Nayan Saxena
@SaxenaNayan
Followers
3,018
Following
509
Media
59
Statuses
313
Teaching machines how to think, express, and create – one model at a time.
Toronto, Ontario
Joined October 2021
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
GLOBAL SPOTLIGHT APO NATTAWIN
• 229487 Tweets
Morara
• 94613 Tweets
天使の日
• 73279 Tweets
علي النبي
• 71342 Tweets
Bomas
• 49947 Tweets
Hayırlı Cumalar
• 42784 Tweets
SB19 IS ACER READY
• 42090 Tweets
Kasmuel
• 33172 Tweets
Michel Blanc
• 26886 Tweets
#يوم_الجمعه
• 26275 Tweets
#リアルアキバボーイズ武道館
• 22676 Tweets
感謝マルチガチャ
• 21955 Tweets
ワートリ
• 21480 Tweets
Tariq
• 19818 Tweets
名探偵ピカチュウ
• 19329 Tweets
最大4体ゲット
• 17591 Tweets
人達同士
• 17337 Tweets
TJ Kajwang
• 15259 Tweets
BINI LAGING FUDGEE
• 12668 Tweets
所信表明演説
• 12209 Tweets
Last Seen Profiles
Pinned Tweet
Amazing article on our recent work by
@kevinroose
from The New York Times (
@nytimes
), this quote captures the essence well:
"Eventually, if you take advantage of the web, the web will start shutting its doors."
1
4
12
Awesome new project, Mirage, a superoptimizer that discovers highly optimized GPU kernels for LLMs and more. Its 𝜇Graphs found speedups 2x faster than FlashAttention and 1.5x faster than FlashInfer. For LoRA, it merges three matmuls into one GPU kernel, outpacing Triton by 3x.
3
16
92
Finally got a real chair after 5 years. Sitting on this feels unbelievable
🥳🥲🥹
10
2
76
✨ Introducing ToDo : Token Downsampling for Efficient Generation of High-Resolution Images ! With
@Ethan_smith_20
&
@aningineer
, we present a training free method that accelerates diffusion inference upto 4.5x while maintaining image fidelity.
ToDo
Token Downsampling for Efficient Generation of High-Resolution Images
Attention mechanism has been crucial for image diffusion models, however, their quadratic computational complexity limits the sizes of images we can process within reasonable time and memory constraints.
3
31
155
1
9
72
🌟 Excited to finally share that
@LeonardoAi
has been acquired by
@Canva
! Our work on Phoenix, the best image generation foundation model they've seen, played a key role. Special kudos to
@Ethan_smith_20
,
@advadnoun
,
@aningineer
,
@sami_ede
, Martin Bell, and the entire team!
🚀 Exciting news: Welcome,
@LeonardoAi_
! 🚀
Since 2013, Canva's mission has been to empower everyone to design. Today, we're thrilled to join forces with Leonardo AI, a leader in generative AI. 🎉
Leonardo’s team and tech will boost our AI capabilities, enhancing our products
120
106
441
6
11
68
✨ Excited to announce Phoenix, a new foundational image generation model! Over the past few months, I've worked with the amazing
@LeonardoAi_
research team, training and bringing it to life. Huge kudos to
@Ethan_smith_20
,
@advadnoun
,
@aningineer
,
@sami_ede
& the entire team!
We're incredibly excited to announce Phoenix, our own foundational model now in preview for ALL users.
👉 Prompt adherence ^ n
Experience exceptional prompt adherence - and it's only getting better from here.
👉 Coherent text in image
Generates clear, accurate text within
59
101
551
2
11
62
The stuff I get to work on as part of my job sometimes feels unreal ! Working on AI art is really satisfying and delving into the technical details has been quite insightful ✨
Here is a preview of what is next to come within Dream by
@WOMBO
🌈
5
4
59
Canada now has the most forward-looking national compute and AI adoption strategy. This is amazing!
4
4
54
The resemblance is uncanny! Amazing work by
@GaryMarcus
and
@Rahll
highlighting the plagiarism issue in generative AI.
