DEDUPEDplatform. TimesFM 3.0 pretrained weights are under a non-commercial license and the README says commercial or production use of self-hosted weights is not permitted.
Who it is forAn analyst who needs a forecast today and has no model.
The problemFitting a model per series is slow and needs tuning.
The outcomePoint forecasts and nine quantiles for each series you pass.
What you need first
- Python with pip, and `pip install timesfm[torch]`, or `pip install timesfm[mlx]` on Apple silicon
- A CUDA GPU for the PyTorch examples, which set device to cuda
- Your series as NumPy arrays
Steps
- Install it with
pip install timesfm[torch] - Create a ModelConfig with checkpoint_path set to google/timesfm-3.0-pytorch and build a TimesFM3Evaluator as in the README
- Call forecaster.predict_batch with your list of 1D NumPy arrays, a horizon and return_quantiles=True
- Read outputs[0].forecast and outputs[0].quantiles, which have shapes like (12,) and (12, 9)
Prompt for your agent
Using google-research/timesfm, install `pip install timesfm[torch]`, forecast the next 12 steps of each series in <my CSV> with TimesFM3Evaluator and return_quantiles, and write the medians and deciles to a new CSV.
Caveat. Weights up to version 2.5 remain Apache 2.0 if you need a permissive option.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for google-research/timesfm and "Forecast a series without training anything"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer