Charlie Ruan @charlie_ruan
MSCS @CSDatCMU | prev @CornellCIS Joined August 2014-
Tweets41
-
Followers252
-
Following120
-
Likes88
#LLama3 on #SteamDeck ! All thanks to the #MLC_LLM library that enables you to effortlessly compile #llms for Cuda, ROC and Vulcan and since Steam Deck is using AMD GPU it wasn’t a problem at all. I managed to run Llama-3-8B-Instruct-q4f16 with almost 15 tokens/s! #localllms
Deploy #Llama3 on $100 Orange Pi with GPU acceleration through MLC LLM. Try it out on your Orange Pi 👉 blog.mlc.ai/2023/08/09/GPU…
#Llama3 🦙🦙 running fully locally on iPad without internet connnection. credits to @ruihanglai and the team
Deploy #Llama3 locally with native GPU acceleration on CUDA/ROCm/Vulkan/Metal with MLC LLM. Check out llm.mlc.ai/docs/ for quick start instructions.
Llama 3 from @AIatMeta is now up on WebLLM! Try it on webllm.mlc.ai with local inference accelerated by @WebGPU. Or start building your local agent with the web-llm package -- everything in-browser!
More updates from WebLLM coming soon!
The #MediaPipe / Core ML team at Google launched a new #LLM API for Google's Gemma 2.5B param model today capable of running in browser with #WebAI! Pretty cool, so I whipped up a little web demo that goes beyond just talking to it like a chat bot...
Besides Android, iOS, and web browsers, Gemma is also supported in MLC LLM on various GPUs! A single model definition does it all -- thanks to the ML compiler infra lead by @junrushao and many others! Try it in Google Colab: github.com/mlc-ai/noteboo…
Google's Gemma model is now supported on Android using MLC LLM. Here is a demo of 4-bit quantized Gemma-2b model running on Samsung S23. Thanks to @ruihanglai and many others for bringing Gemma support to MLC! github.com/mlc-ai/mlc-llm llm.mlc.ai/docs/deploy/an…
Run Gemma model locally on iPhone - we get blazing fast 20 tok/s for 2B model. This shows amazing potential ahead for Gemma fine-tunes on phones, made possible by the new MLC SLM compilation flow by @junrushao from @OctoAICloud and many other contributors. github.com/mlc-ai/mlc-llm
Fist stab of Gemma support lands in MLC, github.com/mlc-ai/mlc-llm… (by @ruihanglai @charlie_ruan, rick zhou). The single model is defined once, and runs across browser/mobile/server. Made possible by ML compiler infra lead by @junrushao(and @leshenj15). more fun things to drop
🕸 #WebML / #WebAI Monthly #17: Test client side AI models in Headless Chrome, Stable Diffusion in <1s, +Chrome mobile now supports WebGPU: run LLMs on a phone. Have a read, and invite your team / friend to sub. Tag me in your Web AI finds for next edition linkedin.com/pulse/web-ml-m…
And two AMD RX 7900s with ROCm as well -- identical deployment process to 2x RTX 4090!
And two AMD RX 7900s with ROCm as well -- identical deployment process to 2x RTX 4090! https://t.co/NvNAqtAQ9U
Thats just wild! I can now run the #Mistral7B-Instruct, a 7 BILLION PARAM #LLM, directly in my MOBILE browser on my GPU thanks to #WebGPU. #WebAI #WebLLM Link in Thread 👇🏼
Run locally. Only way to keep your privacy. Among other advantages.
Run locally. Only way to keep your privacy. Among other advantages.
GustavEiffel @GustavEffel
2 Followers 606 Following⿻ barton 🦺𑗊 @bmorphism
2K Followers 4K Following applied categorical duck cyberneticist • building for agencies in the 21st century • inventor of the operadic cognitive diagram cognitive continuation standardShaista Shabbir @ShaistaSha26148
5 Followers 48 FollowingNassboko @NASSBOKO
56 Followers 378 Following 20 y développer web full Stack / intelligence artificielle 🧑💻🤖Oscar Tong 🌳 @0scartong
197 Followers 67 Following 👨🏻💻 Yet another web developer. 🇸🇬 Dweller of Singapore. 👨🏻 Father of two sons. 🙅 An atheist.Yuchen Jin @Yuchenj_UW
239 Followers 214 Following Co-founder & CTO @hyperbolic_labs 🧑🍳 fun AI systems. Previously @octoaicloud building @ApacheTVM, PhD @uwcse. 🤖Baher Hesham Azzam @AzzamBaher
0 Followers 15 FollowingCharlie Cheng-Jie Ji @charlie_jcj02
78 Followers 515 Following Gorilla LLM, CS & DS @ UC Berkeley, Data 100 Lead TA, Working towards LLM Tool Use, AI safetyEva Louise Marie Gabr.. @e681554349
9 Followers 3K FollowingEric Florenzano @ericflo
7K Followers 3K Following Making AR! Used to make VR. Before that made mobile apps. Before that made web things. People first, then products, systems, and data. he/himChord ai @chordaiapp
597 Followers 1K Following iOS / Android app to get chords, voicing & beats for any song 🎸🎹📱 Developed by former PhD students in ML and music enthusiasts. Tweets mostly by @VivienSeguytc Lin @stuser_tc
1 Followers 289 FollowingTessa Long @zhixuan_long
0 Followers 166 FollowingGe Zhang @GeZhang86038849
744 Followers 448 Following Founder: M-A-P(https://t.co/CGWz8Jr9K9) Incoming Ph.D. student: Computer Science @UWaterloo MSc: ECE & DS @UMich BSc: Computer Science @ BUPTredaready @redaready
317 Followers 3K FollowingPratyay Banerjee (ন.. @Neilzblaze007
242 Followers 4K Following I live in the shadows, but I watch everything.Jan @jantxuuu
31 Followers 314 Following solopreneur, ai, web, author of https://t.co/nqZBQfJ4hF newsletter - https://t.co/C43TeTR79e - https://t.co/9K3k09v6ZU - https://t.co/gFyaBpjGGItlu06 @tlu06
231 Followers 5K Following Shame isn't a strong enough emotion to stop us from doing anything at all.TheTendyMan @TheTendyMan
2K Followers 1K Following Researching $ALGO and Algorand related accessories. E-commerce & marketing 🧙. Born with a bad case of parrhesia. Apologies in advance.Hugo Zanini @hugoznn
174 Followers 281 Following Electrical Engineering @UFMG | @GoogleDevExpert in Machine Learning | Tech Product Manager @nubankAshutosh Mehra @ashutoshmehra
1K Followers 5K Following Senior Principal Scientist at Adobe. Working on Acrobat AI Assistant, LLMs, and document ML.Shawn Simister @narphorium
2K Followers 2K Following Working on GitHub Copilot @GitHub. Previously @Google. ML, knowledge graphs and tools for thought 🇨🇦 @[email protected]Deyan Ginev @dginev
1K Followers 390 Following Researching NLP for Math-rich documents. arXiv syntax spelunking. LaTeXML developer. creator of https://t.co/3d4VTPw7LjShawn Yin @shawn_xyin
111 Followers 837 Following AI4science, autonomous scientific discovery, clean energy and sustainabilityDominik Moritz @domoritz
4K Followers 1K Following HTML decoder. Prof @cmuhcii @cmudig and researcher @apple. Interactive vis tools (e.g. @vega_vis). PhD @uwcse @uwdata. Also at https://t.co/LRqaJK44lXDeping Zhang @joebradly
94 Followers 3K FollowingMrinal Deo @mdeo_deo
79 Followers 2K Following Computer engineer in his 40s. Loves computer architecture and rendering.HiMU @imhimu_
195 Followers 2K Following Machine Learning and Language Model R&D • “Useful AI” Enthusiast • Views Expressed Are Solely My Own • 🍁 🇨🇦/ @2Uj6c
0 Followers 5K Followingpotato_salad.cpp @potato_y_salad
410 Followers 3K Following anon-founder/research scientist - the only politics that remotely matter are accs vs. decels. grappling enthusiast. https://t.co/MIJmpB9iRWMengshiun @mengshyu
18 Followers 25 FollowingAbi Raja @_abi_
