TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model developed by Google Research for time-series forecasting.

• Paper: A decoder-only foundation model for time-series forecasting, to appear in ICML 2024.
• Google Research blog
• Hugging Face checkpoint repo

This repo contains the code to load public TimesFM checkpoints and run model inference. Please visit our Hugging Face checkpoint repo to download model checkpoints.

This is not an officially supported Google product.

We recommend at least 16GB RAM to load TimesFM dependencies.

• We have released full pytorch support (excluding PEFT parts).
• Shoutout to @tanmayshishodia for checking in PEFT methods like LoRA and DoRA.
• To install TimesFM, you can now simply do: pip install timesfm.
• Launched finetuning support that lets you finetune the weights of the pretrained TimesFM model on your own data.
• Launched ~zero-shot covariate support with external regressors. More details here.

timesfm-1.0-200m is the first open model checkpoint:

• It performs univariate time series forecasting for context lengths up to 512 timepoints and any horizon lengths, with an optional frequency indicator.
• It focuses on point forecasts, and does not support probabilistic forecasts. We experimentally offer quantile heads but they have not been calibrated after pretraining.
• It requires the context to be contiguous (i.e. no "holes"), and the context and the horizon to be of the same frequency.

Please refer to our result tables on the extended benchmarks and the long horizon benchmarks.

Please look into the README files in the respective benchmark directories within experiments/ for instructions for running TimesFM on the respective benchmarks.

We will be using pyenv and poetry. In order to set these things up please follow the instructions here. Note that the PAX (or JAX) version needs to run on python 3.10.x and the PyTorch version can run on >=3.11.x. Therefore make sure you have two versions of python installed:

pyenv install 3.10
pyenv install 3.11
pyenv versions # to list the versions available (lets assume the versions are 3.10.15 and 3.11.10)

pyenv local 3.10.15
poetry env use 3.10.15
poetry lock
poetry install -E  pax

After than you can run the timesfm under poetry shell or do poetry run python3 ....

pyenv local 3.11.10
poetry env use 3.11.10
poetry lock
poetry install -E  torch

After than you can run the timesfm under poetry shell or do poetry run python3 ....

Note:

1. Running the provided benchmarks would require additional dependencies. Please see the experiments section fro more instructions.

2. The dependency lingvo does not support ARM architectures, and the code is not working for machines with Apple silicon. We are aware of this issue and are working on a solution. Stay tuned.

1. Running the provided benchmarks would require additional dependencies. Please see the experiments folder.

2. The dependency lingvo does not support ARM architectures, and the PAX version is not working for machines with Apple silicon.

On python 3.11 you can install the torch version using:

pip install timesfm[torch]

On python 3.10 you can install the pax version using:

pip install timesfm[pax]

Then the base class can be loaded as,

import timesfm

# For PAX
tfm = timesfm.TimesFm(
      hparams=timesfm.TimesFmHparams(
          backend="gpu",
          per_core_batch_size=32,
          horizon_len=128,
      ),
      checkpoint=timesfm.TimesFmCheckpoint(
          huggingface_repo_id="google/timesfm-1.0-200m"),
  )

# For Torch
tfm = timesfm.TimesFm(
      hparams=timesfm.TimesFmHparams(
          backend="gpu",
          per_core_batch_size=32,
          horizon_len=128,
      ),
      checkpoint=timesfm.TimesFmCheckpoint(
          huggingface_repo_id="google/timesfm-1.0-200m-pytorch"),
  )

Note some of the parameters are fixed to load the 200m model

1. The context_len in hparams here can be set as the max context length of the model. It needs to be a multiplier of input_patch_len, i.e. a multiplier of 32. You can provide a shorter series to the tfm.forecast() function and the model will handle it. Currently, the model handles a max context length of 512, which can be increased in later releases. The input time series can have any context length. Padding / truncation will be handled by the inference code if needed.

2. The horizon length can be set to anything. We recommend setting it to the largest horizon length you would need in the forecasting tasks for your application. We generally recommend horizon length <= context length but it is not a requirement in the function call.

3. backend is one of "cpu", "gpu", case sensitive.

We provide APIs to forecast from either array inputs or pandas dataframe. Both forecast methods expect (1) the input time series contexts, (2) along with their frequencies. Please look at the documentation of the functions tfm.forecast() and tfm.forecast_on_df() for detailed instructions.

In particular regarding the frequency, TimesFM expects a categorical indicator valued in {0, 1, 2}:

• 0 (default): high frequency, long horizon time series. We recommend using this for time series up to daily granularity.
• 1: medium frequency time series. We recommend using this for weekly and monthly data.
• 2: low frequency, short horizon time series. We recommend using this for anything beyond monthly, e.g. quarterly or yearly.

