KO | EN

GitLite

Minimalist Open Source Repository Search Engine

gitlite — search

#dsh #dsh-plugin #deepseek-harness #typescript #ai-agent #codex #developer-tools #claude-code #windows #mcp #python #desktop-app

Preformer

★ 46 Open GitHub ↗

This repository contains the pytorch code for the 2023 ICASSP paper "Preformer: Predictive Transformer with Multi-Scale Segment-wise Correlations for Long-Term Time Series Forecasting”

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Preformer

This repository contains the pytorch code for the 2023 ICASSP paper "[Preformer: Predictive Transformer with Multi-Scale Segment-wise Correlations for Long-Term Time Series Forecasting](https://arxiv.org/pdf/2202.11356.pdf)”.

# Model
The core MSSC module:

![Main Results](https://raw.githubusercontent.com/ddz16/Preformer/main/figs/mssc.png)

# Acknowledgment

This repository uses some code from [Autoformer](https://github.com/thuml/Autoformer) and [Informer](https://github.com/zhouhaoyi/Informer2020). Thanks to the authors for their work!

# Get Started
## Environment

Install Python 3.6, PyTorch 1.9.0.

## Data

We use all the datasets provided by [Autoformer](https://github.com/thuml/Autoformer) directly, so you can download them from [Google drive](https://drive.google.com/drive/folders/1ZOYpTUa82_jCcxIdTmyr0LXQfvaM9vIy) provided in [Autoformer](https://github.com/thuml/Autoformer). All the datasets are well pre-processed and can be used easily. After downloading, put these dataset files (such as ETTh1.csv) in the `./data/` folder.

# Train

```
python run.py --is_training 1 --root_path ./data/ --data_path ETTh1.csv --model_id ETTh1_96_48 --model Preformer --data ETTh1 --features M --seq_len 96 --label_len 48 --pred_len 48 --e_layers 2 --d_layers 1 --factor 4 --enc_in 7 --dec_in 7 --c_out 7 --des 'Exp' --itr 1 --n_heads 4 --d_model 32 --d_ff 128
```

# Test

```
python run.py --is_training 0 --root_path ./data/ --data_path ETTh1.csv --model_id ETTh1_96_48 --model Preformer --data ETTh1 --features M --seq_len 96 --label_len 48 --pred_len 48 --e_layers 2 --d_layers 1 --factor 4 --enc_in 7 --dec_in 7 --c_out 7 --des 'Exp' --itr 1 --n_heads 4 --d_model 32 --d_ff 128
```

# Main Results

![Main Results](https://raw.githubusercontent.com/ddz16/Preformer/main/figs/result.png)

# Citation

```
@inproceedings{du2022preformer,
  title={Preformer: Predictive Transformer with Multi-Scale Segment-wise Correlations for Long-Term Time Series Forecasting},
  author={Du, Dazhao and Su, Bing and Wei, Zhewei},
  booktitle={ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  year={2023},
  organization={IEEE}
}
```

Preformer

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README