2K Followers 1K Following working on @picoapps and screenshot-to-code. founder @ peercdn (acquired). ex-palantir, stanford.Yixin Dong @yi_xin_dong
84 Followers 255 Following Undergrad in @sjtu1896, now visiting @uwcse, @ai_sampl. Working on @ApacheTVM, Compiler/MLSys/Distributed Sys. All views are my ownVinh Nguyen @vinhnx
860 Followers 5K Following Learning gen AI/LLMs. Building @vtdotai. Built @ClendarApp. Cars, iOS, Swift, SwiftUI, Photography. My opinions. "Learn by doing"Matthias 🦊 @Matt_heyqq
3K Followers 893 Following https://t.co/ZnAgcFT82V 🦊 | https://t.co/f6ygACnHCT | https://t.co/Uc3sDfaAge w/ @Dima_heyqq 🚀 | LL.M. ⚖️ & BA (film) 🎬 | Former filmdirector | Self-taught Developer (he/him)大良造 @wolfofyunmeng
43 Followers 1K FollowingYuchen Jin @Yuchenj_UW
239 Followers 214 Following Co-founder & CTO @hyperbolic_labs 🧑🍳 fun AI systems. Previously @octoaicloud building @ApacheTVM, PhD @uwcse. 🤖Charlie Cheng-Jie Ji @charlie_jcj02
78 Followers 515 Following Gorilla LLM, CS & DS @ UC Berkeley, Data 100 Lead TA, Working towards LLM Tool Use, AI safetyGe Zhang @GeZhang86038849
744 Followers 448 Following Founder: M-A-P(https://t.co/CGWz8Jr9K9) Incoming Ph.D. student: Computer Science @UWaterloo MSc: ECE & DS @UMich BSc: Computer Science @ BUPTShawn Simister @narphorium
2K Followers 2K Following Working on GitHub Copilot @GitHub. Previously @Google. ML, knowledge graphs and tools for thought 🇨🇦 @[email protected]Ashutosh Mehra @ashutoshmehra
1K Followers 5K Following Senior Principal Scientist at Adobe. Working on Acrobat AI Assistant, LLMs, and document ML.Dominik Moritz @domoritz
4K Followers 1K Following HTML decoder. Prof @cmuhcii @cmudig and researcher @apple. Interactive vis tools (e.g. @vega_vis). PhD @uwcse @uwdata. Also at https://t.co/LRqaJK44lXYixin Dong @yi_xin_dong
84 Followers 255 Following Undergrad in @sjtu1896, now visiting @uwcse, @ai_sampl. Working on @ApacheTVM, Compiler/MLSys/Distributed Sys. All views are my ownChuang Gan @gan_chuang
4K Followers 458 Following Faculty Member at UMass Amherst; Principal researcher at MIT-IBM Watson AI Lab; Homepage: https://t.co/oXP6pqXCpoPɾҽɱ Kυɱαɾ Aρ.. @prem_k
4K Followers 2K Following Head of IT Ops - Automation @Cognizant; Exec committee member @imcmontessori Karnataka chapter; Member @kans_india; Own views; QT≠reply. @[email protected]Jiawei Liu @JiaweiLiu_
2K Followers 957 Following Simplifying the making of great software. PhD Student @plfmse @IllinoisCS.Tlopex @Tlopex3
10 Followers 28 FollowingYichuanWang @YichuanM
180 Followers 940 Following 2020 acm class in SJTU interested in mlsys. Prev, intern at NYU with jinyang li. Supervised by Quan Chen in SJTUChris Donahue @chrisdonahuey
5K Followers 1K Following Generative models, musical expression for all. Assistant professor at CMU CSD. Part time research at Google Magenta (views my own)Google DeepMind @GoogleDeepMind
944K Followers 275 Following We’re a team of scientists, engineers, ethicists and more, committed to solving intelligence, to advance science and benefit humanity.Cornell University @Cornell
403K Followers 641 Following Learning. Discovery. Engagement. Join the #Cornell conversation.efxmarty @efxmarty
344 Followers 134 Following ML Engineer at @huggingface Optimization team. efxmarty/fxmarty elsewhereZihao Ye @ye_combinator
830 Followers 454 FollowingShiyi Cao @shiyi_c98
402 Followers 362 Following PhD student @UCBerkeley, MSc @ETH, B.S @sjtu1896, systems, ml, and hpcDacheng Li @DachengLi177
622 Followers 478 Following Intelligence. PhD @Berkeley_EECS @lmsysorg @ucbrise @berkeley_ai, Prev. @Google @SCSatCMU.Ying Sheng @ying11231
4K Followers 487 Following PhD student @Stanford. Large Language Models and Programs. | Do it anywayRobert Scoble @Scobleizer
504K Followers 68K Following Follow me on my new podcast with AI startups, Unaligned. Tech industry color commentator since 1993. Author/Blogger. Former strategist @Microsoft.BrendanEich @BrendanEich
