TLDR Index

Search the archive

Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.

Results

14,500 matches

IT

2023.03.01
cognitiverevolution.substack.com10 min readTLDR

REALFUSION: 360° RECONSTRUCTION OF ANY OBJECT FROM A SINGLE IMAGE

The paper describes a method for reconstructing a full 360◦ photographic model of an object using a single image of it. The approach involves fitting a neural radiance field to the image, but the problem is found to be severely ill-posed. To address this, an off-the-shelf conditional image generator based on diffusion is used, and a prompt is engineered to encourage it to "dream up" novel views of the object. The DreamFusion method is used to fuse the given input view, the conditional prior, and other regularizers in a final, consistent reconstruction.

2023.03.01
arxiv.org15 min readTLDR AI

LANGUAGE IS NOT ALL YOU NEED

MLLM is a multimodal transformer trained on internet scale text/image data. It uses a frozen CLIP image encoder and a Magneto decoder only transformer. It has 1.3B parameters and was trained for 360B tokens. It was trained with TorchScale, a new exciting framework, and finished off with language only instruction tuning. It got 26% on an IQ test, which is better than random chance. It has 73% zeroshot ImageNet classification and positive transfer between modalities. It has a 64k vocab size and no code yet but should be coming soon.

2023.03.01
arxiv.org23 min readTLDR AI

OPENAI’S WHISPER LARGE ON AN IPHONE

Georgi has been working on Whisper.cpp which dramatically speeds up the runtime of OpenAI’s state of the art speech to text model. He has included a WASM version for the web. This latest model uses 4-bit integer quantization. Interestingly, this type of quantization only seems to work on larger models and fails on smaller models.

2023.03.01
github.comGitHub repositoryTLDR AI

CACTI: A FRAMEWORK FOR IMITATION LEARNING

The authors propose a framework for scaling robot learning in multi-task and multi-scene manipulation in kitchen environments, both in simulation and in the real world. The proposed framework, called CACTI, comprises four stages that handle data collection, data augmentation, visual representation learning, and imitation policy training. The approach uses state-of-the-art generative models for data augmentation and pre-trained visual representations to improve training efficiency. Experimental results show that CACTI enables efficient training of a single policy that can perform 10 manipulation tasks involving kitchen objects on a real robot setup, and 18 semantic tasks across 100 layout variations for each individual task in a simulated kitchen environment.

2023.03.01
cacti-framework.github.ioGitHub repositoryTLDR AI

SCENE DIFFUSER

Scene Diffuser is a conditional generative model for 3D scene understanding. This repository is the official implementation of paper "Diffusion-based Generation, Optimization, and Planning in 3D Scenes". SceneDiffuser provides a unified model for solving scene-conditioned generation, optimization, and planning. In contrast to prior work, SceneDiffuser is intrinsically scene-aware, physics-based, and goal-oriented.

2023.03.01
github.comGitHub repositoryTLDR AI

T3-TURBO-AI

T3-Turbo-AI is a starter repo for building full-stack, type-safe, LLM-powered web apps.

2023.03.01
github.comGitHub repositoryTLDR AI

LORA SD MODELS FOR PANORAMA GENERATION

LoRA models can be plugged into a Stable Diffusion pipeline to change out the styles and quality of generated images. This model from Latent Labs allows you to generate panoramas and beautiful 360 scenes.

2023.03.01
civitai.com1 min readTLDR AI

Sponsored

IS WEB3 SERIOUS ABOUT SECURITY?

Too much web3 code is insecure. Hackers know this, and will pounce on any chance to steal cryptocurrency and wreak havoc. That’s why the best web3 developers protect their source code and smart contracts with 3rd-party security audits

2023.03.01
coinspect.comSponsorTLDR Crypto

SUSPENSE

This repository contains APIs to simplify data loading and caching for use with React Suspense. Suspense is a powerful feature that allows developers to declaratively specify loading states for parts of the component tree. It works best when deeply integrated into an application's architecture and combined with the transition API. Suspense makes the 'UI loading state' a first-class declarative concept in the React programming model, allowing developers to build higher-level features on top of it.

2023.02.28
suspense.vercel.appGitHub repositoryTLDR

REMIX UTILS

Remix Utils contains simple utility functions to use with Remix.run. Remix is a full stack web framework that provides a fast, slick, and resilient user experience. It leverages web standards, distributed systems, native browser features, and the Web Fetch API to provide snappy page loads and instant transitions. Remix can run on Cloudflare Workers, serverless, and traditional Node.js environments.

2023.02.28
github.comGitHub repositoryTLDR

2023.02.28
techcrunch.com2 min readTLDR

VOICEMOD RAISES $14.5M TO RIDE THE GENERATIVE AI (SONIC)BOOM

Voicemod has become the leading creator in real-time voice changing and soundboard technology. Their mission is to enable everyone to express themselves through sound. They’ve built expressive and immersive audio tools, making it easy to create unique sonic identities and enable interactions with personalized sounds. This will be huge as the podcast, voiceover, and audiobook spaces have blown up the last few years.

2023.02.28
techcrunch.com5 min readTLDR AI