Shengqiong Wu

Hi @JiyouShin , I got the same error. Have you found any solutions?

In my program, the bug **_seems_** to be triggered that there is insufficient CPU memory when saving the checkpoint.

Hi, @jwzhi I guess you are using the updated Vicuna (not v0) because I also encountered the issue with the updated tokenizer where it tokenizes “[IMG0] [IMG1] [IMG2] [IMG3]” into...

Currently, the model does not support image editing.

The dataset has been released. You can find it in `/data/IT_data/T-T+X_data` file.

Hi, we have already released the new version of the code, including detailed datasets and a modified framework. Please kindly refer to it.

Hi, this is right. The downloaded images are stored in a `tar` file. And the `*_stats.json` file provides information about the download status, including the total number of images, the...

> > 下载 cc3m 数据集时，不断显示错误：'Field 'caption' does not exist in table schema'。 > > 在查看 img2dataset 文档后，发现需要添加以下内容 > > `pip install sed` > > `sed -i '1s/^/caption\turl\n/' Train_GCC-training.tsv` > >...

NExT-GPT also supports other languages like Chinese or Korean, even though we did not use other languages to train this model. I guess this should be contributed to the backbone...

Hi, the released checkpoint includes all training parameters across all three stages. During the inference stage, we indeed load the pre-trained parameters from "7b_tiva_v0" after model initialization. Please refer to...