Contamination MapTime series / Model
TimesFM 2.0
- Organisation
- Google Research
- Family
- TimesFM · version 2.0
- Released
- 2024-12-30 (GitHub README 'Update - Dec. 30, 2024' announcing the 500m TimesFM-2.0 checkpoint (https://raw.githubusercontent.com/google-research/timesfm/master/v1/README.md); HF repo google/timesfm-2.0-500m-jax created 2024-12-20)
- Checkpoints
- google/timesfm-2.0-500m-jax, google/timesfm-2.0-500m-pytorch
- Weights
- Open · Apache-2.0
- Links
- PaperModel cardCode
Training data disclosure: Full list
Level determined for 44% of benchmark cells.
- TestTest data seen2
- SeenSeen, split unclear19
- TrainTraining data seen0
- Related data seen1
- NoNot seen0
- ?Unknown27
“It is important to list all the data sources in order to enable fair benchmarking.”
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
Training corpora
-
LOTSA (Large-scale Open Time Series Archive)
· 104 sources (complete list) · pre-training
· homepage
Evidence that TimesFM 2.0 uses it
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “pretraining data that many other models are pretrained on”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “LOTSA Pretrain”
huggingface.co · model card · Pretraining Data table, Download Source column · retrieved 2026-10-15 Contents of this corpus
-
BDG-2 Panther · split: unknown
HF config(s) in Salesforce/lotsa_data: bdg-2_panther.
Evidence
“BDG-2 Panther”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
BDG-2 Fox · split: unknown
HF config(s) in Salesforce/lotsa_data: bdg-2_fox.
Evidence
“BDG-2 Fox”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
BDG-2 Rat · split: unknown
HF config(s) in Salesforce/lotsa_data: bdg-2_rat.
Evidence
“BDG-2 Rat”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
BDG-2 Bear · split: unknown
HF config(s) in Salesforce/lotsa_data: bdg-2_bear.
Evidence
“BDG-2 Bear”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
Low Carbon London · split: unknown
HF config(s) in Salesforce/lotsa_data: lcl.
Evidence
“Low Carbon London”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
SMART · split: unknown
HF config(s) in Salesforce/lotsa_data: smart.
Evidence
“Low Carbon London SMART”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
IDEAL · split: unknown
HF config(s) in Salesforce/lotsa_data: ideal.
Evidence
“IDEAL”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
Sceaux · split: unknown
HF config(s) in Salesforce/lotsa_data: sceaux.
Evidence
“Sceaux”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
Borealis · split: unknown
HF config(s) in Salesforce/lotsa_data: borealis.
Evidence
“Borealis”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 -
Buildings900K · split: full
Both the BuildingsBench pre-train and test splits (by PUMA) are included. HF config(s) in Salesforce/lotsa_data: buildings_900k.
Evidence
“Buildings900K”
arxiv.org · paper · Appendix A, Table 8 (BuildingsBench) · retrieved 2026-10-15 “Emami et al. (2023) introduce a pre-train/test split based on Public Use Microdata Area, but we use include both splits in LOTSA for pre-training.”
arxiv.org · paper · Appendix A, BuildingsBench paragraph · retrieved 2026-10-15 -
CMIP6 · split: unknown
HF config(s) in Salesforce/lotsa_data: cmip6 (+ cmip6_1850 ... cmip6_2010).
Evidence
“CMIP6 ERA5”
arxiv.org · paper · Appendix A, Table 9 (ClimateLearn) · retrieved 2026-10-15 -
ERA5 · split: unknown
HF config(s) in Salesforce/lotsa_data: era5 (+ era5_1989 ... era5_2018).
Evidence
“CMIP6 ERA5”
arxiv.org · paper · Appendix A, Table 9 (ClimateLearn) · retrieved 2026-10-15 -
Azure VM Traces 2017 · split: train
Train split is the CloudOps TSF pre-train/test split of Woo et al. (2023), not GIFT-Eval or long-horizon. HF config(s) in Salesforce/lotsa_data: azure_vm_traces_2017.
Evidence
“Azure VM Traces 2017”
arxiv.org · paper · Appendix A, Table 10 (CloudOps TSF) · retrieved 2026-10-15 “We follow their pre-train/test split and only include the pre-train time series in LOTSA, holding out the test set.”
arxiv.org · paper · Appendix A, CloudOps TSF paragraph · retrieved 2026-10-15 -
Borg Cluster Data 2011 · split: train
Train split is the CloudOps TSF pre-train/test split of Woo et al. (2023), not GIFT-Eval or long-horizon. HF config(s) in Salesforce/lotsa_data: borg_cluster_data_2011.
Evidence
“Borg Cluster Data 2011”
arxiv.org · paper · Appendix A, Table 10 (CloudOps TSF) · retrieved 2026-10-15 “We follow their pre-train/test split and only include the pre-train time series in LOTSA, holding out the test set.”
arxiv.org · paper · Appendix A, CloudOps TSF paragraph · retrieved 2026-10-15 -
Alibaba Cluster Trace 2018 · split: train
Train split is the CloudOps TSF pre-train/test split of Woo et al. (2023), not GIFT-Eval or long-horizon. HF config(s) in Salesforce/lotsa_data: alibaba_cluster_trace_2018.
Evidence
“Alibaba Cluster Trace 2018”
arxiv.org · paper · Appendix A, Table 10 (CloudOps TSF) · retrieved 2026-10-15 “We follow their pre-train/test split and only include the pre-train time series in LOTSA, holding out the test set.”
arxiv.org · paper · Appendix A, CloudOps TSF paragraph · retrieved 2026-10-15 -
Taxi · split: unknown
HF config(s) in Salesforce/lotsa_data: taxi_30min.
Evidence
“Taxi Uber TLC Daily”
arxiv.org · paper · Appendix A, Table 11 (GluonTS) · retrieved 2026-10-15 -
Uber TLC Daily · split: unknown
HF config(s) in Salesforce/lotsa_data: uber_tlc_daily.
Evidence
“Uber TLC Daily”
arxiv.org · paper · Appendix A, Table 11 (GluonTS) · retrieved 2026-10-15 -
Uber TLC Hourly · split: unknown
HF config(s) in Salesforce/lotsa_data: uber_tlc_hourly.
Evidence
“Uber TLC Hourly”
arxiv.org · paper · Appendix A, Table 11 (GluonTS) · retrieved 2026-10-15 -
Wiki-Rolling · split: unknown
HF config(s) in Salesforce/lotsa_data: wiki-rolling_nips.
Evidence
“Wiki-Rolling”
arxiv.org · paper · Appendix A, Table 11 (GluonTS) · retrieved 2026-10-15 -
M5 · split: unknown
HF config(s) in Salesforce/lotsa_data: m5.
Evidence
“Wiki-Rolling M5”
arxiv.org · paper · Appendix A, Table 11 (GluonTS) · retrieved 2026-10-15 -
LargeST (PeMS California) · split: unknown
Same PeMS source as PEMS03/04/07/08, PEMS Bay and long-horizon Traffic (per the paper). HF config(s) in Salesforce/lotsa_data: largest (+ largest_2017 ... largest_2021).
Evidence
“LargeST Transport 5T”
arxiv.org · paper · Appendix A, Table 12 (LargeST) · retrieved 2026-10-15 “Since we use such a large amount of dataset from the same source, we can no longer consider forecasting on any of these datasets as an out-of-distribution or zero-shot forecasting task anymore, and thus omit the Traffic dataset of the LSF benchmark from our evaluations.”
arxiv.org · paper · Appendix A, LargeST paragraph · retrieved 2026-10-15 -
PEMS03 · split: unknown
HF config(s) in Salesforce/lotsa_data: PEMS03.
Evidence
“PEMS03”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
PEMS04 · split: unknown
HF config(s) in Salesforce/lotsa_data: PEMS04.
Evidence
“PEMS04”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
PEMS07 · split: unknown
HF config(s) in Salesforce/lotsa_data: PEMS07.
Evidence
“PEMS07”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
PEMS08 · split: unknown
HF config(s) in Salesforce/lotsa_data: PEMS08.
Evidence
“PEMS08”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
PEMS Bay · split: unknown
HF config(s) in Salesforce/lotsa_data: PEMS_BAY.
Evidence
“PEMS Bay”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
Los-Loop · split: unknown
HF config(s) in Salesforce/lotsa_data: LOS_LOOP.
Evidence
“Los-Loop”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
Loop Seattle (LibCity) · split: unknown
GIFT-Eval paper Table 23 lists this dataset among those 'that LOTSA collection and our GIFT-Eval has in common'. HF config(s) in Salesforce/lotsa_data: LOOP_SEATTLE.
