LAION (@laion_ai) header
LAION (@laion_ai) profile picture

100% TRULY OPEN AI.

The internet
Joined September 2021
Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)Photo by LAION (@laion_ai)
LAION retweeted
LAION-Audio-XL is number 1 trending audio dataset of HF! huggingface.co/datasets/laio…
1
4
1
16
2,429
Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me. I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA. I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation. Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO. Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours. Jensen
56,407
116,154
23,661
1,097,756
49,501,623
Christoph (@ChristophSchuh6) made a niffty audio segmentation demo using the CLAP2 models we released last week that he shared on his YouTube. He doesn't post much here these days so check it out: youtube.com/watch?v=yr59UH0G…
1
3
17
2,296
AI can write a song in seconds... But which ones would people prefer? TuneJury can help, and it just got new ears! 🎧 With CLAP2 from @laion_ai, it: • listens to up to 2 min • agrees with human listeners more often • scores higher on musicality benchmarks Thanks @foerster_ai and the whole CLAP2 team for the early Christmas gift to TuneJury! 🙏
[1/6] 🔊 We release CLAP2 and LAION-Audio-XL: 365M audio clips (1.39M hours) with 707M captions, used to train five CLAP2 models from scratch. 🏆CLAP2-L reaches 53.6% on MAEB-30, the best CLAP model on the leaderboard. arxiv.org/abs/2611.11592
2
4
32
5,288
We're actively working to remove fake entries and prevent trolls from abusing our page. Thanks for your patience. Also, this is our one and only X account. Please don't trust any other accounts claiming to represent this page. Thanks for your understanding and support! 💚 #standwithjensen #bringjensenback standwithjensen.com/
196
642
52
5,750
292,409
I released OpenCLIP 4.2 earlier this week to support the new CLAP2 variable length audio-text models. OpenCLIP 4.x brings a host of other changes aside from the new CLAP2 and CLAP audio-text support. I've added: • OpenMaMMUT (a dual contrastive + generative image-text family w/ open weights). I fixed some pooling bugs in a set of MaMMUT2 models, along w/ corresponding CoCa pooling fixes in CoCa2. • GenLIP / GenLAP models: pure generative decoder only models with a NaFlex pipeline feeding a simple input projection + norm layer for image / audio, text embed + norm for text modalities. • NaFlexViT encoder support for all of the models. CLAP2 is an audio encoder that's a ViT w/ log-mel patches. MaMMUT2 and CoCa2 support NaFlexViT too. • All models can now use more recent, 'ModernText' transformers w/ RoPE, SwiGLU, gated attention, variable / long context length, etc. ModernText encoders / decoders are typically paired with a TikTokenizer. • The data pipeline was enhanced to support NaFlex patchification (variable length, variable size patches) and variable text length handling, including length bucketing to recover some efficiency with varying image and text patch sequences. • Completely refactored training code w/ FSDP2 support, torch.compile, and shortly PyTorch 2.15 varlen attention will be supported for a nice NaFlex throughput boost.
🤖 Made with AI
6
8
46
2,707
[1/6] 🔊 We release CLAP2 and LAION-Audio-XL: 365M audio clips (1.39M hours) with 707M captions, used to train five CLAP2 models from scratch. 🏆CLAP2-L reaches 53.6% on MAEB-30, the best CLAP model on the leaderboard. arxiv.org/abs/2611.11592
5
29
3
149
14,597
Who did this shit lmao
345
2,075
367
13,865
1,004,243
18 of the 20 best open models today are Mixture-of-Experts. Dense is basically over. And yet the know-how to train a good MoE still lives mostly in the heads of a handful of teams. So we spent 800k GPU hours running the experiments ourselves and wrote everything down. Super happy to finally release "The Ultimate MoE Training Handbook", free and open, organized around the questions training teams hit: - how many experts, how big, shared or not? - why is half my step spent in all-to-all, and how do I hide it? - why is my run 25x slower with no error in the logs? - softmax, sigmoid or sqrt(softplus) router, and why keep it in fp32? - aux loss or DeepSeek's loss-free bias? - upcycle a dense checkpoint or train from scratch? - can I serve with fewer experts per token than I trained with? ⚠️ Warning: it's long (~70k words, budget a few days, sorry) but with 55 interactive figures, code you can run and numbers from our own runs. Read it here: huggingface.co/spaces/huggin…
140
624
82
4,562
277,404
We turned @OpenAI's math problems into open-source RL environments on @huggingface! Early and rough (the verifier only accepts the exact original formalization, so an equivalent proof can still score 0), but this is what it looks like when a research release becomes executable training infrastructure for everyone. This is an exciting direction imo: turning open research into open executable environments that anyone can build on to train better open models! huggingface.co/datasets/Fine…
122
169
31
1,383
72,726
▼概要 約56億件の公開TikTok動画データを収録した大規模データセット「tiktok-5.6B-videos」がHugging Faceで公開。2014年から2026年10月までのメタデータを月別に分割して提供。 ▼ポイント ・キャプション、ハッシュタグ、音声情報、TikTok Shopリンク、エンゲージメント統計を収録 ・2014年から2026年10月3日時点のスナップショットとして55億9746万2038件の動画情報を網羅 ・データ形式は月ごとのParquetファイルを採用 ・DuckDB等を用いて、必要なカラムのみを抽出する効率的なクエリ処理に対応 ・外部サイト(datasocial.ai)にて、クリエイター情報や日次更新データ、スクレイパーのコードなども提供 #TikTok #HuggingFace URLはリプ⬇️
9
186
21
2,209
215,314
LAION retweeted
SPARE, a short film generated with the full version of Kling 4.0. Finally here. Thanks to Kling AI for early access. Made with full Kling 4.0 and Omni Referenc - Native 30s output - Sharper sound and picture - Tighter lip-sync - 21:9 cinematic Kling 4.0 launches in October.
