# Timesfm Forecast

> Usa a pronosticar series temporales con TimesFM 3.0.

- Skill: `ntizar/timesfm-forecast` (Agent Skill)
- Install (CLI): `npx skillmds@latest add ntizar/timesfm-forecast`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ntizar/timesfm-forecast/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: ntizar (https://skillmd.com/u/ntizar)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/ntizar/timesfm-forecast

---


# TimesFM — forecasting de series temporales (API 3.0)

> ⚠️ Corrección 2026-09-05 (auditoría): la v1 usaba `timesfm.TimesFm`/`model.forecast()` (TimesFM v1/v2). La versión actual **3.0** usa `from timesfm3 import TimesFM3Evaluator, ModelConfig` y `forecaster.predict_batch(...)`. Install real: `pip install timesfm[torch]`.

**Repo:** `https://github.com/google-research/timesfm` (Python, ~31K⭐).

## When to Use

- Cuando pidas **pronosticar una serie temporal** (modelo fundacional de Google) para tus datos.

## Uso (API 3.0)

```bash
pip install "timesfm[torch]"
```

```python
from timesfm3 import TimesFM3Evaluator, ModelConfig
model = TimesFM3Evaluator(config=ModelConfig(...))
forecasts = model.predict_batch(inputs, horizon=..., return_quantiles=True)
```

## Pitfalls

- **Install** `pip install "timesfm[torch]"`, no `pip install timesfm`.
- API **3.0**: `TimesFM3Evaluator`/`predict_batch`; no `TimesFm/forecast/forecast_with_quantiles` (v1/v2).

## Verificación

- `predict_batch(inputs, horizon=N)` y comprobar el forecast + quantiles.