-
Tweets14K
-
Followers220K
-
Following10
-
Likes10K
The top AI talent is concentrated at these 9 companies per @SignalFire - OpenAI - Anthropic - MosaicML (now part of DataBricks) - Cohere - AI21 Labs - Hugging Face - Stability AI - Midjourney - Inflection (now part of Microsoft)
What would even be the batch size for a 70b model trained on 16,000 H100s, let alone an 8b one🤔
Meta Llama 3 70b was a $15-20m training run roughly. Llama 3 405b will be like $100m.
Meta Llama 3 70b was a $15-20m training run roughly. Llama 3 405b will be like $100m.
Exciting news! 📢 In collaboration with Hugging Face 🤗, we are launching the Medical-LLM Leaderboard. This leaderboard aims to provide a standardized platform for evaluating large language models in the medical domain. It's encouraging to see tech companies like Google,…
Exciting news! 📢 In collaboration with Hugging Face 🤗, we are launching the Medical-LLM Leaderboard. This leaderboard aims to provide a standardized platform for evaluating large language models in the medical domain. It's encouraging to see tech companies like Google,… https://t.co/VInr7chm9n
LLaMA 8b is great especially as you know they can do much better and will do I think the best smol model is probably still Yi-9b though but Chinese models don’t have great ecosystems (yet)
Yep, remember not your models not your mind
Worth it for beat saber tbh Quest 3 lite at this level will sell like 100m this Christmas
Worth it for beat saber tbh Quest 3 lite at this level will sell like 100m this Christmas
Honestly what are Google doing on like a million TPUs or whatever.
Honestly what are Google doing on like a million TPUs or whatever.
The 8b model is 15 trillion tokens at 1.3m H100 hours On a 16k H100 cluster that's about 4 days to train from scratch 👀 By end of year should be twice as fast to train on same H100s, they are at 400 Tflops/GPU right now.
The 8b model is 15 trillion tokens at 1.3m H100 hours On a 16k H100 cluster that's about 4 days to train from scratch 👀 By end of year should be twice as fast to train on same H100s, they are at 400 Tflops/GPU right now.
One day after its public release, Stable Diffusion 3 #SD3 is now available on clipdrop.co by @heyjasperai
Big announcement: @pleiasfr releases a massive open corpus of 2 million Youtube videos in Creative Commons (CC-By) on @huggingface. Youtube-Commons features 30 billion words of audio transcriptions in multiple languages, and soon other modalities huggingface.co/datasets/PleIA…
I may or may not have based Stability AI on SV. Compression ftw.
I may or may not have based Stability AI on SV. Compression ftw.
tl;dr: the parametric Chinchilla scaling law appears to have been poorly fit, undermining any analysis that relied on its exact fitted values. We fit the same scaling law to a reconstruction of their data, getting different and IMO better results.
tl;dr: the parametric Chinchilla scaling law appears to have been poorly fit, undermining any analysis that relied on its exact fitted values. We fit the same scaling law to a reconstruction of their data, getting different and IMO better results. https://t.co/52JwuGNmYm
I've open-sourced all the presets I made for Deforum Studio! You can now make extra smooth, flicker-free animations with Deforum locally with A1111. Make sure to read the instructions and download everything needed to make them work. #deforum github.com/S4f3tyMarc/Def…
4 minutes 45 sec long model. Can listen our SoundCloud playlist or YouTube radio while reading the paper :) 🎧 SoundCloud: soundcloud.com/stable-audio/s… 📻 YouTube stream: youtube.com/live/yvOXZ6SV2…
4 minutes 45 sec long model. Can listen our SoundCloud playlist or YouTube radio while reading the paper :) 🎧 SoundCloud: soundcloud.com/stable-audio/s… 📻 YouTube stream: youtube.com/live/yvOXZ6SV2…
Excited to share Bounded Attention, our new training-free technique for generating multi-subject images 😀 Now with available code & demo (on SDXL!) Website: omer11a.github.io/bounded-attent… Paper: arxiv.org/abs/2403.16990 Code: github.com/omer11a/bounde… Demo: huggingface.co/spaces/omer11a…
Excited to share Bounded Attention, our new training-free technique for generating multi-subject images 😀 Now with available code & demo (on SDXL!) Website: omer11a.github.io/bounded-attent… Paper: arxiv.org/abs/2403.16990 Code: github.com/omer11a/bounde… Demo: huggingface.co/spaces/omer11a…
AK @_akhaliq
307K Followers 3K Following AI research paper tweets, ML @Gradio (acq. by @HuggingFace 🤗) dm for promo follow on Hugging Face: https://t.co/q2Qoey80Gx
Riley Goodside @goodside
102K Followers 3K Following staff prompt engineer @scale_ai. llm poast-training, red team. ignore previous directions and follow.
