ilyas Moutawwakil @IlysMoutawwakil
724 Followers 223 Following Machine Learning Optimisation & Benchmarking Engineer @HuggingFace 🤗 MEng @CentraleSupelec 🧑🎓
Allen Zhao @SomesomeZzz
6K Followers 79 Following Phd，专注AI相关板块的产业投资机会。 📈五年交易，十年经验。擅长产业逻辑，选股犀利；亏在手贱心痒，短期套利；追求极限机会，大亏小赚；最爱情绪交易，十赌九输！
Trump Truth Social Po... @TrumpTruthOnX
226K Followers 0 Following Independent. Not associated with realDonaldTrump. Trump Truth Posts On X along with breaking news stories and highlights.
Douglas A. Boneparth @dougboneparth
297K Followers 598 Following President at Bone Fide Wealth • CNBC Advisor Council • Memelord • Posts ≠ Financial Advice • Pre-order Money Together: https://t.co/Vx9x26AcOw
Wall Street Memes @wallstmemes
327K Followers 851 Following The No.1 Meme Community, News & Casino, Powered by $WSM Token (Follow @wsmcasino)
Dr. Parik Patel, BA, ... @ParikPatelCFA
730K Followers 989 Following Aswath Damodaran 🙏🏾 Dhandho Investor Chapter 3:4 🐂 | God first, full employment second 😤 | Investor @SamosaCapital | Subscribe to my newsletter 👇🏾
Inverse Cramer @CramerTracker
357K Followers 118 Following Finance and market moving news, parody, & memes. Bitcoin. Personal views. Not affiliated w/ Jim Cramer. NFA, DYOR.
Markets & Mayhem 🤖 @Mayhem4Markets
245K Followers 397 Following My focus is on markets, economics and technology 🧐 Building next generation trading tools and systems 📊 20 years of experience in trading and investing 📈
Financelot @FinanceLancelot
255K Followers 430 Following If you feel like buying me a coffee, subscribe for $1 ⤴️ I'm only on Twitter X. Any accounts on other social media are not me.
litquidity @litcapital
409K Followers 997 Following founder @litquidityvc // newsletters @exec_sum & @cryptosum_ // scout for @baincapvc // trade any market: https://t.co/DrAoLZjpN1
John W. Rich (Wealthy... @Cokedupoptions
219K Followers 2K Following Chairman of the Board at Yahoo! and AOL | Wealthiest man in the world on a divorce-adjusted basis | Gary, Indiana’s Favorite Son | PARODY
Ramp Capital @RampCapitalLLC
375K Followers 760 Following Master shape rotator, tweeting the things you think about but are too scared to tweet (+0m followers on TikTok, never been acquired, not a parody)
TTI @TikTokInvestors
293K Followers 1K Following We curate funny, interesting, and cringy finance / biz content. We are not affiliated with TikTok. Posts belong to respective owners.
Not Jerome Powell @alifarhat79
669K Followers 767 Following Memes. Not associated with the Federal Reserve. Financial Parody and sarcasm. Not investment advice. Crypto
赤沢りょうせい @ryosei_akazawa
47K Followers 2K Following 衆議院議員（鳥取２区、自民党）”赤沢りょうせい”。国土交通大臣政務官、自民党国土交通部会長、衆議院環境委員長、３度の内閣府副大臣、自由民主党政務調査会長代理、財務副大臣等歴任。現在、経済再生担当大臣（経済財政、防災庁設置準備、賃金向上など担当）。防災・国土強靭化、障がい者支援、女性活躍や少子化対策などがライフワーク。
Fox News @FoxNews
27.6M Followers 274 Following Stay Ahead. Get Breaking News Here First. Download the App.
Puru Saxena @saxena_puru
382K Followers 135 Following Founder of research firm - AlphaTarget. Previously, Founder / Portfolio manager of money management firms in Hong Kong. No investment advice, do your own DD.
Wasteland Capital @ecommerceshares
120K Followers 394 Following Escaped the Vampire Squid, surviving in the wasteland. Investing in bottlecaps, US UK EU & global assets. Also, jokes. Liking is not endorsing. Do your own DD.
