README.md

[![Gradio](https://img.shields.io/badge/Gradio-Online%20Demo-blue)](http://103.170.5.190:7860/)
[![Open in OpenXLab](https://cdn-static.openxlab.org.cn/app-center/openxlab_app.svg)](https://openxlab.org.cn/apps/detail/openxlab-app/LISA)

# LISA: Reasoning Segmentation via Large Language Model

<font size=7><div align='center'><b>LISA</b>: Large <b>L</b>anguage <b>I</b>nstructed <b>S</b>egmentation <b>A</b>ssistant</div></font>

<font size=7><div align='center' > <a href=https://arxiv.org/pdf/2308.00692.pdf>**Paper**</a> | <a href="https://huggingface.co/xinlai">**Models**</a> | [**Training**](#training) | [**Inference**](#inference) | [**Local Deployment**](#deployment) | [**Dataset**](#dataset) | <a href="http://103.170.5.190:7860/">**Online Demo**</a></div></font>

<!-- <p align="center"> <img src="imgs/teaser.jpg" width="100%"> </p> -->

<table class="center">
<tr>
  <td style="text-align:center;"><b>Input</b></td>
  <td style="text-align:center;"><b>Output</b></td>
  <td style="text-align:center;"><b>Input</b></td>
  <td style="text-align:center;"><b>Output</b></td>
</tr>
<tr>
  <td><img src="imgs/obama.jpg"></td>
  <td><img src="vis_output/obama.jpg"></td>
  <td><img src="imgs/trump.jpg"></td>              
  <td><img src="vis_output/trump.jpg"></td>
</tr>
<tr>
  <td width=25% style="text-align:center;color:gray;">"Who was <ins>the president of the US</ins> in this image? Please output segmentation mask and explain the reason."</td>
  <td width=25% style="text-align:center;">"Sure, the segmentation result is [SEG]. The President of the United States in the image is <ins>President Obama.</ins>”</td>
  <td width=25% style="text-align:center;color:gray">"Who was <ins>the president of the US</ins> in this image? Please output segmentation mask and explain why."</td>
  <td width=25% style="text-align:center;">"Sure, [SEG]. In the image, the President of the United States is <ins>President Trump.</ins>"</td>
</tr>

<tr>
  <td><img src="imgs/stand_higher.jpg"></td>
  <td><img src="vis_output/stand_higher.jpg"></td>
  <td><img src="imgs/camera_lens.jpg"></td>              
  <td><img src="vis_output/camera_lens.jpg"></td>
</tr>
<tr>
  <td width=25% style="text-align:center;color:gray;">"What can <ins>make the woman stand higher</ins>? Please output segmentation mask and explain why."</td>
  <td width=25% style="text-align:center;">"Sure, [SEG]. The woman is <ins>standing higher by using a ladder</ins>..."</td>
  <td width=25% style="text-align:center;color:gray">"Can you segment <ins>the camera lens that is more suitable for photographing nearby objects</ins> in this image?"</td>
  <td width=25% style="text-align:center;">"Sure, [SEG]."</td>
</tr>

<tr>
  <td><img src="imgs/dog_with_horn.jpg"></td>
  <td><img src="vis_output/dog_with_horn.jpg"></td>
  <td><img src="imgs/wash_hands.jpg"></td>              
  <td><img src="vis_output/wash_hands.jpg"></td>
</tr>
<tr>
  <td width=25% style="text-align:center;color:gray;">"Can you segment <ins>the unusual part</ins> in this image and explain why."</td>
  <td width=25% style="text-align:center;">"Sure, it is [SEG]. In the image, the unusual part is <ins>the dog wearing a reindeer antler headband</ins>..."</td>
  <td width=25% style="text-align:center;color:gray">"Where to <ins>wash hands</ins> in this image? Please output segmentation mask."</td>
  <td width=25% style="text-align:center;">"Sure, [SEG]."</td>
</tr>

<tr>
  <td><img src="imgs/jackma.jpg"></td>
  <td><img src="vis_output/jackma.jpg"></td>
  <td><img src="imgs/blackpink.jpg"></td>              
  <td><img src="vis_output/blackpink.jpg"></td>
</tr>
<tr>
  <td width=25% style="text-align:center;color:gray;">"Can you segment <ins>the founder of Alibaba</ins> in this image and explain why?"</td>
  <td width=25% style="text-align:center;">"Sure, it is [SEG]. In the image, the man wearing a red tie and sitting on one of the chairs is <ins>Jack Ma</ins>, the co-founder of Alibaba Group..."</td>
  <td width=25% style="text-align:center;color:gray">"Please segment <ins>Lisa</ins> in this figure."</td>
  <td width=25% style="text-align:center;">"Sure, [SEG]."</td>
</tr>
</table>

<p align="center"> <img src="imgs/fig_overview.jpg" width="100%"> </p>

## News
- [x] [2023.8.30] Release three new models [LISA-7B-v1](https://huggingface.co/xinlai/LISA-7B-v1), [LISA-7B-v1-explanatory](https://huggingface.co/xinlai/LISA-7B-v1-explanatory), and [LISA-13B-llama2-v1-explanatory](https://huggingface.co/xinlai/LISA-13B-llama2-v1-explanatory). Welcome to check them out!
- [x] [2023.8.23] Refactor code, and release new model [LISA-13B-llama2-v1](https://huggingface.co/xinlai/LISA-13B-llama2-v1). Welcome to check it out!
- [x] [2023.8.9] Training code is released!
- [x] [2023.8.4] [Online Demo](http://103.170.5.190:7860/) is released! 
- [x] [2023.8.4] [*ReasonSeg* Dataset](https://drive.google.com/drive/folders/125mewyg5Ao6tZ3ZdJ-1-E3n04LGVELqy?usp=sharing) and the [LISA-13B-llama2-v0-explanatory](https://huggingface.co/xinlai/LISA-13B-llama2-v0-explanatory) model are released! 
- [x] [2023.8.3] Inference code and the [LISA-13B-llama2-v0](https://huggingface.co/xinlai/LISA-13B-llama2-v0) model are released. Welcome to check them out!
- [x] [2023.8.2] [Paper](https://arxiv.org/pdf/2308.00692.pdf) is released and GitHub repo is created.

**LISA: Reasoning Segmentation via Large Language Model [[Paper](https://arxiv.org/abs/2308.00692)]** <br />
[Xin Lai](https://scholar.google.com/citations?user=tqNDPA4AAAAJ&hl=zh-CN),
[Zhuotao Tian](https://scholar.google.com/citations?user=mEjhz-IAAAAJ&hl=en),
[Yukang Chen](https://scholar.google.com/citations?user=6p0ygKUAAAAJ&hl=en),
[Yanwei Li](https://scholar.google.com/citations?user=I-UCPPcAAAAJ&hl=zh-CN),
[Yuhui Yuan](https://scholar.google.com/citations?user=PzyvzksAAAAJ&hl=en),
[Shu Liu](https://scholar.google.com.hk/citations?user=BUEDUFkAAAAJ&hl=zh-CN),
[Jiaya Jia](https://scholar.google.com/citations?user=XPAkzTEAAAAJ&hl=en)<br />

## Abstract
In this work, we propose a new segmentation task --- ***reasoning segmentation***. The task is designed to output a segmentation mask given a complex and implicit query text. We establish a benchmark comprising over one thousand image-instruction pairs, incorporating intricate reasoning and world knowledge for evaluation purposes. Finally, we present LISA: Large-language Instructed Segmentation Assistant, which inherits the language generation capabilities of the multi-modal Large Language Model (LLM) while also possessing the ability to produce segmentation masks.
For more details, please refer to the [paper](https://arxiv.org/abs/2308.00692).

