TLDR Index
TLDR AI

TLDR AI 2023-04-05

5 sections14 entriesinline urlcomplete parse

Parse notes

  • Subscription/account footer truncated before parsefooter_truncated

01HEADLINES & LAUNCHES

02RESEARCH & INNOVATION

ANOTHER LLAMA MODEL JOINS THE ZOO

Similar to Alpaca from Stanford, Koala was trained on dialog and instruction data. However, they didn’t use distillation from the OpenAI models, but sourced their data from freely available sources on the web. Their model actually seems comparable to strong closed source models and the code to fine-tune along with the dataset are available for further study.

bair.berkeley.edu8 min read

03ENGINEERING & RESOURCES

FOLLOW YOUR POSE: TEXT-TO-VIDEO GENERATION USING POSE-FREE VIDEOS

This paper proposes a two-stage training scheme for generating text-editable and pose-controllable character videos, using easily obtainable datasets and a pre-trained text-to-image model. The method successfully generates pose-controllable character videos and retains the editing abilities of the pre-trained model.

follow-your-pose.github.ioGitHub repository

FAST REAL-TIME VIDEO ENHANCEMENT

ReBotNet is a video model that is designed to enhance live streams and video conferences in real-time. They use a dual-branch network that learns different types of video abstractions like location and time based ones. They then use a common decoder network to output the enhanced frames. It runs at 50 fps, uses about 5gb of memory and performs on par with models 8x larger.

jeya-maria-jose.github.io3 min read

BABYAGI

Babyagi is a Python AI-powered task management system.

github.comGitHub repository

04MISCELLANEOUS