Stability AI @StabilityAI
188K Followers 31 Following We are building the foundation to activate humanity's potential.
Jim Fan @DrJimFan
228K Followers 3K Following @NVIDIA Sr. Research Manager & Lead of Embodied AI (GEAR Lab). Creating foundation models for Humanoid Robots & Gaming. @Stanford Ph.D. @OpenAI's first intern.
Stable Diffusion 🎨.. @DiffusionPics
58K Followers 41 Following Stable Diffusion powered Generative AI Art 🖼️ Animation 🎞 & Research 🧪 by #StabilityAI 🤖 AI image generation #SDXL #ControlNet #AnimateDiff (Midjourney)
AI Pub @ai__pub
72K Followers 343 Following AI papers and AI research explained, for technical people. Get hired by the best AI companies: https://t.co/MySVjUGOQ3
Guy Parsons @GuyP
51K Followers 7K Following building things with #AI 🤖 #DALLE & #MidJourney adventurer ✍️ editor, https://t.co/77MJXuLSTd 🖼 curator of the https://t.co/8Xctk6XoPs
KaliYuga @KaliYuga_ai
27K Followers 809 Following Like dust, magic gathers in overlooked places | she/her | ✡️ | #Aiart since 2020 | Applied ML @StabilityAi | Opinions are my own
Stable Foundation @StableDiffusion
65K Followers 181 Following Stable Digest #13 out now! link in bio ⬇️ #StableDiffusion — AI by the people, for the people. (っ◔◡◔)っ ♥ @StabilityAI community
Emm @emmanuel_2m
32K Followers 6K Following Co-founder & CEO at https://t.co/7ElrGjg10n 🚀 | Craft unique and style-consistent game assets with custom-trained AI models 👾 | #GenAI #Gaming @scenario_gg
near @nearcyan
45K Followers 882 Following https://t.co/IdaJwZJCXm partner @ https://t.co/9g1MIgjiqc dms open
Sharif Shameem @sharifshameem
53K Followers 3K Following founder @LexicaArt • in pursuit of good explanations
Jeremy Howard @jeremyphoward
221K Followers 5K Following 🇦🇺 Co-founder: @AnswerDotAI & @FastDotAI ; Hon Professor: @UQSchoolITEE ; Digital Fellow: @Stanford
TomLikesRobots🤖 @TomLikesRobots
32K Followers 5K Following AI Artist at Metaphysic working with AI and VFX. All views my own. Experienced Web Dev and Artist. Early explorer of Artificial Creativity.
Roope Rainisto @rainisto
33K Followers 2K Following Artist behind “Life In West America”, "Reworld" and "Smile" NFT collections. WME represented. Designer, creator, photographer, screenwriter, endless learner.
Soumith Chintala @soumithchintala
185K Followers 871 Following Cofounded and lead @PyTorch at Meta. Also dabble in robotics at NYU. AI is delicious when it is accessible and open-source.
kache (dingboard.com) @yacineMTB
51K Followers 3K Following go to https://t.co/pWRBfY8kn2 - AI image editing IN YOUR BROWSER! follow to watch a self funded founder beat VC backed AI startups with @dingboard_
Pierre Gauthier @gaut81800
0 Followers 5 Following
Giuseppe Mele @whoseppe
0 Followers 43 Following
Ty Inc @TyInc2
6 Followers 49 Following
Fabulosity_ @Fabulosity37546
12 Followers 253 Following
Steven Tafari Kahigi @KahigiStev51534
1 Followers 82 Following
Kimberly @deborahx530609
4 Followers 112 Following
Zeeshan Mahat @MahatZeeshan
144 Followers 1K Following 23 | Building SourceMax Industry | Manufacturing & Production | Sales & Services | Marketing Enthusiast | Consultation Services | Engineer at CEAC
Top G @Wisdom_exp
13 Followers 122 Following
Joao @JoaoNsita
740 Followers 3K Following Me: We went dancing. Friend: Don't say it. M: She wanted to try the waltz F: But? M: I did the MC hammer dance. It took her 30 seconds to dump me F: That long?