Evidence
“Loop Seattle”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
SZ-Taxi (LibCity) · split: unknown
GIFT-Eval paper Table 23 lists this dataset among those 'that LOTSA collection and our GIFT-Eval has in common'. HF config(s) in Salesforce/lotsa_data: SZ_TAXI.
Evidence
“SZ-Taxi”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
Beijing Subway · split: unknown
HF config(s) in Salesforce/lotsa_data: BEIJING_SUBWAY_30MIN.
Evidence
“Beijing Subway”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
SHMetro · split: unknown
HF config(s) in Salesforce/lotsa_data: SHMETRO.
Evidence
“SHMetro”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
HZMetro · split: unknown
HF config(s) in Salesforce/lotsa_data: HZMETRO.
Evidence
“HZMetro”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
M_DENSE (LibCity) · split: unknown
No HF config named Rotterdam; the HF dataset and the uni2ts LibCity builder list 'M_DENSE' in its place (13 LibCity entries in both). See open_questions. Matched to M_DENSE: the published LOTSA data and the uni2ts LibCity builder list M_DENSE in place of 'Rotterdam', and the GIFT-Eval paper (Table 23) lists m_dense among datasets shared with LOTSA. To be confirmed.
Evidence
“Rotterdam”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
Q-Traffic · split: unknown
HF config(s) in Salesforce/lotsa_data: Q-TRAFFIC.
Evidence
“Q-Traffic”
arxiv.org · paper · Appendix A, Table 13 (LibCity) · retrieved 2026-10-15 -
London Smart Meters · split: full
HF config(s) in Salesforce/lotsa_data: london_smart_meters_with_missing.
Evidence
“London Smart Meters”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “London Smart Meters, Wind Farms, Wind Power, Solar Power, Oikolab Weather, Covid Mobility: Originally not included in the Monash benchmark.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Wind Farms · split: full
HF config(s) in Salesforce/lotsa_data: wind_farms_with_missing.
Evidence
“Wind Farms”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “London Smart Meters, Wind Farms, Wind Power, Solar Power, Oikolab Weather, Covid Mobility: Originally not included in the Monash benchmark.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Wind Power · split: full
HF config(s) in Salesforce/lotsa_data: wind_power.
Evidence
“Wind Power”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “London Smart Meters, Wind Farms, Wind Power, Solar Power, Oikolab Weather, Covid Mobility: Originally not included in the Monash benchmark.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Solar Power · split: full
HF config(s) in Salesforce/lotsa_data: solar_power.
Evidence
“Solar Power”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “London Smart Meters, Wind Farms, Wind Power, Solar Power, Oikolab Weather, Covid Mobility: Originally not included in the Monash benchmark.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Oikolab Weather · split: full
HF config(s) in Salesforce/lotsa_data: oikolab_weather.
Evidence
“Oikolab Weather”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “London Smart Meters, Wind Farms, Wind Power, Solar Power, Oikolab Weather, Covid Mobility: Originally not included in the Monash benchmark.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Elecdemand · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: elecdemand.
Evidence
“Elecdemand”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Covid Mobility · split: full
HF config(s) in Salesforce/lotsa_data: covid_mobility.
Evidence
“Covid Mobility”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “London Smart Meters, Wind Farms, Wind Power, Solar Power, Oikolab Weather, Covid Mobility: Originally not included in the Monash benchmark.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Kaggle Web Traffic Weekly · split: full
HF config(s) in Salesforce/lotsa_data: kaggle_web_traffic_weekly.
Evidence
“Kaggle Web Traffic Weekly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “Extended Web Traffic, Kaggle Web Traffic Weekly: We include the extended version of Web Traffic which contains overlap with the original Web Traffic dataset.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Extended Web Traffic · split: full
HF config(s) in Salesforce/lotsa_data: extended_web_traffic_with_missing.
Evidence
“Extended Web Traffic”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “Extended Web Traffic, Kaggle Web Traffic Weekly: We include the extended version of Web Traffic which contains overlap with the original Web Traffic dataset.”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
M1 Yearly · split: full
HF config(s) in Salesforce/lotsa_data: m1_yearly.
Evidence
“M1 Yearly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
M1 Quarterly · split: full
HF config(s) in Salesforce/lotsa_data: m1_quarterly.
