_pranavan's profile picture. GPU poor :(

Pranav

@_pranavan

GPU poor :(

Pranav 已轉發

NDA = not doing anything

aditiitwt's tweet image. NDA = not doing anything

Pranav 已轉發

Get your wife pregnant in February. Paternity leave in November. Thank me later.

‘GTA 6’ has been delayed to November 19, 2026.

DiscussingFilm's tweet image. ‘GTA 6’ has been delayed to November 19, 2026.


Pranav 已轉發

hmmm

whatRanjuSaid's tweet image. hmmm

🔥🚀

Hiring for a new team building computer control agents. Join us to build Grok5 / macrohard later this year. DM me! Will send out a job post soon too.



Pranav 已轉發

me trying to keep up with new AI tools

DataChaz's tweet image. me trying to keep up with new AI tools

Pranav 已轉發
AdrianDittmann's tweet image.

Been waiting for this moment! 🔥🚀

I resisted AI for too long Living in denial Now it is game on @xAI @Tesla @SpaceX



AI will become super intelligent very soon, leaving time for humans to think about their existence. As a result, a new collective intelligence will be born, which will restore peace among humans, and it will also show us the path to eternal happiness.


Pranav 已轉發

I’ve never seen this app so distilled

NaderLikeLadder's tweet image. I’ve never seen this app so distilled

Pranav 已轉發

you know you wrote mega paper when it has its own wikipedia page.

Hesamation's tweet image. you know you wrote mega paper when it has its own wikipedia page.

Pranav 已轉發

Apple’s brilliant new AI strategy is to prove it doesn’t exist.


RIP resume generators

_pranavan's tweet image. RIP resume generators

Fu*king Legend 🫡

BRO IM ON FUCKING FIRESHIPPPPPPPP

im_roy_lee's tweet image. BRO IM ON FUCKING FIRESHIPPPPPPPP


Depth Any Camera (DAC) by @33yuliangguo redefines zero-shot metric depth estimation, excelling across fisheye, 360, and perspective images. Trained only on perspective data, it delivers state-of-the-art accuracy with smaller datasets. Seamless generalization, fewer annotations,…


HaWoR by @JiankangDeng is a cutting-edge method for reconstructing 3D hand motions in world coordinates from egocentric videos. Combining adaptive SLAM and a novel motion infiller achieves state-of-the-art performance, even when hands move out of view. hawor-project.github.io


ProTracker by @Tingyang030707 delivers robust, accurate long-term tracking of arbitrary video points by probabilistically integrating optical flow and semantic features. It outperforms existing methods, handling occlusions and maintaining smooth trajectories. A game-changer for…


EvTexture sets a new benchmark in video super-resolution by leveraging event signals for texture enhancement. Its iterative module delivers unmatched detail restoration, achieving up to 4.67dB gain on Vid4. A breakthrough in VSR texture refinement! dachunkai.github.io/evtexture.gith…


VideoAnydoor by DAMO Academy is a groundbreaking AI for zero-shot video object insertion, offering high fidelity detail preservation and precise motion control. It enables applications like virtual try-ons, face swapping, logo insertion, and multi region editing, all without task…


TokenHMR by @saidwivedi redefines 3D human pose estimation with a novel token-based representation and adaptive loss scaling, achieving state-of-the-art accuracy on in-the-wild benchmarks. A game-changer for precise and robust human mesh recovery. tokenhmr.is.tue.mpg.de


RTMPose and RTMW by @OpenMMLab are high-performance, real-time models for multi-person 2D/3D pose estimation. RTMPose achieves up to 75.8% AP on COCO with over 430 FPS, while RTMW delivers over 940 FPS on GPUs, making them ideal for efficient and accurate real-world applications.…


Loading...

Something went wrong.


Something went wrong.