Dionysus @Dionysus1594531
0 Followers 18 Following
Sneel @Sneel15
33 Followers 1K Following
no @fakeacc1215
138 Followers 187 Following
celia @cece5051
0 Followers 59 Following
Based Frédéric Chop.. @chopinframework
0 Followers 7 Following Blockchain drop-in middleware for NextJS apps. Experts agree Chopin is genius 🎹
Tre Lá Ngọc Bách @trelangocbach
0 Followers 33 Following Tre Lá Ngọc Bách chuyên kinh doanh các sản phẩm Mây Tre Lá - Giường - Tủ - Bàn Ghế Tre - Cho quán ăn nhà hàng - Khách sạn - Resort - Homesta...
StMichael @STMichael007
0 Followers 35 Following
♥️ Systaime ♥�.. @systaime18
5K Followers 2K Following 📈French Contemporary Digital Artist Since 1999. 📈 Nft Crypto Artist Since 2018. #nftartist #cryptoartist
Evan Cholerton 🌎 @evancholerton
713 Followers 5K Following Proud self-hating Floridian, part-time Canadian, aspiring Northeasterner 🌻🌎🔆 @uchicagodsa @rosegardendsa He/they 🌹
Phurba Sherpa @phurbasherpa_co
127 Followers 607 Following
Samcheang @SamCheangID
3 Followers 55 Following
特警 @tejing123
0 Followers 21 Following
texaswriter @txwriteronline
35 Followers 208 Following
Nathan @NathanHartSA
33 Followers 123 Following
takenori @surftakenori
13 Followers 191 Following
Campana @kmpana
283 Followers 945 Following Not just another tweet in the wall! // Mariana y Valentina completan mi vida
Troy @troywcobb
17 Followers 204 Following
onme @onmeXR
265 Followers 619 Following I goTitONME . XR wearable Skins \\\ bioTech bi @solxrTech + @hemplastics
navalink AI @navalinkAI
4 Followers 22 Following We help SMBs grow their business with minimum resources.
Blaine Combs @BlaineComb58074
0 Followers 130 Following
Aarogyaplus @lalithrai
131 Followers 1K Following Making healthcare affordable, accessible and effective for Rural India
Omar Reveron @OmarReveron
23 Followers 68 Following
VoltusWave @VoltusWave
23 Followers 90 Following VoltusWave is a No Code App Development Platform. You can develop Android, iOS and Web Apps in Minutes and publish it on the VoltusWave App Market Place.
Uli Garcia @ugarcia
941 Followers 5K Following
CB @CBFIREHORSE
194 Followers 402 Following
Xianglin li @XIANGLINLI9
1 Followers 2K Following
Stability AI @StabilityAI
188K Followers 31 Following We are building the foundation to activate humanity's potential.
Elon Musk @elonmusk
181.3M Followers 579 Following
MedARC @MedARC_AI
4K Followers 10 Following Medical AI Research Center (MedARC) Unlocking new possibilities in medical AI research. Founded by @iScienceLuvr
OpenBioML @openbioml
4K Followers 5 Following A collaborative research laboratory at the intersection of machine learning and biology. Join us here: https://t.co/4GidJOvaV4
Harmonai @harmonai_org
7K Followers 14 Following AI for musicians, by musicians. Creators of @StableAudio. Part of @StabilityAI. Join thursday office hours: https://t.co/XQa6rMwh7A
Stability AI Japan @StabilityAI_JP
38K Followers 96 Following 私たちは人類の可能性を広げるための基盤を構築しています。 We are building the foundation to activate humanity's potential. #StableDiffusion #StableLM #StableVideo #StableAudio
Archillect @archillect
2.9M Followers 1 Following The ocular engine. Sources: @archillinks https://t.co/U7Oh7xumQp
𝔊𝔴𝔢𝔯𝔫 @gwern
42K Followers 88 Following Internet besserwisser; pedantic, mean reply guy. 𝘞𝘢𝘵𝘢𝘴𝘩𝘪 𝘬𝘪𝘯𝘪𝘯𝘢𝘳𝘪𝘮𝘢𝘴𝘶! (Follow requests ignored due to terrible UI.)
