ibm-granite/granite-timeseries-ttm-r3

Granite-TTM-R3 Model Card

Building on top of TTM-R1 and TTM-R2, we introduce the next generation of TinyTimeMixer under the Granite time-series foundation model family — Granite-TTM-R3. This release incorporates several novel tiny-neural architectural innovations designed to push the limits of accuracy in high-speed forecasting, a critical requirement for real-world production deployments.

---

Supported Tasks & Capabilities

Granite-TTM-R3 is a family of pretrained models supporting multiple real-world forecasting scenarios:

• Zero-shot forecasting across unseen datasets
• Few-shot adaptation effective with as few as ~1K samples
• Full fine-tuning for domain-specific optimization
• Multivariate time-series forecasting
• Exogenous / control variable integration
• High-throughput batch inference for production systems

---

Architecture Overview

Granite-TTM adopts a mixture-of-experts paradigm composed of models with varying complexities [1-35M and Lite: 1-18M parameters] — coupled with a lightweight routing mechanism that automatically selects or blends the most suitable expert based on input data characteristics.

This enables adaptive model selection, improving both accuracy and efficiency across diverse time-series scenarios.

The architecture is built on efficient mixer-based designs that avoid expensive self-attention. Instead, Granite-TTM-R3 leverages linear gating-based attention mechanisms to capture temporal dependencies with significantly lower computational overhead.

This combination allows Granite-TTM-R3 to deliver scalable, adaptive, and ultra-fast forecasting performance suitable for real-time and large-scale deployments.

---

Accuracy & Speed

Granite-TTM-R3 is engineered to achieve a strong balance between state-of-the-art accuracy and extreme inference efficiency, making it well-suited for real-world, high-throughput deployments.

Accuracy on GIFT-Eval

• Maintains top-tier performance on the GIFT-Eval leaderboard.
• Fine-tuned FM:
  • MASE: 0.718 | CRPS: 0.514
  • (Lite) MASE: 0.719 | CRPS: 0.514
• Pre-trained FM:
  • MASE: 0.727 | CRPS: 0.523
  • (Lite) MASE: 0.733 | CRPS: 0.524

---

Inference Throughput

Granite-TTM-R3 delivers orders-of-magnitude faster inference compared to existing popular SOTA models.

🖥️ GPU Throughput

• Typical SOTA models: ~20–500 samples/sec
• Granite-TTM-R3: ~7,500 samples/sec
• Granite-TTM-R3 Lite: ~18,000 samples/sec

💻 CPU Throughput

• Typical SOTA models: ~1–20 samples/sec
• Granite-TTM-R3: ~180 samples/sec
• Granite-TTM-R3 Lite: ~800 samples/sec

👉 Granite-TTM-R3 models achieves ~15–50× speedup over many existing approaches, without compromising accuracy, setting a new benchmark for fast and reliable time-series forecasting.

---

What’s New in Granite-TTM-R3

• Trend–Residual Decomposition
  Separately models long-term trends and high-frequency residuals for improved structural learning.

• Three-Stage Pre-Training
  Sequential training of trend, residual, and joint components for better stability and convergence.

• Student-Teacher Pretraining
  Student-teacher based pretraining for stable learning in noisy datasets.

• Enhanced Data Augmentation
  Structured perturbations improve robustness across domains.

• Improved Normalization Strategy
  Stabilizes training across scale shifts and heterogeneous datasets.

• GLU Gating
  Dynamic information flow control within mixer blocks.

• Multi-Resolution Temporal Layer
  Captures dependencies across short-, medium-, and long-term horizons.

• FFT-Based Embeddings
  Incorporates frequency-domain signals to model periodicity and seasonality.

• Register Tokens
  Learnable global tokens to encode sequence-level semantics.

• Multi-Quantile Forecasting Head
  Enables probabilistic forecasting with multiple quantiles.

• Refined Loss Weighting
  Balances trend, residual, and quantile objectives for improved calibration.

---

⚡ Why Granite-TTM-R3

Granite-TTM-R3 is purpose-built for production-grade time-series systems where:

• Low latency is critical for real-time inference
• High scale is required for millions of forecasts
• Data is heterogeneous and evolving
• Compute constraints exist, including CPU-friendly deployment

It provides a practical combination of compactness, speed, and accuracy, making it suitable for industrial deployment.

---

Example Recipes and Notebooks

To be released.

---

Publication

• Granite-TTM-R3 Paper: To be released.
• Base TTM Paper: Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series

---

Training Data

• Select Datasets from GiftEvalPretrain and GIFT-Eval Train, using non-leaking historical context.
• Custom synthesized data based on KernelSynth.

---

Model Card Authors

Vijay Ekambaram, Arindam Jati, Haoxiang Qiu, Takayuki Katsuki, Tomoya Sakai, Priyanshul Govil, Pankaj Dayama

---

Citation

Please cite the following paper if you intend to use this model or its associated architectures and approaches in your work.

@inproceedings{ekambaram2024tinytimemixersttms,
  title={Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series},
  author={Vijay Ekambaram and Arindam Jati and Pankaj Dayama and Sumanta Mukherjee and Nam H. Nguyen and Wesley M. Gifford and Chandra Reddy and Jayant Kalagnanam},
  booktitle={Advances in Neural Information Processing Systems (NeurIPS 2024)},
  year={2024},
}