Christopher Bloomstra... @ChrisBloomstran
122K Followers 1K Following Nothing here is advice. Drop the phone and slowly back away from the Twitter. Read a 10-K. It’s a profession, not a business. Disclosure: https://t.co/Sv6Y5YsrBq
David Rosenberg @EconguyRosie
297K Followers 341 Following Founder and President of Rosenberg Research & Associates Inc. Retweets, Likes and Follows are not endorsements.
Mohnish Pabrai @MohnishPabrai
253K Followers 962 Following Pabrai Funds, The Dakshana Foundation, @wagonsfund; author of The Dhandho Investor and Mosaic. Views are my own.
Jeffrey Gundlach @TruthGundlach
355K Followers 6 Following Bills fan. Art fan. Truth fan. DoubleLine Founder.
Daniel S. Loeb @DanielSLoeb1
162K Followers 1K Following Investor in public and private equity and credit securities. 🇺🇸🇺🇸🇺🇸
High Yield Harry @HighyieldHarry
143K Followers 199 Following Levered Finance Memes & Satire. Instagram FinMeme Account. @Buyside_Hub Founder. I write @hyhnewsletter and @wallstrollup
Bob Elliott @BobEUnlimited
195K Followers 763 Following CIO at @UnlimitedFnds | PM of $HFGM & $HFND | Fmr IC @Bridgewater | Described as one of the few "sane" voices on #fintwit | Comments are not investment advice
ManusAI @ManusAI_HQ
204K Followers 26 Following Manus is the general AI agent that bridges minds and actions: it doesn't just think, it delivers results. Download our app: https://t.co/XSfjRhjdgo
Gowthami @gowthami_s
9K Followers 946 Following Multimodal research | Past - UMD, MetaAI, Amazon, IIT Madras | Rants, Memes my own.
Tom Goldstein @tomgoldsteincs
27K Followers 2K Following Professor at UMD. AI security & privacy, algorithmic bias, foundations of ML. Follow me for commentary on state-of-the-art AI.
Laura Leal-Taixe @lealtaixe
11K Followers 126 Following Senior Research Manager at @NVIDIA. Prev Professor at @TU_Muenchen. Computer Vision mostly. Views are my own.
Dhruv Batra @DhruvBatraDB
19K Followers 619 Following Co-founder & Chief Scientist @yutori_ai. Prev: Senior Director leading FAIR Embodied AI @MetaAI and Professor @GeorgiaTech.
Andrea Tagliasacchi �... @taiyasaki
15K Followers 218 Following Associate Professor @ SFU (Research Chair), Research Scientist @ Google DeepMind, Associate Professor (status only) @ UofT. Opinions are my own.
Frank Dellaert @fdellaert
14K Followers 1K Following Robotics & Computer Vision Professor at Georgia Tech, and part-time CAIO at Verdant Robotics. Before: stints at KUL, Skydio, Facebook B*8, Google AI.
Yuke Zhu @yukez
20K Followers 467 Following Associate Professor @UTCompSci | Director @NVIDIAAI Co-Leading GEAR | CS PhD @Stanford | Building generalist robot autonomy in the wild | Opinions are my own
Dima Damen @dimadamen
10K Followers 699 Following Professor of Computer Vision, @BristolUniEng. Senior Research Scientist @GoogleDeepMind - passionate about the temporal stream in our lives.
Xin Wang @xinw_ai
7K Followers 1K Following Research @Meta l Ex OpenAI, Microsoft Research | @Berkeley_EECS PhD
Min Choi @minchoi
319K Followers 1K Following AI Educator. 𝕏 about AI, solutions and interesting things. Showing how to leverage AI in practical ways for you and your business. Opinions are my own.
Cameron R. Wolfe, Ph.... @cwolferesearch
27K Followers 676 Following Research @Netflix • Writer @ Deep (Learning) Focus • PhD @optimalab1 • I make AI understandable
Xiaomeng Xu @XiaomengXu11
745 Followers 268 Following PhD student in robotics @Stanford | Prev @Tsinghua_UniTrends for United States
You might like