Evidence
“M1 Quarterly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
M1 Monthly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: m1_monthly.
Evidence
“M1 Monthly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
M3 Yearly · split: full
HF config(s) in Salesforce/lotsa_data: monash_m3_yearly.
Evidence
“M3 Yearly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
M3 Quarterly · split: full
HF config(s) in Salesforce/lotsa_data: monash_m3_quarterly.
Evidence
“M3 Quarterly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
M3 Monthly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: monash_m3_monthly.
Evidence
“M3 Monthly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
M3 Other · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: monash_m3_other.
Evidence
“M3 Other”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
M4 Yearly · split: full
HF config(s) in Salesforce/lotsa_data: m4_yearly.
Evidence
“M4 Yearly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
M4 Quarterly · split: full
HF config(s) in Salesforce/lotsa_data: m4_quarterly.
Evidence
“M4 Quarterly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
M4 Monthly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: m4_monthly.
Evidence
“M4 Monthly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
M4 Weekly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: m4_weekly.
Evidence
“M4 Weekly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
M4 Hourly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: m4_hourly.
Evidence
“M4 Hourly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
M4 Daily · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: m4_daily.
Evidence
“M4 Daily”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
NN5 Daily · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: nn5_daily_with_missing.
Evidence
“NN5 Daily”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
NN5 Weekly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: nn5_weekly.
Evidence
“NN5 Weekly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Tourism Yearly · split: full
HF config(s) in Salesforce/lotsa_data: tourism_yearly.
Evidence
“Tourism Yearly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
arxiv.org · paper · Appendix A, Monash exceptions list · retrieved 2026-10-15 -
Tourism Quarterly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: tourism_quarterly.
Evidence
“Tourism Quarterly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Tourism Monthly · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: tourism_monthly.
Evidence
“Tourism Monthly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
CIF 2016 · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: cif_2016_12, cif_2016_6.
Evidence
“CIF 2016”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Traffic Weekly (Monash) · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: traffic_weekly.
Evidence
“Traffic Weekly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Traffic (PeMS San Francisco Bay, 862 sensors) · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. Monash Traffic Hourly; whether it is the same series as long-horizon 'Traffic' is not stated by the Moirai paper (see open_questions). HF config(s) in Salesforce/lotsa_data: traffic_hourly.
Evidence
“Traffic Hourly”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Australian Electricity Demand · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: australian_electricity_demand.
Evidence
“Australian Electricity Demand”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Rideshare · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: rideshare_with_missing.
Evidence
“Rideshare”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Saugeen River Flow · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: saugeenday.
Evidence
“Saugeen”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Sunspot · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: sunspot_with_missing.
Evidence
“Sunspot”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Temperature Rain · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: temperature_rain_with_missing.
Evidence
“Temperature Rain”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Vehicle Trips · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: vehicle_trips_with_missing.
Evidence
“Vehicle Trips”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Weather · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. This is Monash 'Weather', not the 10-minute Weather dataset of the long-horizon benchmark. HF config(s) in Salesforce/lotsa_data: weather.
Evidence
“Vehicle Trips Weather”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 “Note that the “Weather” from Monash is a different dataset from that used in the out-of-distribution evaluations.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Car Parts · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: car_parts_with_missing.
Evidence
“Car Parts”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
FRED MD · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: fred_md.
Evidence
“FRED MD”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Pedestrian Counts · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: pedestrian_counts.
Evidence
“Pedestrian Counts”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Hospital · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: hospital.
Evidence
“Hospital”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
COVID Deaths · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: covid_deaths.
Evidence
“COVID Deaths”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
KDD Cup 2018 · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: kdd_cup_2018_with_missing.
Evidence
“KDD Cup 2018”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Bitcoin · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: bitcoin_with_missing.
Evidence
“Bitcoin”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
US Births · split: train
Train region only: the final Monash forecast horizon (Godahewa et al., 2021) is held out as test. HF config(s) in Salesforce/lotsa_data: us_births.
Evidence
“US Births”
arxiv.org · paper · Appendix A, Table 14 (Monash) · retrieved 2026-10-15 “apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
arxiv.org · paper · Appendix A, Monash paragraph · retrieved 2026-10-15 -
Covid19 Energy · split: unknown
HF config(s) in Salesforce/lotsa_data: covid19_energy.
