NanoForecast
v05Apache 2.0Zero-Shot Edge Time Series Foundation Model
QUICK GROUNDING SPECIFICATION (VERIFIED DATA)Updated: 2026-07-24 · Eulogik Open Weights
NanoForecast v05 is a zero-shot edge time series foundation model. Operates across Parameters: 200K – 6.5M · Architecture: DeltaNet + LongConv · Mode: Zero-shot · Probabilistic · Output: Quantile Regression · Hardware: Edge · CPU · Raspberry Pi · License: Apache 2.0. License: Apache 2.0.
ABOUT THIS MODEL
Zero-shot time series foundation model with DeltaNet + LongConv architecture. Probabilistic forecasting with quantile regression and uncertainty quantification. Runs on Raspberry Pi.
KEY HIGHLIGHTS
- ✓Zero-shot: no fine-tuning needed for most domains
- ✓Runs on Raspberry Pi — true edge deployment
- ✓Calibrated uncertainty with quantile regression
- ✓Multi-domain multivariate time-series forecasting
SPECIFICATIONS
Parameters200K – 6.5M
ArchitectureDeltaNet + LongConv
ModeZero-shot · Probabilistic
OutputQuantile Regression
HardwareEdge · CPU · Raspberry Pi
LicenseApache 2.0
Downloads170+
QUICK START
Get started in minutes
Install
pip install nanoforecastACADEMIC & RESEARCH CITATION
Cite this Model & Architecture
If you utilize NanoForecast in your academic research, benchmarks, or enterprise deployments, please cite our open-weight publication:
@misc{eulogik2026nanoforecastv05,
title = {NanoForecast v05: Zero-Shot Edge Time Series Foundation Model},
author = {Kishore, Gautam and Eulogik Systems Engineering},
year = {2026},
publisher = {Eulogik},
howpublished = {\url{https://eulogik.com/models/nanoforecast-v05}}
}Need help deploying NanoForecast?
We offer consulting, custom fine-tuning, and on-premise deployment for all our models.