TurkForecast-FM-Chronos2-LoRA-v1

Turkish-vertical LoRA fine-tune of amazon/chronos-2 (120M T5 encoder, Apache-2.0). First publicly-released Apache-2.0 Turkish-vertical time-series foundation model (TSFM) fine-tune.

Distinction from prior Turkish ML work

TurkForecast-FM is the first publicly-released Apache-2.0 Turkish-vertical time-series foundation model (TSFM) fine-tune. We explicitly distinguish from three related but non-overlapping prior works:

1. TabiBERT (arXiv:2512.23065, Boğaziçi University / VNGRS-AI, December 2025) is a Turkish ModernBERT language encoder pretrained from scratch on text — different modality (NLP, not time series).
2. Ertürk et al. (MDPI Applied Sciences 16(6):2760, March 2026) train a supervised LSTM+BLSTM+GRU+Transformer hybrid from scratch on EPİAŞ electricity data — not a TSFM fine-tune (no pre-trained foundation backbone).
3. NIEXCHE/chronos-t5-small-fine-tuned-v1 (Fevzi Kılaş, HuggingFace) is a Chronos-T5 fine-tune on a 15M-row multi-domain proprietary dataset — not claimed as Turkish-vertical.

Özet (TR): TurkForecast-FM, Türkçe dikey alana özgü, açık-kaynak (Apache-2.0) ilk TSFM ince-ayarıdır; yukarıdaki üç çalışma sırasıyla farklı modalite (dil), train-from-scratch paradigma ve çok-alanlı kapsam nedeniyle bu nişin dışındadır.

GIFT-Eval Results (v1.1, 2026-05-19)

Evaluated on the full GIFT-Eval benchmark (97 canonical (dataset, term) entries across 7 domains × 3 forecast horizons):

Metric	v1.1 (current)	v1.0 baseline	Δ%
Mean MASE[0.5]	1.1328	1.1486	−1.38%
Mean wQL	0.1997	0.2054	−2.76%

By forecast term (v1.1)

Term	n	Mean MASE	Mean wQL
short	55	1.1603	0.1819
medium	21	1.0728	0.2201
long	21	1.1207	0.2262

By domain (v1.1)

Domain	n	Mean MASE	Mean wQL
Econ/Fin	6	1.7409	0.0350
Energy	32	0.9734	0.1630
Healthcare	5	2.0393	0.0449
Nature	15	1.1707	0.2267
Sales	4	0.7579	0.5528
Transport	15	0.6901	0.1003
Web/CloudOps	20	1.3573	0.3303

Full per-dataset metrics: all_results.csv.

v1.1 changelog (vs v1.0)

• Q4 per-domain conformal calibration (Tier-2 DR-6 Round 1 LOCK): Variant-B per-quantile temperature scaling + Romano joint-score CQR with empirical-Bayes hierarchical shrinkage (κ=50) on pure-Chronos baseline. Short-term wQL gain across 7 domains; Web/CloudOps −9.2% mean wQL vs v1.0.
• Horizon-conditional CQR refinement (DR-6 Round 2 3/4-agent endorsement): Per-(dataset, horizon, quantile) Bayesian-shrunk delta on top of Q4 v1; modest incremental lift (34 datasets covered, 20 skipped due to insufficient calibration window).
• Dataset name + V4-bucket alias fixes (Final v3 PASS 5/5 gate): SHA1-deterministic pipeline-short matching expanded 26 → 51 canonical rows, eliminating canonical-fallback dilution.
• MASE regression fixed: v1.0-intermediate P6.0 medium/long substitution caused +46% MASE blowup; v1.1 preserves canonical mid/long baseline (matched bitwise to README floor).
• Small-magnitude target override (bizitobs_service/H, solar/10T): Two datasets with sub-0.01 target magnitudes trigger relative wQL explosion in conformal calibration (Mondrian retune diagnosed +350% wQL on bizitobs_service). Both datasets bypass the calibration pipeline and preserve canonical baseline predictions to avoid noise amplification.

