Skip to main content

Time Series Forecasting Tool for Machine Learning Researchers with Nerual Process

Project description

Sibyl: A Time-Seires Forecasting Package

Sibyl is a package for forecasting time series data with PyTorchLightning which is a lightweight PyTorch wrapper for Machine Learning Researchers.

Sibyl is released and maintained by Microsoft Substrate Connectivity Team.

Models

  • Seq to Seq
  • Long-short term memory
  • Neural process
  • Attentive neural process

Installation

Sibyl is avaialbe on PyPi, feel free to get it via pip:

pip install mssibyl

Quick Start

UPDATING ...

import pandas as pd
from mssibyl.models import BaseModel
from mssibyl.trainer import BaseTrainer
from mssibyl.datasets.utils import make_timeline

df = pd.read_csv("./data.csv")
df.columns = ["ds", "y"]

df_with_future = make_timeline(df, freq="H", periods=300)

args = {
    "lookback": 48,
    "forecast_horizon": 1,
    "data": df_with_future,
    "target": "y",
}

sibyl_model = BaseModel(args)
trainer = BaseTrainer(
    max_epochs=100, gpus=1, show_progress_bar=False, early_stop_callback=True
)
trainer.fit(sibyl_model)
preds = sibyl_model.predict(interval=True, include_history=True)

preds.to_csv("res.csv", index=False)

Changelog

Version 0.1 (coming)

  • Initial release

License

Sibyl is licensed under the MIT license

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mssibyl-0.1.1.dev0.tar.gz (1.3 MB view details)

Uploaded Source

File details

Details for the file mssibyl-0.1.1.dev0.tar.gz.

File metadata

  • Download URL: mssibyl-0.1.1.dev0.tar.gz
  • Upload date:
  • Size: 1.3 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.1.1 pkginfo/1.5.0.1 requests/2.23.0 setuptools/46.1.3 requests-toolbelt/0.9.1 tqdm/4.46.0 CPython/3.7.5

File hashes

Hashes for mssibyl-0.1.1.dev0.tar.gz
Algorithm Hash digest
SHA256 aca2d3def418f2a50a4d0c2b6c34c349752a11880b9a087f206c3c444e07d1f2
MD5 ff4cee7771117cd92641d061431bb567
BLAKE2b-256 94c1bde613d6b45c2ae2374fee3762976f92afc2c3ed661910b56ab4fb25b19b

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page