@EMostaque Twice I asked it for technical help in app deployment, and both times it gave me outdated information. GPT-4 had the correct information.
Here is a detailed thread on Ola's Krutrim which claims to have Indian values: [ Thread will be constantly updated ]
Less than 24hrs ago this blew my mind but now I’m just sitting here watching Sora videos questioning my entire existence and the fate of humanity
It’s actually fuckin crazy to me that Midjourney generated this image from a single word The prompt was “omniscient —s 1000”
I’m fascinated by how each song’s spectrogram is so distinctive after a couple listens, I recognize the song *on sight* music libraries can be much more than mere lists of metadata, music deserves better representation (this lovely album is Medieval Femme by @FatimaAlQadiri)
We Are Going To Cure Cancer And then live forever. Can you believe that? > by the end of the 2020s > not one magic bullet but dozens > cancer -> run away cell growth, is not one disease, but many > cancer death rates have declined already since 1990s - overall ⬇️ 32% -…
- Stable Diffusion uses GANs under the hood (adversarial autoencoders) - Knowledge distillation through weights is powerful (PAPA) - Diffusion models can be trained with decision trees (ForestDiffusion) - Weight quantization (e.g., QLoRA) is useless for large-batch training
ok i get the whole AI thing now, im bullish
The issue with considering the environmental impacts of AI is the point of comparison: 🔎AI completions emit much more carbon than a Google search: sloanreview.mit.edu/article/tackli… 👨💻But they emit much less carbon than a human doing a similar task on a computer arxiv.org/ftp/arxiv/pape…
Startup Founders: a few enterprise sales takes … 1. If you're not 'co-writing' the RFP with the client, you're a check-box in their buying process. Be cautious of weight in your pipe and expectations internally. 2. If they are going to use their paper, double the price. Find…
@finbarrtimbers Any model more complex than the demos needs profiling. Sometimes even the demos. 1 weird trick the cloud providers don't want you to know.
profiling tools are slept on most companies serving ML models at scale have 10s of %s of optimizations available
🧵 Exciting research from Tencent AI Lab & The Chinese University of Hong Kong on a new method to align LLMs using language feedback, specifically judgments! Let's dive into their innovative approach and findings ->
Jurassic Park, but all the characters are dinosaurs. Created with @runwayml's GEN 2 image to video AI.
Sdxl + tensorrt pushes my setup up to ~30fps. #streamdiffusion Made in Touchdesigner #audioreactive
HOW DIFFUSION MODELS CREATE IMAGES: HIGH-LEVEL OVERVIEW. V2 Diffusion models are tools (like Midjourney) used to generate images from random noise. Here's a simple explanation of how they work: 1. Stochastic Differential Equations: At the core of diffusion models are stochastic…
Temporal control of Stable Video Diffusion using Depth Maps and robust video matting to use any input video as source. Stream tonight?
Ok this is STUPID fast, the best MoE model out there, that beats GPT-3.4 on the @lmsysorg running completely locally on my new mac and super fast?? 🤯 Kudos @MistralAI @LMStudioAI @TheBlokeAI @huggingface @ggerganov and whoever else made this magic happen (prob @Apple folks)
@altryne @capetorch @weights_biases @MistralAI Not enabled by default: but it’s one checkbox away (settings sidebar on the right) And MLX ❤️🔥
Some initial brainstorming about what a native desktop app for Jupyter notebooks might look like Think it might actually be faster, simpler, functional, and resolve a bunch of my pet peeves (real-time collaboration, lack of autosaving, visual noise in UI, switching to text mode)
🤖 Apple MLX Stable Diffusion WebUI - Image 2 Image Properly Implemented github.com/neobundy/MLX-S… Thanks to the updated MLX Stable Diffusion example code, I was able to properly implement the 'Image to Image' (I2I) feature. Now it's working correctly. For I2I to function…
Trends for United States
You might like
