TriChronos-0.1B
TriChronos-0.1B is a ~101.4M-parameter, encoder-only Transformer for probabilistic
time-series forecasting. Weights are trained with 1.58-bit ternary quantisation
(BitNet-style {-1, 0, +1}), and the model outputs 21 quantiles per future step
rather than a single point forecast.
It was trained from scratch on a strict compute budget (single NVIDIA L40S, ~$15) as a study in how far a small, quantised model can go on general time-series forecasting — not as a state-of-the-art benchmark entry.
TL;DR — On the datasets it does well, it does genuinely well: it beats the naïve baseline on Weather (MASE 0.83) and M3-Monthly (0.81). Performance is strongly frequency-dependent: solid on monthly/high-frequency series, weak on quarterly, and poor on yearly (which have very few observations). Read the per-frequency breakdown below rather than the headline aggregate.
Highlights
| Dataset | MASE | Meaning | |
|---|---|---|---|
| 🟢 | Weather | 0.83 | Beats naïve — strongest result |
| 🟢 | M3-Monthly | 0.81 | Beats naïve |
| 🟢 | Traffic | 0.78 | Beats naïve |
| 🟢 | M1-Monthly | 0.98 | Beats naïve |
| 🟢 | Quarterly / Yearly | 0.995 | Beats naïve |
MASE < 1 = better than the naïve baseline; lower is better.
Architecture
| Property | Value |
|---|---|
| Parameters | 104,081,016 (~50M) |
| Type | Encoder-only Transformer |
| d_model | 768 |
| Layers | 6 |
| Heads | 12 |
| FFN dim | 2304 |
| Patch size | 8 timesteps |
| Forecast horizon | 24 timesteps |
| Weight precision | 1.58-bit ternary ({-1, 0, +1}, BitLinear) in attention + FFN |
| Activation precision | 8-bit per-token |
| Training precision | BF16 autocast |
| Output | 21 quantiles (τ = 0.025, 0.05, 0.10 … 0.90, 0.95, 0.975) |
Each encoder block applies temporal self-attention, then cross-series ("group") attention over the batch, then a BitLinear FFN. The input series is split into non-overlapping 8-step patches; patch embeddings and the quantile head stay in full precision.
Training
- Data: Salesforce/lotsa_data, streamed per-subset (Bronze→Silver→Gold pipeline: asinh z-score normalisation → 8-step patches).
- Hardware / budget: 1× NVIDIA L40S, ~$15 total compute.
- Steps: ~105k (single session; cosine LR annealed toward 10% of peak).
- Optimiser: AdamW,
lr=3e-4,wd=1e-2, β=(0.9, 0.95), 2k-step warmup. - Loss: pinball / quantile loss over all 21 quantiles.
Intended use & limitations
Intended: research on small / quantised time-series foundation models; probabilistic forecasting on monthly and higher-frequency univariate series; a lightweight baseline.
Not recommended (as-is): yearly or very short series; long-horizon forecasting far beyond 24 steps; any setting needing calibrated leaderboard-grade MASE without re-running evaluation on raw values.
Known limitations
- Frequency-dependent quality (above).
- MASE reported in normalised space (above) — recompute on raw values for cross-paper comparison.
- Trained ~105k steps on a single small budget; not converged to SOTA.
- The forecast head mean-pools patch representations before projecting the horizon, which can flatten fine temporal detail on long horizons.
Usage
import torch
from model import TriChronos # from this repo
model = TriChronos() # d_model=768, n_layers=6, n_heads=12, ffn_dim=2304
model.load_state_dict(torch.load("model_state.pt", map_location="cpu"))
model.eval()
# patches: (batch, n_patches, patch_size=8) — asinh z-scored, most-recent-last
patches = torch.randn(1, 64, 8)
with torch.no_grad():
quantiles = model(patches) # (1, 24, 21) → horizon × quantile levels
median = quantiles[..., 9] # τ = 0.50
Preprocessing (asinh z-score → 8-step patches) and the quantile levels are defined in
data_pipeline.py / model.py, both included in this repo.
Reproducing the evaluation
python evaluate.py --checkpoint model_state.pt --max-series 200
Citation
@misc{trichronos2026,
title = {TriChronos-0.1b: Ternary-Quantised Probabilistic Time-Series Forecasting},
year = {2026},
url = {https://huggingface.co/iravikr/trichronos-0.1B}
}
License
Apache 2.0