Evidence
“Covid19 Energy”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
GEF12 · split: unknown
HF config(s) in Salesforce/lotsa_data: gfc12_load.
Evidence
“GEF12”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
GEF14 · split: unknown
HF config(s) in Salesforce/lotsa_data: gfc14_load.
Evidence
“GEF14”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
GEF17 · split: unknown
HF config(s) in Salesforce/lotsa_data: gfc17_load.
Evidence
“GEF17”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
PDB · split: unknown
HF config(s) in Salesforce/lotsa_data: pdb.
Evidence
“GEF17 PDB”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
Spanish · split: unknown
HF config(s) in Salesforce/lotsa_data: spain.
Evidence
“Spanish”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
BDG-2 Hog · split: unknown
HF config(s) in Salesforce/lotsa_data: hog.
Evidence
“BDG-2 Hog”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
BDG-2 Bull · split: unknown
HF config(s) in Salesforce/lotsa_data: bull.
Evidence
“BDG-2 Bull”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
BDG-2 Cockatoo · split: unknown
HF config(s) in Salesforce/lotsa_data: cockatoo.
Evidence
“BDG-2 Cockatoo”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
ELF · split: unknown
HF config(s) in Salesforce/lotsa_data: elf.
Evidence
“BDG-2 Cockatoo ELF”
arxiv.org · paper · Appendix A, Table 15 (ProEnFo) · retrieved 2026-10-15 -
Subseasonal · split: unknown
HF config(s) in Salesforce/lotsa_data: subseasonal.
Evidence
“Subseasonal Subseasonal Precipitation”
arxiv.org · paper · Appendix A, Table 16 (SubseasonalClimateUSA) · retrieved 2026-10-15 -
Subseasonal Precipitation · split: unknown
HF config(s) in Salesforce/lotsa_data: subseasonal_precip.
Evidence
“Subseasonal Subseasonal Precipitation”
arxiv.org · paper · Appendix A, Table 16 (SubseasonalClimateUSA) · retrieved 2026-10-15 -
KDD Cup 2022 · split: unknown
HF config(s) in Salesforce/lotsa_data: kdd2022.
Evidence
“KDD Cup 2022”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
GoDaddy · split: unknown
HF config(s) in Salesforce/lotsa_data: godaddy.
Evidence
“GoDaddy”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Favorita Sales · split: unknown
HF config(s) in Salesforce/lotsa_data: favorita_sales.
Evidence
“Favorita Sales”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Favorita Transactions · split: unknown
HF config(s) in Salesforce/lotsa_data: favorita_transactions.
Evidence
“Favorita Transactions”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Restaurant (Recruit Restaurant Visitor Forecasting) · split: unknown
GIFT-Eval paper Table 23 lists this dataset among those 'that LOTSA collection and our GIFT-Eval has in common'. HF config(s) in Salesforce/lotsa_data: restaurant.
Evidence
“Restaurant”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Hierarchical Sales (Mancuso et al., 2020) · split: unknown
GIFT-Eval paper Table 23 lists this dataset among those 'that LOTSA collection and our GIFT-Eval has in common'. HF config(s) in Salesforce/lotsa_data: hierarchical_sales.
Evidence
“Hierarchical Sales”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
China Air Quality · split: unknown
HF config(s) in Salesforce/lotsa_data: china_air_quality.
Evidence
“China Air Quality”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Beijing Air Quality · split: unknown
HF config(s) in Salesforce/lotsa_data: beijing_air_quality.
Evidence
“Beijing Air Quality”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Residential Load Power · split: unknown
HF config(s) in Salesforce/lotsa_data: residential_load_power.
Evidence
“Residential Load Power”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Residential PV Power · split: unknown
HF config(s) in Salesforce/lotsa_data: residential_pv_power.
Evidence
“Residential PV Power”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
CDC FluView ILINet (LOTSA) · split: unknown
CDC ILINet source; the paper does not say whether this overlaps the long-horizon 'ILI' dataset (see open_questions). HF config(s) in Salesforce/lotsa_data: cdc_fluview_ilinet.
Evidence
“CDC Fluview ILINet”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
CDC Fluview WHO NREVSS · split: unknown
HF config(s) in Salesforce/lotsa_data: cdc_fluview_who_nrevss.