## Highlights
**LISA** unlocks the new segmentation capabilities of multi-modal LLMs, and can handle cases involving: 
1. complex reasoning; 
2. world knowledge; 
3. explanatory answers; 
4. multi-turn conversation. 

**LISA** also demonstrates robust zero-shot capability when trained exclusively on reasoning-free datasets. In addition, fine-tuning the model with merely 239 reasoning segmentation image-instruction pairs results in further performance enhancement.

## Experimental results
<p align="center"> <img src="imgs/table1.jpg" width="80%"> </p>

## Installation
```
pip install -r requirements.txt
pip install flash-attn --no-build-isolation
```

## Training
### Training Data Preparation
The training data consists of 4 types of data:

1. Semantic segmentation datasets: [ADE20K](http://data.csail.mit.edu/places/ADEchallenge/ADEChallengeData2016.zip), [COCO-Stuff](http://calvin.inf.ed.ac.uk/wp-content/uploads/data/cocostuffdataset/stuffthingmaps_trainval2017.zip), [Mapillary](https://www.mapillary.com/dataset/vistas), [PACO-LVIS](https://github.com/facebookresearch/paco/tree/main#dataset-setup), [PASCAL-Part](https://github.com/facebookresearch/VLPart/tree/main/datasets#pascal-part), [COCO Images](http://images.cocodataset.org/zips/train2017.zip)

    Note: For COCO-Stuff, we use the annotation file stuffthingmaps_trainval2017.zip. We only use the PACO-LVIS part in PACO. COCO Images should be put into the `dataset/coco/` directory.

3. Referring segmentation datasets: [refCOCO](https://web.archive.org/web/20220413011718/https://bvisionweb1.cs.unc.edu/licheng/referit/data/refcoco.zip), [refCOCO+](https://web.archive.org/web/20220413011656/https://bvisionweb1.cs.unc.edu/licheng/referit/data/refcoco+.zip), [refCOCOg](https://web.archive.org/web/20220413012904/https://bvisionweb1.cs.unc.edu/licheng/referit/data/refcocog.zip), [refCLEF](https://web.archive.org/web/20220413011817/https://bvisionweb1.cs.unc.edu/licheng/referit/data/refclef.zip) ([saiapr_tc-12](https://web.archive.org/web/20220515000000/http://bvisionweb1.cs.unc.edu/licheng/referit/data/images/saiapr_tc-12.zip)) 

    Note: the original links of refCOCO series data are down, and we update them with new ones. If the download speed is super slow or unstable, we also provide a [OneDrive link](https://mycuhk-my.sharepoint.com/:f:/g/personal/1155154502_link_cuhk_edu_hk/Em5yELVBvfREodKC94nOFLoBLro_LPxsOxNV44PHRWgLcA?e=zQPjsc) to download. **You must also follow the rules that the original datasets require.**

4. Visual Question Answering dataset: [LLaVA-Instruct-150k](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K/blob/main/llava_instruct_150k.json)

5. Reasoning segmentation dataset: [ReasonSeg](https://github.com/dvlab-research/LISA#dataset)

Download them from the above links, and organize them as follows.

```
├── dataset
│   ├── ade20k
│   │   ├── annotations
│   │   └── images
│   ├── coco
│   │   └── train2017
│   │       ├── 000000000009.jpg
│   │       └── ...
│   ├── cocostuff
│   │   └── train2017
│   │       ├── 000000000009.png
│   │       └── ...
│   ├── llava_dataset
│   │   └── llava_instruct_150k.json
│   ├── mapillary
│   │   ├── config_v2.0.json
│   │   ├── testing
│   │   ├── training
│   │   └── validation
│   ├── reason_seg
│   │   └── ReasonSeg
│   │       ├── train
│   │       ├── val
│   │       └── explanatory
│   ├── refer_seg
│   │   ├── images
│   │   |   ├── saiapr_tc-12 
│   │   |   └── mscoco
│   │   |       └── images
│   │   |           └── train2014
│   │   ├── refclef
│   │   ├── refcoco
│   │   ├── refcoco+
│   │   └── refcocog
│   └── vlpart
│       ├── paco
│       │   └── annotations
│       └── pascal_part
│           ├── train.json
│           └── VOCdevkit
```