This categorical value should be directly provided with the array inputs. For dataframe inputs, we convert the conventional letter coding of frequencies to our expected categories, that

• 0: T, MIN, H, D, B, U
• 1: W, M
• 2: Q, Y

Notice you do NOT have to strictly follow our recommendation here. Although this is our setup during model training and we expect it to offer the best forecast result, you can also view the frequency input as a free parameter and modify it per your specific use case.

Examples:

Array inputs, with the frequencies set to low, medium and high respectively.

import numpy as np
forecast_input = [
    np.sin(np.linspace(0, 20, 100)),
    np.sin(np.linspace(0, 20, 200)),
    np.sin(np.linspace(0, 20, 400)),
]
frequency_input = [0, 1, 2]

point_forecast, experimental_quantile_forecast = tfm.forecast(
    forecast_input,
    freq=frequency_input,
)

pandas dataframe, with the frequency set to "M" monthly.

import pandas as pd

# e.g. input_df is
#       unique_id  ds          y
# 0     T1         1975-12-31  697458.0
# 1     T1         1976-01-31  1187650.0
# 2     T1         1976-02-29  1069690.0
# 3     T1         1976-03-31  1078430.0
# 4     T1         1976-04-30  1059910.0
# ...   ...        ...         ...
# 8175  T99        1986-01-31  602.0
# 8176  T99        1986-02-28  684.0
# 8177  T99        1986-03-31  818.0
# 8178  T99        1986-04-30  836.0
# 8179  T99        1986-05-31  878.0

forecast_df = tfm.forecast_on_df(
    inputs=input_df,
    freq="M",  # monthly
    value_name="y",
    num_jobs=-1,
)

We now have an external regressors library on top of TimesFM that can support static covariates as well as dynamic covariates available in the future. We have an usage example in notebooks/covariates.ipynb.

Let's take a toy example of forecasting sales for a grocery store:

Task: Given the observed the daily sales of this week (7 days), forecast the daily sales of next week (7 days).

Product: ice cream
Daily_sales: [30, 30, 4, 5, 7, 8, 10]
Category: food
Base_price: 1.99
Weekday: [0, 1, 2, 3, 4, 5, 6, 0, 1, 2, 3, 4, 5, 6]
Has_promotion: [Yes, Yes, No, No, No, Yes, Yes, No, No, No, No, No, No, No]
Daily_temperature: [31.0, 24.3, 19.4, 26.2, 24.6, 30.0, 31.1, 32.4, 30.9, 26.0, 25.0, 27.8, 29.5, 31.2]

Product: sunscreen
Daily_sales: [5, 7, 12, 13, 5, 6, 10]
Category: skin product
Base_price: 29.99
Weekday: [0, 1, 2, 3, 4, 5, 6, 0, 1, 2, 3, 4, 5, 6]
Has_promotion: [No, No, Yes, Yes, No, No, No, Yes, Yes, Yes, Yes, Yes, Yes, Yes]
Daily_temperature: [31.0, 24.3, 19.4, 26.2, 24.6, 30.0, 31.1, 32.4, 30.9, 26.0, 25.0, 27.8, 29.5, 31.2]

In this example, besides the Daily_sales, we also have covariates Category, Base_price, Weekday, Has_promotion, Daily_temperature. Let's introduce some concepts:

Static covariates are covariates for each time series.

• In our example, Category is a static categorical covariate,
• Base_price is a static numerical covariates.

Dynamic covariates are covaraites for each time stamps.

• Date / time related features can be usually treated as dynamic covariates.
• In our example, Weekday and Has_promotion are dynamic categorical covariates.
• Daily_temperate is a dynamic numerical covariate.

Notice: Here we make it mandatory that the dynamic covariates need to cover both the forecasting context and horizon. For example, all dynamic covariates in the example have 14 values: the first 7 correspond to the observed 7 days, and the last 7 correspond to the next 7 days.

We can now provide the past data of the two products along with static and dynamic covariates as a batch input to TimesFM and produce forecasts that take into the account the covariates. To learn more, check out the example in notebooks/covariates.ipynb.

We have provided an example of finetuning the model on a new dataset in notebooks/finetuning.ipynb.

If you would like to submit a PR please make sure that you use our formatting style. We use yapf for formatting with the following options,

[style]
based_on_style = google
# Add your custom style rules here
indent_width = 2
spaces_before_comment = 2

Please run yapf --in-place --recursive <filename> on all affected files.