Evidence
“CDC Fluview WHO NREVSS”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15 -
Project Tycho · split: unknown
HF config(s) in Salesforce/lotsa_data: project_tycho.
Evidence
“Project Tycho”
arxiv.org · paper · Appendix A, Table 17 (Others) · retrieved 2026-10-15
-
BDG-2 Panther · split: unknown
Other declared training data
-
Google Trends · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). ~22k head queries; hourly Jan 2018-Dec 2019, daily/weekly/monthly Jan 2007-Dec 2021 (Table 1 rows: Trends hourly/daily/weekly/monthly). Not publicly released.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We download the search interest over time for these queries in hourly, daily, weekly and monthly granularities to form our dataset.”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “Trends hourly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
Wiki Pageviews · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Table 1 rows: Wiki hourly/daily/weekly/monthly.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We download all pageview data from Jan. 2012 to Nov. 2023, clean and aggregate the views by page into hourly, daily, weekly and monthly granularities, and filter out pageview time-series with excessive zeros.”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “Wiki hourly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
Synthetic Data · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). ARMA, seasonal (sines/cosines), trends, step functions; generator described in Appendix A.8.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We create 3M synthetic time-series each of length 2048 time-points.”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “A.8 Synthetic Data”
arxiv.org · paper · Appendix A.8 · retrieved 2026-10-15 -
M4 Yearly · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Text says 'all the granularities'; Table 1 lists M4 hourly, daily, monthly, quarterly, yearly (no weekly row). The paper says 'all the granularities of the M4 dataset'.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “M4 quarterly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
M4 Quarterly · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Text says 'all the granularities'; Table 1 lists M4 hourly, daily, monthly, quarterly, yearly (no weekly row). The paper says 'all the granularities of the M4 dataset'.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “M4 quarterly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
M4 Monthly · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Text says 'all the granularities'; Table 1 lists M4 hourly, daily, monthly, quarterly, yearly (no weekly row). The paper says 'all the granularities of the M4 dataset'.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “M4 quarterly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
M4 Weekly · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Text says 'all the granularities'; Table 1 lists M4 hourly, daily, monthly, quarterly, yearly (no weekly row). The paper says 'all the granularities of the M4 dataset'.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “M4 quarterly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
M4 Daily · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Text says 'all the granularities'; Table 1 lists M4 hourly, daily, monthly, quarterly, yearly (no weekly row). The paper says 'all the granularities of the M4 dataset'.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “M4 quarterly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
M4 Hourly · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Text says 'all the granularities'; Table 1 lists M4 hourly, daily, monthly, quarterly, yearly (no weekly row). The paper says 'all the granularities of the M4 dataset'.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “M4 quarterly”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
Electricity (ECL, 321 clients) · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Hourly and 15-minute; reference [ZZP+ 21] is Informer (Zhou et al. 2021). Split not stated. Hourly (321-client) and 15-minute (UCI) versions.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 -
ElectricityLoadDiagrams20112014 (UCI) · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Hourly and 15-minute; reference [ZZP+ 21] is Informer (Zhou et al. 2021). Split not stated. Hourly (321-client) and 15-minute (UCI) versions.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 -
Traffic (PeMS San Francisco Bay, 862 sensors) · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Hourly; reference [ZZP+ 21] (Informer). Split not stated.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 -
Weather (not yet matched to a dataset) · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). 10-minute granularity, 'used for evaluations in [ZZP+ 21]'. Split not stated.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “We also add the 10-minute granularity Weather dataset used for evaluations in [ZZP+ 21].”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “Weather [ZZP+ 21]”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
LibCity (not yet matched to a dataset) · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). All 15-min granularity traffic time-series from LibCity [WJJ+ 23].
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “In addition, we add all the 15 min granularity traffic time-series from [WJJ+ 23].”
