Poster Generation User Guide

February 24, 2026 · View on GitHub

Poster Generation User Guide

This project aims to automatically generate academic posters from research papers. The main workflow includes: data preprocessing → bullet point generation → poster generation.

1. Environment Setup

Install Dependencies

Install Python dependencies:

pip install -r requirements.txt

Install LibreOffice (required for document conversion):

sudo apt install libreoffice

2. Data Preparation

1. Download and Process Dataset

Download the Paper2Poster dataset and process it using MinerU.

After processing, the directory structure should be:

paper2poster/
└── parsed_papers/
    └── {paper_id}/
        ├── {paper_id}.pdf
        └── poster.png

Where:

{paper_id}.pdf: Original research paper
poster.png: Corresponding reference poster image

2. Prepare Assets

Download the following directory from the original Paper2Poster repository:

assets/poster_data

Copy it into this project:

paper2poster/assets/poster_data

Make sure the directory structure is consistent, otherwise template or style loading may fail.

3. Bullet Point Generation

Bullet points summarize key information from papers and serve as intermediate results for poster generation.

1. Configure API Key

This project uses Alibaba Cloud DashScope API by default. Please set the environment variable first:

export DASHSCOPE_API_KEY=your_api_key

Replace your_api_key with your actual API key.

2. Run Bullet Generation Script

Execute the following command:

python gen_bullet.py \
  --summ_provider aliyun \
  --summ_model qwen3-vl-8b-instruct \
  --gpu_id 0 \
  --poster_path /path/to/parsed_papers/{paper_id} \
  --output_root ./bullet_output

Parameter description:

--summ_provider: Model service provider (default: aliyun)
--summ_model: Model name
--gpu_id: GPU device ID
--poster_path: Input paper directory path
--output_root: Output directory for bullet points

Generated results will be saved in:

./bullet_output/

4. Poster Generation

After obtaining bullet points, you can generate the final poster.

Run the following command:

python gen_poster.py \
  --input_dir ./bullet_output \
  --output_dir ./poster_output

Parameter description:

--input_dir: Bullet output directory
--output_dir: Final poster output directory

Results will be saved in:

./poster_output/

5. Complete Workflow Example

Example of a full pipeline:

# 1. Install dependencies
pip install -r requirements.txt
sudo apt install libreoffice

# 2. Set API key
export DASHSCOPE_API_KEY=your_api_key

# 3. Generate bullet points
python gen_bullet.py \
  --summ_provider aliyun \
  --summ_model qwen3-vl-8b-instruct \
  --gpu_id 0 \
  --poster_path ./parsed_papers/xxxx \
  --output_root ./bullet_output

# 4. Generate poster
python gen_poster.py \
  --input_dir ./bullet_output \
  --output_dir ./poster_output

6. Notes

Make sure the parsed_papers directory structure is correct, otherwise the program may fail.
Ensure your API key is valid and has sufficient quota.
If GPU resources are limited, set --gpu_id to -1 to use CPU (if supported).
Linux environment is recommended for better compatibility with LibreOffice and dependencies.

7. Examples

Below are example posters generated by our framework, along with the corresponding paper titles and links.

A Context-Integrated Transformer-Based Neural Network for Auction Design

Paper Link

Flexible Attention-Based Multi-Policy Fusion for Efficient Deep Reinforcement Learning

Paper Link