5
6
52
✨Train your LoRA's upto ~3x faster with our ToDo: Token Downsampling approach, with inference now available on Stable Diffusion WebUI (A1111) !
Much thanks to feffy380 for the implementation :
1
11
42
✨Meta just announced new MTIAv2 chip sets setting new standards:
- 708TF/s Int8, 353TF/s BF16, 256MB SRAM, 128GB memory, 90W TDP. 72 chips/rack,
- Powered by PyTorch, Triton, on TSMC 5nm.
A fully programmable inference powerhouse!
2
8
38
This is amazing, the new xLSTM is parallelizable, with improved exponential gating. Strong results on large-scale language modeling.
I am so excited that xLSTM is out. LSTM is close to my heart - for more than 30 years now. With xLSTM we close the gap to existing state-of-the-art LLMs. With NXAI we have started to build our own European LLMs. I am very proud of my team.
49
382
2K
2
2
39
The total solar eclipse today was so beautiful ! I managed to capture a rare peak through the Toronto clouds of the crescent 🌙
1
4
39
✨ Structured completions using LLMs may have finally been solved and that too 100x faster with 10,000x less overhead !
Amazing work by the awesome team
@RysanaAI
. Been following their work since last year, and super happy to see how far they have come.
AI should be not just smarter but much, much faster. Early last year, we set out to bring the power of large transformers to a level of reliability and performance that feels real-time.
That means <200ms (as close to instant, ideally) for as many workloads as possible. Watch:
11
13
146
5
3
34
We are in a bubble…
"A new wireless mouse with a dedicated AI button that launches Logi AI Prompt Builder"
Guys...this is going too far
83
83
2K
1
4
35
Bittensor, is a great example of decentralized AI where peers rank each other by training networks which learn the value of their neighbors where high ranking peers are monetarily rewarded with additional weight in the network $TAO :
Not going to beat centralized AI with more centralized AI.
All in on
#DecentralizedAI
Lots more 🔜
253
325
2K
3
4
28
March in a nutshell:
- CEO of
@inflectionAI
resigns and joins
@Microsoft
- Creators of Stable Diffusion resign from
@StabilityAI
- President of Vietnam Resigns
- Prime Ministers of Ireland, Bulgaria & Haiti Resign
- First Minister of Wales Resigns
4
3
27
🎉 ToDo: Token Downsampling is now on Hugging Face Spaces. Try it out!
Thanks to
@aningineer
for putting it together, and
@Gradio
for sharing our work 🤩
📢What if you are no longer limited by image size due to memory constraints
This Paper introduces training-free method 𝐓𝐨𝐃𝐨 that downsamples key & value tokens to accelerate SD inference (2x to 4.5x) for high res like 2048x2048 images.
Outperforms in throughput & fidelity.
6
31
123
1
4
23
✨Excited to share 2 short papers on Neural Architecture Search! To appear, as student abstracts in AAAI proceedings.
In this work, with Robert Wu &
@JainRohan16
, we explore:
- Better pre-optimized search space generation.
- Vulnerabilities with the search space design.
(1/3)
3
7
23
Grok Has been Open Sourced:
magnet:?xt=urn:btih:5f96d43576e3d386c9ba65b883210a393b68210e&tr=https%3A%2F%%2Fannounce.php%3Fpasskey%3Decac4c57591b64a7911741df94f18b4b&t
0
3
21
Flash attention 3 is out with faster and better efficiency for long context length and an increase of utilization from 35% to 75% for H100s. Also 1.5x-2x faster with lower error rate for FP8 benchmarked at 2.6x ! Great work by
@tri_dao
and team.
1
1
21
New model just dropped. Mistral 8x 22B
magnet:?xt=urn:btih:9238b09245d0d8cd915be09927769d5f7584c1c9&dn=mixtral-8x22b&tr=udp%3A%2F%%3A1337%2Fannounce&tr=http%3A%2F%%3A1337%2Fannounce
272
822
6K
1
0
20
Definitely worth a read:
1
1
19
Loving
@getairchat
by
@naval
, once the floodgates open would love to invite some of you all to the platform !