arxiv.org · paper · Section 5 (Pretraining Details) · retrieved 2026-10-15 “LibCity [WJJ+ 23]”
arxiv.org · paper · Table 1 · retrieved 2026-10-15 -
Favorita Sales · split: unknown
Inherited from the TimesFM 1.0 pretraining set (model card). Appears only as a row of Table 1; not described in the text.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “Favorita Sales”
arxiv.org · paper · Table 1 (Composition of TimesFM pretraining dataset) · retrieved 2026-10-15 -
azure_vm_traces · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “azure_vm_traces”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
residential_load_power · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “residential_load_power”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
borg_cluster_data · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “borg_cluster_data”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
residential_pv_power · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “residential_pv_power”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
q_traffic · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “q_traffic”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
london_smart_meters_with_missing · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “london_smart_meters_with_missing”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
taxi_30min · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “taxi_30min”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
solar_power · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “solar_power”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
wind_power · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “wind_power”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
kdd2022 · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “kdd2022”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
LargeST (PeMS California) · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “largest”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
era5 · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “era5”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
buildings · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “buildings”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
cmip6 · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “cmip6”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
china_air_quality · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “china_air_quality”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
beijing_air_quality · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “beijing_air_quality”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
subseasonal · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “subseasonal”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
kaggle_web_traffic_weekly · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “kaggle_web_traffic_weekly”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
cdc_fluview_who_nrevss · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “cdc_fluview_who_nrevss”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15 -
godaddy · split: unknown
Download Source: LOTSA Pretrain.
Evidence
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
huggingface.co · model card · Pretraining Data · retrieved 2026-10-15 “godaddy”
huggingface.co · model card · Pretraining Data table (Download Source: LOTSA Pretrain) · retrieved 2026-10-15
Benchmark cells
GIFT-Eval
Unknown Jena Weather (GIFT-Eval) · training data not disclosed in enough detail
Unknown BizITObs - Application · training data not disclosed in enough detail
Unknown BizITObs - Service · training data not disclosed in enough detail
Unknown BizITObs - L2C · training data not disclosed in enough detail
Unknown Bitbrains - Fast Storage · training data not disclosed in enough detail
Unknown Bitbrains - rnd · training data not disclosed in enough detail
Seen, split unclear Restaurant (Recruit Restaurant Visitor Forecasting) · the dataset was used; the source does not say which part (via LOTSA (Large-scale Open Time Series Archive))
Evidence (4)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Restaurant”
Unknown ETT1 (GIFT-Eval) · training data not disclosed in enough detail
Unknown ETT2 (GIFT-Eval) · training data not disclosed in enough detail
Seen, split unclear Loop Seattle (LibCity) · the dataset was used; the source does not say which part (via LOTSA (Large-scale Open Time Series Archive))
Evidence (4)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Loop Seattle”
Seen, split unclear SZ-Taxi (LibCity) · the dataset was used; the source does not say which part (via LOTSA (Large-scale Open Time Series Archive))
Evidence (4)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“SZ-Taxi”
Seen, split unclear M_DENSE (LibCity) · the dataset was used; the source does not say which part (via LOTSA (Large-scale Open Time Series Archive))
Evidence (4)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Rotterdam”
Unknown Solar (LSTNet, 137 Alabama PV plants) · training data not disclosed in enough detail
Seen, split unclear Hierarchical Sales (Mancuso et al., 2020) · the dataset was used; the source does not say which part (via LOTSA (Large-scale Open Time Series Archive))
Evidence (4)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Hierarchical Sales”
Test data seen M4 Yearly · the whole dataset was used (via LOTSA (Large-scale Open Time Series Archive))
Evidence (6)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“M4 Yearly”
“For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
“M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
Test data seen M4 Quarterly · the whole dataset was used (via LOTSA (Large-scale Open Time Series Archive))
Evidence (6)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“M4 Quarterly”
“For the following datasets, we include their entirety in LOTSA without a held-out test set for the following reasons:”
