DeepSeq3 is a propsed framework for sequencial circuit analysis.
This repository provides a quick start workflow: simply prepare your data and run the provided script.

This project requires Python 3.8+ and CUDA 11.8 for GPU acceleration (if using PyTorch with GPU).

git clone https://github.com/zhoujy22/DeepSeq3.git
cd <your-repo-directory>

All required Python packages are listed in requirements.txt. Install them using:

pip install --upgrade pip
pip install -r requirements.txt

You can use raw_data/example/example.bench as an example to prepare your data For comgraph experiments

python src/prepare_dataset_comgraph.py

For SNG experiments

python src/prepare_dataset_sng.py

After installing dependencies and preparing your data in data/train_stage1 and data/train_stage2, you can quickly run the experiments:

The main startup script ds3_stage1.sh will run the stage_1 default experiment, and ds3_stage2.sh is for the SNG phase:

bash run/ds3_stage1.sh
bash run/ds3_stage2.sh

This will automatically load datasets from data/ and start training the default model.