2
1
19
Link to the Open AI Event Livestream, starts at 10am PT. Potential new updates include a voice assistant and updated GPT-4(.5?).
0
4
15
Humane AI pin was
@TIME
Magazine's best inventions for 2023 without the product even being tested/reviewed. You really can't make this stuff up.
i think the fate of Humane seems like an all too-common failure in ai startups. not making market/data driven decisions
telling people what they want rather than listening to what they want.
granted people don't always know what they want, but there's definitely ways to scope
2
0
6
0
0
13
🚀
@GoogleAI
just revealed Gemini. The GPT-4 competitor comes in 3 models — Ultra, Pro, and Nano.
This is the FIRST multimodal AI to outperform humans on the MMLU, scoring >90%.
Report from
@GoogleDeepMind
:
🧵Heres everything you need to know:
1
2
11
📃Paper here:
Also available for you to try on Gradio Spaces:
📢What if you are no longer limited by image size due to memory constraints
This Paper introduces training-free method 𝐓𝐨𝐃𝐨 that downsamples key & value tokens to accelerate SD inference (2x to 4.5x) for high res like 2048x2048 images.
Outperforms in throughput & fidelity.
6
31
123
0
6
12
Special shoutout to
@parshantdeep
,
@paul_bridger
@_sshahid_
, and
@CafeSamosaBlue
who've been working really hard to push the boundaries of what is possible in bringing the next AI art revolution!
0
0
12
🌐 We built this model on an entirely new architecture that enables exceptional prompt adherence and consistent character generation.
1
1
12
🥲
RIP Jim Simons
"I did a lot of math. I made a lot of money, and I gave almost all of it away. That's the story of my life."
110
2K
10K
0
0
12
Everything came to a standstill today and we were blessed with a rainbow this evening in Toronto for a few minutes 🌈
If only more people looked up…
0
0
11
@Ash_Stewart_
@maystansstuff
Lmao I managed to get one shot
The total solar eclipse today was so beautiful ! I managed to capture a rare peak through the Toronto clouds of the crescent 🌙
1
4
39
0
0
11
Happy to share that our work, led by
@steveddev
, outlining a pioneer dataset for Nigerian Sign Language (5000 images), has been accepted as an oral paper to the 31st International Joint Conference on Artificial Intelligence!
Check out the preprint here :
My research journey until recently have been haphazard at best - independent work in a not-so-research-inclined climate while leveraging heavily on
@ml_collective
resources[PTO]
Excited to post that this work got accepted + oral at IJCAI-ECAI 2022, AI for Social Good Track!🥳 1/
16
8
62
1
3
11
Bullish on
@cognition_labs
, but the inflated valuation with no VC ROI in sight is scary. I hope a few bad examples don’t set the precedent for the next AI winter and hinder long term growth. Even
@facebook
was valued 8.5MM inflation adjusted in its first year by
@peterthiel
.
Cognition Labs (creator of Devin) is now valued at a $2B
The startup is 6 months old and has no revenue
AI agents are so back
164
357
3K
1
0
11
🛑Open AI API and ChatGPT are down and we can expect further downtime moving forward.
If there are any startups struggling with the situation and need help switching/navigating over to Azure or open source models (Llama 2/Mistral etc) my DM's are open and I'd be happy to help !
0
1
9
Prompt: A beautiful painting of a singular lighthouse, shining its light across a tumultuous sea of blood by Greg Rutkowski and Thomas Kinkade, Trending on Artstation.
1
0
9
Now thats impressive! The movement and mechanics look unreal.
0
1
9
Proud to have worked on this with some exceptional co-authors:
@ShayneRedford
@RobertMahari
@ArielNLee
@campbellslund
@CaimingXiong
@luis_in_brief
@BlancheMinerva
@hanlinliii
@daphneipp
@sarahookr
@jad_kabbara
@alex_pentland
@didaoh
@naana_om
@TobinSouth
and many more!
1
1
9
Absolutely thrilled about Prof. Avi Wigderson's Turing Award – a true genius in computational theory. Sharing lunch with him
@HLForum
last year was unforgettable. His profound insights on non-determinism were incredibly inspiring.