182K Followers 2K Following Co-founder & CEO @Brave Software (https://t.co/NV4bmd6vxq) and @attentiontoken (https://t.co/XhGIrdBJWu). Co-founded Mozilla & Firefox. Created JavaScript.Chrome @googlechrome
7.3M Followers 105 Following The official Twitter account for the Google Chrome browser.Jay Wang @Jay4w
889 Followers 521 Following ML PhD student @PoloDataClub @GeorgiaTech 🐝 | Researching human-centered AI 🔍 | Prev. intern @GoogleAI @Apple @MSFTResearch 🌱 | @UWMadison '19 🦡Guillaume Lample @GuillaumeLample
37K Followers 648 Following Cofounder & Chief Scientist https://t.co/hLfvKLkFHd (@MistralAI). Working on LLMs. Ex @MetaAI | PhD @Sorbonne_Univ_ | MSc @CarnegieMellon | X11 @PolytechniqueMistral AI @MistralAI
91K Followers 0 Following Fast, open-source and secure language models. Join us https://t.co/INALdNGvCPAlex Atallah @xanderatallah
48K Followers 3K Following Experiments: @openrouterai + @windowai_io. Previously: co-founder & CTO @OpenSea.Xenova @xenovacom
6K Followers 284 Following Bringing the power of machine learning to the web. Currently working on Transformers.js (@huggingface 🤗)Jason Mayes @jason_mayes
9K Followers 1K Following #WebAI Lead @Google (#DevRel) (13yr Googler). I talk about: #WebML, #JS, #Chrome, #WebDev, #MediaPipe web, #Google #AI, #tech, #innovation. Opinions are my own.🤷 Nico Martin @nic_o_martin
2K Followers 1K Following Proud author of at least one todo app, several untitled projects and countless hello-worlds // owner of 0 NFTs // @GoogleDevExpert for #WebTechnologies 💚 #PWAAbubakar Abid @abidlabs
12K Followers 1K Following Hind Rajab. 5 yrs old. She + 14,000 children killed by Israeli forces. PLEASE don't be silent. Take 5 min to call your reps and urge peace (link in bio)Peiyuan Liao 🐇 @LiaoPeiyuan
2K Followers 212 Following CTO @rabbit_hmi, aka. Alexander; mono no aware ('the pathos of things')Jesse Lyu @jessechenglyu
29K Followers 290 Following founder and ceo @rabbit_hmi board @jugendingenieur any crypto relates to @rabbit_hmi or r1 is a scam.rabbit inc. @rabbit_hmi
84K Followers 1 Following rabbit brings the future of human-machine interface. order r1, your pocket companion, now.Tohid @tohid
77 Followers 2K FollowingSimon Guo 🦝 @simonguozirui
1K Followers 4K Following Incoming CS PhD student @Stanford and curr training models at @cohere | 🎓 @Berkeley_EECS | prev built things at @ @anyscalecompute @nvidiaMachine Learning Dept.. @mldcmu
17K Followers 370 Following The top education and research institution in the 🌎 for #AI and #machinelearning | Research → https://t.co/jUD0hZ8SFx | Learn more ↓#LLama3 on #SteamDeck ! All thanks to the #MLC_LLM library that enables you to effortlessly compile #llms for Cuda, ROC and Vulcan and since Steam Deck is using AMD GPU it wasn’t a problem at all. I managed to run Llama-3-8B-Instruct-q4f16 with almost 15 tokens/s! #localllms
🚨 Web Dev / #WebAI folk! 🚨 #Chrome wants feedback on new idea for translation API running on device (downloads AI model locally with new API features). We need your feedback! Looks cool - can run offline +private Docs: github.com/explainers-by-… Proposal: github.com/w3ctag/design-…
Meta's Segment Anything Model (SAM) can now run in your browser w/ WebGPU (+ fp16), meaning up to 8x faster image encoding (10s → 1.25s)! 🤯⚡️ Video is not sped up! Everything runs 100% locally thanks to 🤗 Transformers.js and onnxruntime-web! 🔗 Demo: hf.co/spaces/Xenova/…
Run #Llama3 locally on your android device using MLC-LLM! Download the mlc-llm APK from github.com/mlc-ai/binary-…. For more details, check out llm.mlc.ai/docs/
Deploy #Llama3 on $100 Orange Pi with GPU acceleration through MLC LLM. Try it out on your Orange Pi 👉 blog.mlc.ai/2023/08/09/GPU…
#Llama3 🦙🦙 running fully locally on iPad without internet connnection. credits to @ruihanglai and the team
Deploy #Llama3 locally with native GPU acceleration on CUDA/ROCm/Vulkan/Metal with MLC LLM. Check out llm.mlc.ai/docs/ for quick start instructions.
