Images for training - Githubissues

PKU-YuanGroup / MoE-LLaVA

Mixture-of-Experts for Large Vision-Language Models

https://arxiv.org/abs/2401.15947

Apache License 2.0

1.99k stars 127 forks source link

Images for training #4

Closed phellonchen closed 9 months ago

phellonchen commented 9 months ago

Could you share the sampled images for Moe training, i.e. the images for Stage II SViT-157k, LVIS-220k LRV-331k, MIMIC-IT-256k

LinB203 commented 9 months ago

here to download. We have packaged the data for downloading.

SizeWu commented 8 months ago

Hi! Thanks for preparing and open-sourcing all of these datasets. I noticed that you use the LA images in MIMIC-IT. May I know wether you applied in-context tuning when training the model or the LA images were just used in ordinary QA format?