Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
  • Loading branch information
scofield7419 authored May 13, 2024
1 parent a6f2426 commit d8dc1ef
Showing 1 changed file with 2 additions and 6 deletions.
8 changes: 2 additions & 6 deletions documents/chapter6/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -192,11 +192,7 @@ pip install -r requirements.txt

#### 2.4.1 加载预训练的NExT-GPT模型checkpoint

- **步骤1**:加载`冻结参数`

[NExT-GPT](https://github.com/NExT-GPT/NExT-GPT) 是基于以下现有模型或模块进行训练的。
请按照以下说明准备checkpoint。

- **步骤1**:加载`冻结参数`[NExT-GPT](https://github.com/NExT-GPT/NExT-GPT) 是基于以下现有模型或模块进行训练的, 请按照以下说明准备checkpoint。
- `ImageBind` 是统一的图像/视频/音频编码器。可以从[此处](https://dl.fbaipublicfiles.com/imagebind/imagebind_huge.pth)下载预训练检查点,版本为`huge`。然后,将`imagebind_huge.pth`文件放置在[[./ckpt/pretrained_ckpt/imagebind_ckpt/huge]](ckpt/pretrained_ckpt/imagebind_ckpt/)
- `Vicuna`:首先按照[[这里]](ckpt/pretrained_ckpt/prepare_vicuna.md)的说明准备LLaMA。然后将预训练模型放置在[[./ckpt/pretrained_ckpt/vicuna_ckpt/]](ckpt/pretrained_ckpt/vicuna_ckpt/)
- `Image Diffusion` 用于生成图像。NExT-GPT 使用版本为`v1-5`[Stable Diffusion](https://huggingface.co/runwayml/stable-diffusion-v1-5)。(_代码里将会自动下载_)
Expand Down Expand Up @@ -225,7 +221,7 @@ bash scripts/app.sh
目前的版本能够支持文字、图像、视频、声音四种模态下任意组合的输入,并任务组合模态的输出。
并且支持多轮上下文交互。

请各位自行运行测试
请各位自行运行测试效果

- **Case-1**:输入T+I,输出T+A

Expand Down

0 comments on commit d8dc1ef

Please sign in to comment.