🏆 We're thrilled to announce the recipient of the 2023
#ACMTuringAward
: Avi Wigderson! Wigderson is recognized for his foundational contributions to the theory of computation. Join us in celebrating his incredible achievements! Learn more here:
@the_IAS
12
250
799
0
0
9
Based on everything I have seen, Grok absolutely sucks. Change my mind.
2
0
8
Having actively been engaged in sparsity research for the past year, this gives me PTSD lol
I'm not really an expert on sparsity, but I enjoy using this template, and reminding about learning-rate, whenever I can. So I will:
12
21
278
0
1
8
NeuralArTS: Structuring Neural Architecture Search with Type Theory:
pdf:
arXiv:
--
Towards One Shot Search Space Poisoning in Neural Architecture Search
pdf:
arXiv:
(3/3)
0
1
8
As the year draws to a close, I'm reflecting on my amazing 2023 journey. Here's highlights from my year beautifully captured through DALL-E 2's lens. Wishing everyone a fantastic year ahead in 2024 !
0
1
8
Try it out here, now in preview for all users:
0
1
6
Apparently OpenAI’s new breakthrough is Q-learning —something that has existed for 30 years at-least 😭
The AI ecosystem is so cringe right now, people are ready to just hype up anything. Unbelievable, how we are celebrating nothingness
2
1
7
AI researchers at Intel are definitely onto something. I wish their foundry business and other verticals were as promising…
🔥Wanted to quantize LLMs with best accuracy & smallest size, Intel Neural Compressor is your choice. We just released v2.6 featuring SOTA LLM quantizer, outperforming GPTQ/AWQ on typical LLMs.
🎯Quantized LLM leaderboard:
Github:
1
34
184
0
0
7
This work would not be possible without the amazing feedback from George-Alexandru Adam (
@uoft
) & Kenyon Tsai (
@VectorInst
)
Special thanks, to
@savvyRL
,
@jasonyo
& the entire
@ml_collective
community for their continued support, insightful discussions and compute grant🙂
(2/3)
1
0
7
🌟 Finally, ChatGPT's uses differ from web domains' primary purposes . Much of the highest quality data comes from commercialized sources like ads & paywalls. However, the WildChat dataset shows ChatGPT is mainly used for creative writing, sexual roleplay, and brainstorming.
1
1
7
Alternatively, for those who prefer Comfy UI there is a node already!
0
1
6
There’s no curse like always getting what you want in life.
1
0
6
🌟The key finding: Access to training data is rapidly declining. In less than a year, tokens revoking consent in C4/RefinedWeb increased significantly: over 5% of all tokens, over 30% from top-2k active domains, and 40%+ from sites with anti-crawling terms.
1
0
5
Check out
@steveddev
's presentation on 2nd December
@MasakhaneNLP
!
The talk will focus on the development of a pioneer dataset for Nigerian Sign Language (5000 images), and outline key results from our preprint :
📢
@MasakhaneNLP
is delighted to host
@steveddev
to presents his fantastic work bringing the benefits of
#NLProc
to more people
@AiDisability
#Diversity
#Inclusion
#Accessibility
Please join our slack to get zoom links at
2
10
26
2
2
5
I am happy to share that AGI has been achieved internally finally.
1
1
5
Here's the cat itself.
0
0
5
🌟 How do restrictions vary by company? OpenAI is restricted the most, with 26% of web domains blocking it. Anthropic and Common Crawl each face 13% restrictions, and Google AI 10%. Cohere (5%) and Meta (4%) are less restricted.
1
0
5
✨Exciting news for users of Google Colab! 🧑💻 They've just started rolling out AI coding features using their new Codey models, built on PaLM 2.
1
0
5
NVIDIA is like the IBM / Xerox of AI, pre-Apple era. I feel the GPU markets will be taken over by another player down the line.
Apple of AI exists. It’s called NVIDIA. Hardware is GPUs. Software is CUDA.
