David @deeplearninpapi

ML Engineer, Researcher Joined January 2023

Tweets

31
Followers

2
Following

166
Likes

104

Surya Dantuluri @sdand

a year ago

By 21, you've lived half your life—perceptually Our brain's internal clock is predominantly dopaminergic: novel experiences slow time, routine accelerates it. This time warp means perceived age outpaces real age(studies prove this) Solution? Just do more cool shit. Blog below!

Sam Altman @sama

8 years ago

88 1K 4K 0 1K

134 1K 12K 2.2M 9K

Download Image

Andrej Karpathy @karpathy

a year ago

⚡️ Excited to share that I am starting an AI+Education company called Eureka Labs. The announcement: --- We are Eureka Labs and we are building a new kind of school that is AI native. How can we approach an ideal experience for learning something new? For example, in the case…

2K 4K 28K 2.4M 8K

Download Image

Patrick Hsu @pdhsu

a year ago

What if we could universally recombine, insert, delete, or invert any two pieces of DNA? In back-to-back @Nature papers, we report the discovery of bridge RNAs and 3 atomic structures of the first natural RNA-guided recombinase - a new mechanism for programmable genome design

187 1K 5K 1.3M 2K

Download Video

arXiv Sound @ArxivSound

a year ago

``LiveSpeech: Low-Latency Zero-shot Text-to-Speech via Autoregressive Modeling of Audio Discrete Codes,'' Trung Dang, David Aponte, Dung Tran, Kazuhito Koishida, ift.tt/ErQFjm1

0 2 25 3K 14

Rohan Paul @rohanpaul_ai

a year ago

@MistralAI In the past 24 hours we got Mixtral 8x22B, Unsloth crazy performance upgrades, an entire new architecture from Google outperforming Transformers (Griffin), Command R+ support in llama.cpp, and news of Llama 3! If this is not a revolutionary time to live, what is.

11 31 291 30K 69

Daniel Han @danielhanchen

a year ago

Had a look through @grok's code: 1. Attention is scaled by 30/tanh(x/30) ?! 2. Approx GELU is used like Gemma 3. 4x Layernoms unlike 2x for Llama 4. RMS Layernorm downcasts at the end unlike Llama - same as Gemma 5. RoPE is fully in float32 I think like Gemma 6. Multipliers are 1…

26 223 1K 233K 952

Download Image

Jason Wei @_jasonwei

2 years ago

It was an honor to give a guest lecture yesterday at Stanford’s CS330 class, "Deep Multi-Task and Meta-Learning"! I discussed a few very simple intuitions for how I personally think about large language models. Slides: docs.google.com/presentation/d… Here are the six intuitions: (1)…

22 276 2K 277K 2K

Download Image

Alex Volkov (Thursd/AI) @altryne

2 years ago

Watching @karpathy presentation from today and taking twitter notes, come along for the ride: If you're like only the practical tips, skip to #32 @karpathy starts with stages: 1 - Pre-training - months x thousands of GPUs 2, 3, 4 - Finetuning stages that take hours or days 1/

64 637 3K 1.1M 4K

Download Image

Shashank Prasanna @shshnkp

2 years ago

New blogpost! a visual primer on how @PyTorch 2.0 compiler technologies for graph capture, IRs, operator fusions and automatic C++ and @NVIDIAAIDev GPU code generation. This is your one stop shop to grok PyTorch 2.0's torch.compile() API Summary 🧵👇 medium.com/towards-data-s…

7 85 353 88K 214

Sam Altman @sama

2 years ago

we had a significant issue in ChatGPT due to a bug in an open source library, for which a fix has now been released and we have just finished validating. a small percentage of users were able to see the titles of other users’ conversation history. we feel awful about this.

567 559 7K 3.4M 597

Mathieu @miniapeur

3 years ago

3 173 1K 60K 29

Download Image

Sharon Zhou @realSharonZhou

3 years ago

The more of an expert you are, the more you feel you’re “late to the party” when something new comes out. I felt late to AI in 2015, late to generative AI in 2017. Idk who needs to hear this but: it’s still just beginning. You’re not late at all. Come make some noise.

22 67 978 103K 91

François Chollet @fchollet

3 years ago

We're just at the start of the golden age of Applied ML.

46 197 2K 297K 102

David @deeplearninpapi

3 years ago

Hello @memdotai 👋 just setting up my Twitter flow. Here's my validation code: ewr4e7whbUagfNq1MAc2

0 0 1 19 0

Bojan Tunguz @tunguz

3 years ago

Good morning everyone, except for the companies that are searching for junior prompt engineer with a minimum of 5 years of experience.

12 16 323 32K 1

Andrej Karpathy @karpathy

3 years ago

debugging in Python: - `print()`s alone: too simple - `import pdb; pdb.set_trace()`: too complex - `import code; code.interact(local=locals())`: just right simply drops you into interpreter, perfect for 95% of debugging

132 674 7K 1.0M 3K

Lior⚡ @LiorOnAI

3 years ago

GoogleAI just released "Muse", a text-to-image generation/editing model via Masked Generative Transformers: - Achieves new SOTA - Zero-shot, Mask-free editing - Zero-shot Inpainting/Outpainting - 900M params 📄 Paper: arxiv.org/abs/2301.00704 ⚙️ Project: muse-model.github.io