Want to try Llama 3 with @LangChainAI.js 🦜🔗? The 8b model is available on @Cloudflare WorkersAI! Or, if you're GPU rich, pull and run it locally with @ollama! 🌤️ Cloudflare docs: js.langchain.com/docs/integrati… Blog: blog.cloudflare.com/meta-llama-3-a… 🦙 Ollama docs: js.langchain.com/docs/integrati…
@charlie_ruan That would certainly be convenient for folk making things on Glitch / Codepen etc ✅
Wow. #WebAI folk are fast! #Llama3 8B model already ported to run entirely in the web browser, client side, on device. Capture below is from my NVIDIA 1070 GPU (quite old now) + still runs fast enough! Incredible work by MLC folk (hat tip to @charlie_ruan) webllm.mlc.ai
Llama 3 running natively in browser accelerated by @WebGPU
Llama 3 from @AIatMeta is now up on WebLLM! Try it on webllm.mlc.ai with local inference accelerated by @WebGPU. Or start building your local agent with the web-llm package -- everything in-browser!
While there is a lot of enterprise funding into #LAM, looks like the travails of #RPA cannot be avoided. We need #LLMs to call APIs like how TaskMatrix.ai attempts to do. And made to run in-browser, leveraging #WebLLM & js version of @LangChainAI (by @Hacubu).
So many groups in industry working on web agents/action models/benchmarks now! @GoogleDeepMind -> Pix2Act & WebGUM @ServiceNowRSRCH -> WorkArena + Case Study @MithrilSecurity -> LaVague Salesforce Research -> AgentOhana
🌟 My fully-local chat with PDF repo doubled in stars this week - now at more than 1.2k! So I revisited WebLLM and was able to add browser-only mode! Here, the entire @LangChainAI ingest + RAG flow runs with the model weights cached in browser storage! webml-demo.vercel.app
Model is @Microsoft Phi-2 via WebLLM. @charlie_ruan has done incredible work on webllm.mlc.ai in the ~9 months! Vectorstore is still local @dawchihliou Voy Embeddings run locally via @xenovacom Transformers.js. Check out the full repo here: github.com/jacoblee93/ful…
WGSL (WebGPU Shading Language) absolutely steals the show in Chrome 123. ➡️ DP4a built-in functions support ➡️ Unrestricted pointer parameters ➡️ Syntax sugar for dereferencing composites Check out developer.chrome.com/blog/new-in-we… now!
We just added some new features to the 🤗 Transformers.js WebGPU Embedding Benchmark: - fp16 and int8 support - Ability to change models - Lazy model loading - Options to select which tests to run On my device, I got >100x speedup with fp16 on WebGPU! ⚡️ huggingface.co/spaces/Xenova/…
Introducing the 🤗 Transformers.js WebGPU Embedding Benchmark! ⚡️ How much does WebGPU speed up ML models running locally in your browser? Try it out and share your results! 🚀👇 🔗 huggingface.co/spaces/Xenova/…
Introducing the 🤗 Transformers.js WebGPU Embedding Benchmark! ⚡️ How much does WebGPU speed up ML models running locally in your browser? Try it out and share your results! 🚀👇 🔗 huggingface.co/spaces/Xenova/…
The #MediaPipe / Core ML team at Google launched a new #LLM API for Google's Gemma 2.5B param model today capable of running in browser with #WebAI! Pretty cool, so I whipped up a little web demo that goes beyond just talking to it like a chat bot...