30
34
550
4
1
5
We don’t know if it’s the best, you gotta release it first for us to see that.
0
1
5
So seems like
@trsohmers
the ex-director of technology from
@GroqInc
went ahead and made his own inference box,
@positron_ai
debuted at NeurIPS.
Llama 2 70B performance:
Batch 1 → 480 tokens/sec/user
Batch 8 → 1,280 tokens/sec/batch
(160 tokens/sec/user)
1
0
5
Thanks
@mozilla
for featuring our work and supporting the open future of the web!
Can
#AI
training undermine the Open Web?
The Data Provenance Initiative thinks so.
The
#DataFuturesLab
awardee has found that blunt restrictions from the websites that feed popular training datasets could have an impact on the data-economy⤵️
0
14
38
0
0
5
Having seen the tech in action, can definitely say its amazing. Worth checking it out !
0
0
4
@Ethan_smith_20
Kinda similar but I ran a bunch of experiments on depth augmented activations so basically modifying the activation function as you go deeper into the network. For instance changing the angle within ReLU. Saw very marginal training improvements on toy networks.
1
0
4
Discovering that the most profound victories are those whispered, not announced. A melody in the soul is more profound than the world's loudest applause.
BRAG ABOUT SOMETHING YOU’RE PROUD OF ACCOMPLISHING IN 2023 ✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨
1K
470
6K
0
0
4
Learn more about this exciting news here 👉
0
1
5
“AI will not replace developers”
-
@thomasdomke
, CEO Github
Was an absolute pleasure listening to the
@ashtom
talking about the future of AI today and future of work
@CollisionHQ
1
0
4
2
0
4
Had a lovely time reading through the little book of deep learning by
@francoisfleuret
. Definitely recommended for people who want a handbook that covers everything you need to know in ~100pages !
0
1
4
I've been at NVIDIA for 6 years and 3 promotions. My compensation has gone up 500% since I started.
Work hard at a good company you believe in, and you will be rewarded.
338
121
2K
1
1
4
@steveddev
Congratulations Steven ! I am sure this is the first in many amazing papers to come :)
1
0
4
e/acc
1
0
4
🎨 One of the key features of Pheonix is coherent text in image. You can generate clear, accurate text within images, perfect for banners, posters, and logos.
1
0
4
@savvyRL
@peterdavidfagan
@andrey_kurenkov
@k_saifullaah
@steveddev
@cchoi314
Thank you too for organizing this !
0
0
3
Gemini AI's prowess goes beyond benchmarks. For instance, it can reads messy handwriting in physics problems, converts them to mathematical typesetting, identifies errors, and provides correct solutions. This breakthrough in multimodal reasoning opens thrilling possibilities.
1
0
4
This is literally from 2018. I was there in the audience. I am not sure why people on LinkedIn and Twitter are hyping up - this retinal myopathy use-case from 5 years ago.
Sundar Pichai has unveiled a revolutionary
#healthcare
technology that employs
#AI
and eye scans to predict cardiovascular events with remarkable accuracy. This new development could potentially eliminate the necessity for CT scans, MRIs, and X-rays, enabling doctors to obtain a
203
1K
4K
2
0
4
Read more here:
0
0
4
Diffusion models need to become better at text generation in images.
LLMs need to become better at structured JSON completions.
Hopefully we solve these soon.
1
1
4
What a wonderful group of people ! Truly grateful to
@ZEISS_Group
for sponsoring us for this opportunity to spend a week in Germany and meet various laureates.
#HLF23
Great to meet all the Abbe grant holders of
@HLForum
as well as the program managers of
#CZS
for lunch. It was particularly interesting to learn about the connection to
@ZEISS_Group
and
@SCHOTT_AG
, how the foundation operates or the career path of those working at
#CZS
.
#HLF23
0
2
8
0
0
3
This approach performs on par or better when compared to previous approaches like ToMe while being closer to the baseline in image quality and fidelity.
We begin at ToMe, a previous method used to reduce attention computation for stable diffusion. It works by merging tokens with the highest cosine similarity thus reducing the number of tokens input to attention.
1
0
4
1
1
2