Chronos-t5-small Open-source Time Series Forecasting Model - Free Deployment to Facilitate Probabilistic Prediction Tasks

Chronos T5 Small

Developed by amazon

Chronos is a family of pre-trained time series forecasting models based on language model architectures. It converts time series into token sequences through quantization and scaling for training, suitable for probabilistic forecasting tasks.

Climate Model

Transformers

Open Source License:Apache-2.0 #Time Series Forecasting #Probabilistic Forecasting #Language Model Architecture

Downloads 22.8M

Release Time : 2/21/2024

Model Overview

Chronos-T5 is a pre-trained model specifically designed for time series forecasting. It uses an improved version of the T5 architecture, training on time series data converted into token sequences with cross-entropy loss to generate probabilistic predictions.

Model Features

Probabilistic Forecasting

Generates future trajectories through multiple sampling, providing probabilistic predictions rather than single-point estimates.

Pre-trained Foundation Model

Pre-trained on extensive public time series data and synthetic data, offering broad applicability.

Efficient Architecture

Utilizes an improved T5 architecture with a smaller vocabulary (4096 tokens) for higher parameter efficiency.

Multi-Scale Support

Adapts to time series data of varying scales through quantization and scaling mechanisms.

Model Capabilities

Time Series Forecasting

Probabilistic Forecasting

Multi-Step Forecasting

Handling Non-Stationary Time Series

Use Cases

Business Forecasting

Sales Forecasting

Predict future product sales volumes.

Generates predictions with confidence intervals.

Demand Forecasting

Forecast future demand for products or services.

Economic Indicator Forecasting

Air Passenger Forecasting

Predict future air passenger numbers.

Example shows 12-month forecast results.

🚀 Chronos-T5 (Small)

Chronos is a family of pretrained time series forecasting models. It transforms time - series data into tokens, trains a language model on these tokens, and then generates probabilistic forecasts. This model has been trained on a large amount of public time - series data and synthetic data.

🚀 Quick Start

Updates

🚀 Feb 14, 2025: Chronos - Bolt & original Chronos models are now available on Amazon SageMaker JumpStart! Check out the tutorial notebook to learn how to deploy Chronos endpoints for production use in a few lines of code.
🚀 Nov 27, 2024: We have released Chronos - Bolt⚡️ models that are more accurate (5% lower error), up to 250 times faster and 20 times more memory - efficient than the original Chronos models of the same size. Check out the new models here.

Chronos models transform input time series into token sequences through scaling and quantization. Then, a language model is trained using the cross - entropy loss. After training, probabilistic forecasts are obtained by sampling multiple future trajectories based on historical context. For more details, please refer to the paper Chronos: Learning the Language of Time Series.

Fig. 1: High - level depiction of Chronos. (Left) The input time series is scaled and quantized to obtain a sequence of tokens. (Center) The tokens are fed into a language model which may either be an encoder - decoder or a decoder - only model. The model is trained using the cross - entropy loss. (Right) During inference, we autoregressively sample tokens from the model and map them back to numerical values. Multiple trajectories are sampled to obtain a predictive distribution.

✨ Features

Chronos models are based on language model architectures. They are trained on a large corpus of publicly available time series data and synthetic data. This enables them to provide probabilistic forecasts by sampling multiple future trajectories given historical context.

📦 Installation

To perform inference with Chronos models, install the package in the GitHub companion repo by running:

pip install git+https://github.com/amazon-science/chronos-forecasting.git

💻 Usage Examples

Basic Usage

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import torch
from chronos import ChronosPipeline

pipeline = ChronosPipeline.from_pretrained(
  "amazon/chronos-t5-small",
  device_map="cuda",
  torch_dtype=torch.bfloat16,
)

df = pd.read_csv("https://raw.githubusercontent.com/AileenNielsen/TimeSeriesAnalysisWithPython/master/data/AirPassengers.csv")

# context must be either a 1D tensor, a list of 1D tensors,
# or a left-padded 2D tensor with batch as the first dimension
context = torch.tensor(df["#Passengers"])
prediction_length = 12
forecast = pipeline.predict(context, prediction_length)  # shape [num_series, num_samples, prediction_length]

# visualize the forecast
forecast_index = range(len(df), len(df) + prediction_length)
low, median, high = np.quantile(forecast[0].numpy(), [0.1, 0.5, 0.9], axis=0)

plt.figure(figsize=(8, 4))
plt.plot(df["#Passengers"], color="royalblue", label="historical data")
plt.plot(forecast_index, median, color="tomato", label="median forecast")
plt.fill_between(forecast_index, low, high, color="tomato", alpha=0.3, label="80% prediction interval")
plt.legend()
plt.grid()
plt.show()

📚 Documentation

Architecture

The models in this repository are based on the T5 architecture. The only difference is in the vocabulary size: Chronos - T5 models use 4096 different tokens, compared to 32128 of the original T5 models, resulting in fewer parameters.

Property	Details
Model Type	The models in this repository are based on the T5 architecture.
Training Data	Chronos models have been trained on a large corpus of publicly available time series data, as well as synthetic data generated using Gaussian processes.

Model	Parameters	Based on
chronos-t5-tiny	8M	t5-efficient-tiny
chronos-t5-mini	20M	t5-efficient-mini
chronos-t5-small	46M	t5-efficient-small
chronos-t5-base	200M	t5-efficient-base
chronos-t5-large	710M	t5-efficient-large

📄 License

This project is licensed under the Apache - 2.0 License.

📚 Citation

If you find Chronos models useful for your research, please consider citing the associated paper:

@article{ansari2024chronos,
    title={Chronos: Learning the Language of Time Series},
    author={Ansari, Abdul Fatir and Stella, Lorenzo and Turkmen, Caner and Zhang, Xiyuan, and Mercado, Pedro and Shen, Huibin and Shchur, Oleksandr and Rangapuram, Syama Syndar and Pineda Arango, Sebastian and Kapoor, Shubham and Zschiegner, Jasper and Maddix, Danielle C. and Mahoney, Michael W. and Torkkola, Kari and Gordon Wilson, Andrew and Bohlke - Schneider, Michael and Wang, Yuyang},
    journal={Transactions on Machine Learning Research},
    issn={2835-8856},
    year={2024},
    url={https://openreview.net/forum?id=gerNCVqqtR}
}

🔧 Technical Details

For details on Chronos models, training data and procedures, and experimental results, please refer to the paper Chronos: Learning the Language of Time Series.

🔒 Security

See CONTRIBUTING for more information.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご