README.md

---
library_name: granite_tsfm
base_model: ibm-granite/granite-timeseries-ttm-r2
tags:
- ttm4hvac
- tsfm
- digital twin
- hvac
- energy
- experiment
license: apache-2.0
papers:
- title: "Transfer learning of building dynamics digital twin for HVAC control with Time-series Foundation Model"
  url: https://arxiv.org/abs/XXXX.XXXXX
  authors: "Ferran Aran Domingo"
datasets:
- gft/ttm4hvac-source-default-train
- gft/ttm4hvac-target-heat-test
- gft/ttm4hvac-target-cool-test
pipeline_tag: time-series-forecasting
---
# TTM4HVAC –  TinyTimeMixer for HVAC dynamics modeling 

This repository contains the **TTM4HVAC – source-default** fine-tuned TinyTimeMixer model.

It corresponds to the **“source-default” experiment** described in the TTM4HVAC paper, where the model is trained only on **default rule-based control (RBC)** trajectories across all simulated source buildings and climates.

This model is part of the TTM4HVAC family, but **is not the main recommended checkpoint**.  
For full documentation, usage examples, and the best-performing model, please visit:

👉 **Main model card:** https://huggingface.co/gft/ttm4hvac