### Pre-trained weights

#### LLaVA
To train LISA-7B or 13B, you need to follow the [instruction](https://github.com/haotian-liu/LLaVA/blob/main/docs/MODEL_ZOO.md) to merge the LLaVA delta weights. Typically, we use the final weights `LLaVA-Lightning-7B-v1-1` and `LLaVA-13B-v1-1` merged from `liuhaotian/LLaVA-Lightning-7B-delta-v1-1` and `liuhaotian/LLaVA-13b-delta-v1-1`, respectively. For Llama2, we can directly use the LLaVA full weights `liuhaotian/llava-llama-2-13b-chat-lightning-preview`.

#### SAM ViT-H weights
Download SAM ViT-H pre-trained weights from the [link](https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth).

### Training
```
deepspeed --master_port=24999 train_ds.py \
  --version="PATH_TO_LLaVA" \
  --dataset_dir='./dataset' \
  --vision_pretrained="PATH_TO_SAM" \
  --dataset="sem_seg||refer_seg||vqa||reason_seg" \
  --sample_rates="9,3,3,1" \
  --exp_name="lisa-7b"
```
When training is finished, to get the full model weight:
```
cd ./runs/lisa-7b/ckpt_model && python zero_to_fp32.py . ../pytorch_model.bin
```

### Merge LoRA Weight
Merge the LoRA weights of `pytorch_model.bin`, save the resulting model into your desired path in the Hugging Face format:
```
CUDA_VISIBLE_DEVICES="" python merge_lora_weights_and_save_hf_model.py \
  --version="PATH_TO_LLaVA" \
  --weight="PATH_TO_pytorch_model.bin" \
  --save_path="PATH_TO_SAVED_MODEL"
```

For example:
```
CUDA_VISIBLE_DEVICES="" python3 merge_lora_weights_and_save_hf_model.py \
  --version="./LLaVA/LLaVA-Lightning-7B-v1-1" \
  --weight="lisa-7b/pytorch_model.bin" \
  --save_path="./LISA-7B"
```

### Validation
```
deepspeed --master_port=24999 train_ds.py \
  --version="PATH_TO_LISA_HF_Model_Directory" \
  --dataset_dir='./dataset' \
  --vision_pretrained="PATH_TO_SAM" \
  --exp_name="lisa-7b" \
  --eval_only
```

Note: the `v1` model is trained using both `train+val` sets, so please use the `v0` model to reproduce the validation results. (To use the `v0` models, please first checkout to the legacy version repo with `git checkout 0e26916`.)

 
## Inference 

To chat with [LISA-13B-llama2-v1](https://huggingface.co/xinlai/LISA-13B-llama2-v1) or [LISA-13B-llama2-v1-explanatory](https://huggingface.co/xinlai/LISA-13B-llama2-v1-explanatory):
(Note that `chat.py` currently does not support `v0` models (i.e., `LISA-13B-llama2-v0` and `LISA-13B-llama2-v0-explanatory`), if you want to use the `v0` models, please first checkout to the legacy version repo `git checkout 0e26916`.)
```
CUDA_VISIBLE_DEVICES=0 python chat.py --version='xinlai/LISA-13B-llama2-v1'
CUDA_VISIBLE_DEVICES=0 python chat.py --version='xinlai/LISA-13B-llama2-v1-explanatory'
```
To use `bf16` or `fp16` data type for inference:
```
CUDA_VISIBLE_DEVICES=0 python chat.py --version='xinlai/LISA-13B-llama2-v1' --precision='bf16'
```
To use `8bit` or `4bit` data type for inference (this enables running 13B model on a single 24G or 12G GPU at some cost of generation quality):
```
CUDA_VISIBLE_DEVICES=0 python chat.py --version='xinlai/LISA-13B-llama2-v1' --precision='fp16' --load_in_8bit
CUDA_VISIBLE_DEVICES=0 python chat.py --version='xinlai/LISA-13B-llama2-v1' --precision='fp16' --load_in_4bit
```
Hint: for 13B model, 16-bit inference consumes 30G VRAM with a single GPU, 8-bit inference consumes 16G, and 4-bit inference consumes 9G.

