Sun Oct 30 2022
Thu Oct 27 2022

What Language Model to Train if You Have One Million GPU Hours?

Artificial Intelligence
Natural Language Processing
Machine Learning
NLP applications
Language modeling
Multilingual models

Compares different modeling practices impact on zero-shot generalization

Provides insights on best modeling practices and pre-training corpora for billion-parameter scale language models

https://huggingface.co/bigscience
https://arxiv.org/pdf/2210.15424.pdf
https://arxiv.org/abs/2210.15424
https://twitter.com/arankomatsuzaki/status/1585796195238060032/photo/1

ERNIE-ViLG 2.0

Artificial Intelligence
Computer Vision
Natural Language Processing
Image generation
Text-to-image conversion
Multimodal learning

Proposes a Chinese text-to-image diffusion model to improve image fidelity and text relevancy

Can be used to generate high-quality images with text conditions in various applications, especially in the Chinese-speaking market

https://arxiv.org/pdf/2210.15257.pdf
https://arxiv.org/abs/2210.15257
https://twitter.com/arankomatsuzaki/status/1585794707715633153/photo/1
Wed Oct 26 2022
Tue Oct 25 2022
Mon Oct 24 2022
Thu Oct 20 2022