27
23
9
262
22,911
Palisade interviewed 22 current and former employees from OpenAI, DeepMind, and Anthropic about their personal views and fears around AI development. Today, we’re releasing the first batch of those interviews. Please watch and share.
46
255
67
1,298
274,407
For context, this is about 1.5x the amount of data used to train the original Whisper. You could combine this with YODASv2, VoxPopuli, MLS, People’s Speech v1 and v2 to train a model on roughly 3M hours of open data.
Who wants more speech training data? YODAS v3 is now available @huggingface It’s 1.1M hours - the biggest audio dataset ever. And the first at this scale with stereo audio at 48kHz, with timestamped transcripts and translations. 100+ langs hf.co/blog/espnet/yodasv3 CC-BY-3.0
2
3
1
67
7,007
We've had a lot of fun with Eleven v4. The level of realism keeps surprising us, and we think you'll have fun with it too. At the same time, it's incredibly practical, leading in pronunciation robustness and human preference.
Introducing Eleven v4 and Eleven v4 Turbo, our fastest and most emotive voice models yet. Ranked #1 by Artificial Analysis.
4
9
45
2,869
Who wants more speech training data? YODAS v3 is now available @huggingface It’s 1.1M hours - the biggest audio dataset ever. And the first at this scale with stereo audio at 48kHz, with timestamped transcripts and translations. 100+ langs hf.co/blog/espnet/yodasv3 CC-BY-3.0
23
73
13
457
49,645
LAION retweeted
Shit, Gary Marcus predicts that there is no way AI kills absolutely everyone. Given his track record, we are dead in two years
unavailable
40
39
4
889
39,233
Excited to share our new 𝗼𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲 work, 𝗛𝗔𝗥𝗠𝗢𝗡𝗬, a framework for reconstructing compositional 3D scenes from single indoor images. Much like a 3D artist, our VLM agent builds the scene in a 𝗵𝗶𝗲𝗿𝗮𝗿𝗰𝗵𝗶𝗰𝗮𝗹 𝗼𝗿𝗱𝗲𝗿: first the room structure and camera, then wall-mounted objects, free-standing furniture, and finally dependent decorations. After each stage, it compares its render against the input image, reflects on what went wrong, and iteratively corrects the scene to prevent errors from accumulating. Even with older VLM backbones, our framework remains 𝗰𝗼𝗺𝗽𝗲𝘁𝗶𝘁𝗶𝘃𝗲 𝘄𝗶𝘁𝗵 𝗚𝗣𝗧-𝟲 𝗔𝘀𝘁𝗿𝗮 𝗶𝗻 𝘃𝗶𝘀𝘂𝗮𝗹 𝗾𝘂𝗮𝗹𝗶𝘁𝘆, 𝘄𝗵𝗶𝗹𝗲 𝗮𝗰𝗵𝗶𝗲𝘃𝗶𝗻𝗴 𝗯𝗲𝘁𝘁𝗲𝗿 𝗴𝗲𝗼𝗺𝗲𝘁𝗿𝗶𝗰 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁 for single-image 3D scene reconstruction. We also release 𝗛𝗔𝗥𝗠𝗢𝗡𝗬𝟯𝟬𝟬, a benchmark for single-image 3D scene reconstruction with 300 scenes across three difficulty tiers: easy, medium, and complicated. 🔗 Project: cwchenwang.github.io/harmony 📊 Leaderboard: cwchenwang.github.io/harmony… 🤗 Data: huggingface.co/datasets/Shuf… 📄 Paper: arxiv.org/abs/2609.26793
6
42
3
306
30,676
Me, in a time machine, to Eliezer in 2006: The year is 2026. You stand back to back with Bernie Sanders. Peter Thiel has named you the Antichrist. EY-2006: Oh no. Did somebody invent pharmacological mind control and use it on me, or- Me: Nvidia was on the verge of destroying all things. You had no choice. EY-2006: Are we talking about the computer graphics card manufacturer, or an unrelated supervillian named "Invidia"? Me: The former. It turned out that computer graphics cards contained a surprising amount of world-destroying potential. All attempts at hindering the reckless exploitation of the Graphic Card Force for corporate profit were stymied by the far left, which feared that any attempts to regulate them might lead to regulatory capture. Other attempts were made to prevent Nvidia from selling to foreign companies that resold to communist China, but those attempts were blocked by the far right. Nvidia is now a $5.5 trillion company. EY-2006: ...What are politics like in 2026, exactly? Me: In other news that is mostly unrelated, the decision theory paper you're currently working on will accidentally spark off a transgender vegan murder cult. EY-2006: A WHAT? Why? How? Why? Me: Millions know your name as the greatest of heroes, millions more as the greatest of villains, and other millions know you solely as the greatest author of Harry Potter fanfiction. EY-2006: This is beginning to strain credulity. Me: The New York Post will probably soon publish a story claiming that you keep a harem of submissive mathematicians. Sadly, they will be lying. EY-2006: I'm going to stop believing you now. Me: All of this is taking place under the ominous shadow of Donald Trump.
102
173
52
3,266
156,478