Inference Pipeline

The model ships as a 3-stage inference pipeline (not a monolithic adapter):

1. Base + LoRA adapter — amazon/chronos-2 + LoRA r=32, α=64, dropout=0 on all-linear target_modules
2. V4 router with WiSE-FT α-blend (router_v5_with_alphas_and_calibration_v4.json) — per-(domain, freq) bucket selects FT vs ZS vs blend α ∈ {0, 0.25, 0.5, 0.75, 1.0}; MASE-aware adoption rule (Stage 1.A.2): adopt = (wql_delta<0) OR (mase_delta<−3 AND wql_delta<5) over P5.7 v2 baseline
3. Bucket isotonic calibrators (bucket_isotonic_calibrators.pkl) — per-bucket Kuleshov (ICML 2018) PIT-CDF isotonic regression on 9 quantile levels, with per-domain fallback for low-cardinality buckets

Quick start

from chronos import BaseChronosPipeline

pipe = BaseChronosPipeline.from_pretrained(
    "Verm1ion/turkforecast-fm-chronos2-lora-v1",
    device_map="cuda", torch_dtype="bfloat16",
)
pipe.model.eval()

quantiles, _ = pipe.predict_quantiles(
    inputs=[your_context_array],
    prediction_length=48,
    quantile_levels=[0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9],
)

For full GIFT-Eval reproduction with V4 router + bucket calibrators, see Replication.

Training Recipe

• Base: amazon/chronos-2 (120M params, T5 encoder)
• Adapter: LoRA r=32, lora_alpha=64, lora_dropout=0, target_modules="all-linear", bias="none"
• Training data: 1,260 series total
  • Turkish synthetic telco augmentations
  • Turkish-themed KernelSynth (Chronos-2 synthetic generator with TR-domain priors)
  • TSMixup augmentation across heterogeneous series
• Validation: EPİAŞ (Energy Exchange Istanbul) 2025 out-of-time electricity market clearing prices (4 quarterly folds)
• Optimizer: AdamW (non-fused), lr=2e-4, cosine schedule, batch_size=8, max_steps=300
• Reproducibility (Tier 1.5):
  • torch==2.4.1+cu124, transformers==4.46.3, peft==0.13.2
  • CUBLAS_WORKSPACE_CONFIG=:4096:8, seed=42, torch.use_deterministic_algorithms(True)
  • SDP backend: math-only (enable_flash_sdp(False) + enable_mem_efficient_sdp(False) + enable_math_sdp(True))
  • torch.backends.cuda.matmul.allow_tf32=False, cudnn.deterministic=True

Replication

Full Colab notebook + per-stage builders included in this repo (last-checkpoint/ adapter snapshot + training_args.bin). The complete reproducible pipeline (data load → LoRA train → V4 router fit → bucket calibration → GIFT-Eval eval) is documented in the project decision log and is runnable end-to-end on an A100 40 GB session.

• Replication code available: Yes
• Test data leakage: No (GIFT-Eval canonical test windows held strictly out of training corpus per Salesforce AIRS pretrain manifest)

License

Apache-2.0 (inherited from amazon/chronos-2 base + clean LoRA adapter chain; no NC-restricted ancestor models in training corpus).

Citation

@misc{turkforecastfm2026,
  title  = {TurkForecast-FM-Chronos2-LoRA-v1: First Apache-2.0 Turkish-vertical Time-Series Foundation Model Fine-Tune},
  author = {Karatay, Mert},
  year   = {2026},
  url    = {https://huggingface.co/Verm1ion/turkforecast-fm-chronos2-lora-v1},
  note   = {GIFT-Eval submission v1.1, 2026-05-19}
}

Acknowledgements

• Amazon Science for Chronos-2 (base model + KernelSynth generator)
• Salesforce AI Research for GIFT-Eval (benchmark)
• EPİAŞ (Energy Exchange Istanbul, Enerji Piyasaları İşletme A.Ş.) for public Turkish electricity market data used in out-of-time validation