After that, input the text prompt and then the image path. For example，
```
- Please input your prompt: Where can the driver see the car speed in this image? Please output segmentation mask.
- Please input the image path: imgs/example1.jpg

- Please input your prompt: Can you segment the food that tastes spicy and hot?
- Please input the image path: imgs/example2.jpg
```
The results should be like:
<p align="center"> <img src="imgs/example1.jpg" width="22%"> <img src="vis_output/example1_masked_img_0.jpg" width="22%"> <img src="imgs/example2.jpg" width="25%"> <img src="vis_output/example2_masked_img_0.jpg" width="25%"> </p>

## Deployment
```
CUDA_VISIBLE_DEVICES=0 python app.py --version='xinlai/LISA-13B-llama2-v1 --load_in_4bit'
CUDA_VISIBLE_DEVICES=0 python app.py --version='xinlai/LISA-13B-llama2-v1-explanatory --load_in_4bit'
```
By default, we use 4-bit quantization. Feel free to delete the `--load_in_4bit` argument for 16-bit inference or replace it with `--load_in_8bit` argument for 8-bit inference.


## Dataset
In ReasonSeg, we have collected 1218 images (239 train, 200 val, and 779 test). The training and validation sets can be download from <a href="https://drive.google.com/drive/folders/125mewyg5Ao6tZ3ZdJ-1-E3n04LGVELqy?usp=sharing">**this link**</a>. 

Each image is provided with an annotation JSON file:
```
image_1.jpg, image_1.json
image_2.jpg, image_2.json
...
image_n.jpg, image_n.json
```
Important keys contained in JSON files:
```
- "text": text instructions.
- "is_sentence": whether the text instructions are long sentences.
- "shapes": target polygons.
```

The elements of the "shapes" exhibit two categories, namely **"target"** and **"ignore"**. The former category is indispensable for evaluation, while the latter category denotes the ambiguous region and hence disregarded during the evaluation process. 

We provide a <a href="https://github.com/dvlab-research/LISA/blob/main/utils/data_processing.py">**script**</a> that demonstrates how to process the annotations:
```
python3 utils/data_processing.py
```

Besides, we leveraged GPT-3.5 for rephrasing instructions, so images in the training set may have **more than one instructions (but fewer than six)** in the "text" field. During training, users may randomly select one as the text query to obtain a better model.


## Citation 
If you find this project useful in your research, please consider citing:

```
@article{lai2023lisa,
  title={LISA: Reasoning Segmentation via Large Language Model},
  author={Lai, Xin and Tian, Zhuotao and Chen, Yukang and Li, Yanwei and Yuan, Yuhui and Liu, Shu and Jia, Jiaya},
  journal={arXiv preprint arXiv:2308.00692},
  year={2023}
}
@article{yang2023improved,
  title={An Improved Baseline for Reasoning Segmentation with Large Language Model},
  author={Yang, Senqiao and Qu, Tianyuan and Lai, Xin and Tian, Zhuotao and Peng, Bohao and Liu, Shu and Jia, Jiaya},
  journal={arXiv preprint arXiv:2312.17240},
  year={2023}
}
```

## Acknowledgement
-  This work is built upon the [LLaVA](https://github.com/haotian-liu/LLaVA) and [SAM](https://github.com/facebookresearch/segment-anything).