Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/magic-research/bubogpt
BuboGPT: Enabling Visual Grounding in Multi-Modal LLMs
https://github.com/magic-research/bubogpt
Last synced: 2 months ago
JSON representation
BuboGPT: Enabling Visual Grounding in Multi-Modal LLMs
- Host: GitHub
- URL: https://github.com/magic-research/bubogpt
- Owner: magic-research
- License: bsd-3-clause
- Created: 2023-07-16T14:54:40.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2023-07-21T15:57:00.000Z (about 1 year ago)
- Last Synced: 2024-06-15T05:36:03.736Z (4 months ago)
- Language: Python
- Homepage: https://bubo-gpt.github.io/
- Size: 6.35 MB
- Stars: 481
- Watchers: 10
- Forks: 34
- Open Issues: 11
-
Metadata Files:
- Readme: README.md
- License: LICENSE.md
Awesome Lists containing this project
- awesome-llm-and-aigc - BuboGPT - research/bubogpt?style=social"/> : "BuboGPT: Enabling Visual Grounding in Multi-Modal LLMs". (**[arXiv 2023](https://arxiv.org/abs/2307.08581)**). (Summary)
- awesome-llm-and-aigc - BuboGPT - research/bubogpt?style=social"/> : "BuboGPT: Enabling Visual Grounding in Multi-Modal LLMs". (**[arXiv 2023](https://arxiv.org/abs/2307.08581)**). (Summary)
- awesome-llm-and-aigc - BuboGPT - research/bubogpt?style=social"/> : "BuboGPT: Enabling Visual Grounding in Multi-Modal LLMs". (**[arXiv 2023](https://arxiv.org/abs/2307.08581)**). (Summary)
README
# BuboGPT: Enabling Visual Grounding in Multi-Modal LLMs
*A multi-modal LLM capable of jointly understanding of text, vision and audio and grounding knowledge into visual objects.*
[[Project Page](https://bubo-gpt.github.io/)] [[Arxiv](https://arxiv.org/abs/2307.08581)] [[Demo Video](https://youtu.be/uRdmC3wPz6k)] [[Gradio](https://huggingface.co/spaces/magicr/BuboGPT)] [[Data](https://huggingface.co/datasets/magicr/BuboGPT/tree/main)] [[Model](https://huggingface.co/magicr/BuboGPT-ckpt/tree/main)]
**BuboGPT: Enabling Visual Grounding in Multi-Modal LLMs**
[Yang Zhao*](https://scholar.google.com/citations?user=uPmTOHAAAAAJ&hl=en), [Zhijie Lin*](https://scholar.google.com/citations?user=xXMj6_EAAAAJ&hl=en), [Daquan Zhou](https://scholar.google.com/citations?user=DdCAbWwAAAAJ&hl=en), [Zilong Huang](https://scholar.google.com.sg/citations?user=GW9vw8UAAAAJ&hl=en), [Jiashi Feng](https://sites.google.com/site/jshfeng/) and [Bingyi Kang†](https://scholar.google.com.sg/citations?user=NmHgX-wAAAAJ&hl=en) (*Equal Contribution, †Project Lead)
Bytedance Inc.
[](https://huggingface.co/spaces/magicr/BuboGPT)
## News🔥
2023/07/21 - Huggingface demo released!
## Setup
Clone this repository and navigate to the current folder.
### Environment
Our code is based on Python 3.9, CUDA 11.7 and Pytorch 2.0.1.
```bash
pip3 install -r pre-requirements.txt
pip3 install -r requirements.txt
```
### Models
Follow the [instruction](./PrepareVicuna.md) to prepare the pretrained Vicuna weights,
and update the `llama_model` in `bubogpt/configs/models/mmgpt4.yaml`.
```bash
## get pre-trained checkpoints
mkdir checkpoints && cd checkpoints;
wget https://huggingface.co/spaces/Vision-CAIR/minigpt4/resolve/main/blip2_pretrained_flant5xxl.pth;
wget https://huggingface.co/spaces/xinyu1205/recognize-anything/resolve/main/ram_swin_large_14m.pth;
wget https://github.com/IDEA-Research/GroundingDINO/releases/download/v0.1.0-alpha/groundingdino_swint_ogc.pth;
wget https://huggingface.co/spaces/abhishek/StableSAM/resolve/main/sam_vit_h_4b8939.pth;
wget https://huggingface.co/magicr/BuboGPT-ckpt/resolve/main/bubogpt_7b.pth
```
For training, down load MiniGPT-4 [checkpoint](https://drive.google.com/file/d/1RY9jV0dyqLX-o38LrumkKRh6Jtaop58R/view?usp=sharing) to `checkpoints`.
### Data
#### Stage1
- Image-Text Data: Following [MiniGPT4's instruction](https://github.com/Vision-CAIR/MiniGPT-4/blob/main/dataset/README_1_STAGE.md) to prepare the stage1 dataset.
- Audio-Text Data: Following our [audio data instruction](dataset/README.md#audio-dataset-instruction) to prepare it.
#### Stage2
- MiniGPT4's visual instruction-following data: following their [prepartion doc](https://github.com/Vision-CAIR/MiniGPT-4/blob/main/dataset/README_2_STAGE.md).
- LLaVA's visual instruction-following data: refer to [LLaVA Visual Instruct 150K Dataset Card](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K)
- Our audio instruction-following data: Following our [audio data instruction](dataset/README.md#audio-dataset-instruction) to prepare it.
- Our image-audio sound localization data: Following our [image-audio data instruction](dataset/README.md#image-audio-dataset-instruction) to prepare it.
- Our image-audio negatively paired data: Following our [image-audio data instruction](dataset/README.md#image-audio-dataset-instruction) to prepare it.
### Usage
#### Gradio demo
Run gradio demo with:
```bash
python3 app.py --cfg-path eval_configs/mmgpt4_eval.yaml --gpu-id 0
```
#### Training
Browse the [dataset config](bubogpt/configs/datasets) folder, and replace the `storage` item with `path/to/your/data` for each dataset.
Stage 1: Audio pre-training
```
bash dist_train.sh train_configs/mmgpt4_stage1_audio.yaml
```
Stage2: Multi-modal instruct tuning
- Put `path/to/stage1/ckpt` to `ckpt` in [train_configs/mmgpt4_stage2_mm.yaml](train_configs/mmgpt4_stage2_mm.yaml)
```
bash dist_train.sh train_configs/mmgpt4_stage2_mm.yaml
```
## Demo
#### 1. Image Understanding with Grounding
#### 2. Audio Understanding
#### 3. Aligned Audio-Image Understanding
#### 4. Arbitrary Audio-Image Understanding
For more demonstrations, please refer to the [examples](/examples/).
## Acknowledgement
This codebase is mainly developed based on the following repos:
- https://github.com/Vision-CAIR/MiniGPT-4
- https://github.com/IDEA-Research/GroundingDINO
- https://github.com/xinyu1205/recognize-anything
- https://github.com/facebookresearch/segment-anything