“M1 Yearly, M1 Quarterly, M3 Yearly, M3 Quarterly, M4 Yearly, M4 Quarterly, Tourism Yearly: Some time series in these datasets are too short after train/test splits, thus we do not split them”
Seen, split unclear M4 Monthly · the dataset was used; the source does not say which part
Evidence (3)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
“M4 quarterly”
Seen, split unclear M4 Weekly · the dataset was used; the source does not say which part
Evidence (3)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
“M4 quarterly”
Seen, split unclear M4 Daily · the dataset was used; the source does not say which part
Evidence (3)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
“M4 quarterly”
Seen, split unclear M4 Hourly · the dataset was used; the source does not say which part
Evidence (3)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
“M4 quarterly”
Seen, split unclear Hospital · a training split was used; its overlap with this benchmark's test data is not established (via LOTSA (Large-scale Open Time Series Archive))
Evidence (5)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Hospital”
“apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
Seen, split unclear COVID Deaths · a training split was used; its overlap with this benchmark's test data is not established (via LOTSA (Large-scale Open Time Series Archive))
Evidence (5)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“COVID Deaths”
“apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
Seen, split unclear US Births · a training split was used; its overlap with this benchmark's test data is not established (via LOTSA (Large-scale Open Time Series Archive))
Evidence (5)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“US Births”
“apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
Seen, split unclear Saugeen River Flow · a training split was used; its overlap with this benchmark's test data is not established (via LOTSA (Large-scale Open Time Series Archive))
Evidence (5)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Saugeen”
“apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
Seen, split unclear Temperature Rain · a training split was used; its overlap with this benchmark's test data is not established (via LOTSA (Large-scale Open Time Series Archive))
Evidence (5)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Temperature Rain”
“apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
Seen, split unclear KDD Cup 2018 · a training split was used; its overlap with this benchmark's test data is not established (via LOTSA (Large-scale Open Time Series Archive))
Evidence (5)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“KDD Cup 2018”
“apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
Seen, split unclear Car Parts · a training split was used; its overlap with this benchmark's test data is not established (via LOTSA (Large-scale Open Time Series Archive))
Evidence (5)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“Car Parts”
“apart from the exceptions listed below, we only include the train region in LOTSA, by holding out the final forecast horizon as the test set.”
Seen, split unclear ElectricityLoadDiagrams20112014 (UCI) · the dataset was used; the source does not say which part
Evidence (2)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
Long-horizon forecasting (ETT, Electricity, Traffic, Weather, Exchange, ILI)
Unknown ETTh1 · training data not disclosed in enough detail
Unknown ETTh2 · training data not disclosed in enough detail
Unknown ETTm1 · training data not disclosed in enough detail
Unknown ETTm2 · training data not disclosed in enough detail
Seen, split unclear Electricity (ECL, 321 clients) · the dataset was used; the source does not say which part
Evidence (2)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
Seen, split unclear Traffic (PeMS San Francisco Bay, 862 sensors) · the dataset was used; the source does not say which part
Evidence (2)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“We add all the granularities of the M4 dataset [MSA22], the hourly and 15 minute Electricity and the hourly Traffic datasets (see [ZZP+ 21]).”
Unknown Weather (Jena, 2020; Autoformer) · training data not disclosed in enough detail
Unknown Exchange-Rate · training data not disclosed in enough detail
ILI (CDC influenza-like illness, national) · related dataset CDC FluView ILINet (LOTSA) was used (via LOTSA (Large-scale Open Time Series Archive))
Evidence (7)
“The TimesFM 2.0 series contains the pretraining set of TimesFM 1.0 along with these additional datasets”
“pretraining data that many other models are pretrained on”
“LOTSA Pretrain”
“CDC Fluview ILINet”
“includes the weekly recorded influenza-like illness (ILI) patients data from Centers for Disease Control and Prevention of the United States between 2002 and 2021”
“https://gis.cdc.gov/grasp/fluview/fluportaldashboard.html”
“CDC Fluview ILINet CDC Fluview WHO NREVSS”
Time-series anomaly detection
Unknown TSB-AD-U (TSB-AD univariate) · training data not disclosed in enough detail
Unknown TSB-AD-M (TSB-AD multivariate) · training data not disclosed in enough detail
Unknown UCR Time Series Anomaly Archive (Hexagon ML/UCR, KDD Cup 2021) · training data not disclosed in enough detail
Unknown Server Machine Dataset (SMD) · training data not disclosed in enough detail
Unknown Mars Science Laboratory (MSL) telemetry anomaly dataset · training data not disclosed in enough detail
Unknown Soil Moisture Active Passive (SMAP) telemetry anomaly dataset · training data not disclosed in enough detail
Unknown Pooled Server Metrics (PSM) · training data not disclosed in enough detail
Unknown Secure Water Treatment (SWaT) · training data not disclosed in enough detail
Unknown Yahoo S5 Labeled Anomaly Detection Dataset (Yahoo Webscope S5) · training data not disclosed in enough detail
Unknown Numenta Anomaly Benchmark (NAB) · training data not disclosed in enough detail
Time-series classification
Unknown UCR Time Series Classification Archive (2018) · training data not disclosed in enough detail
Unknown UEA Multivariate Time Series Classification Archive (2018) · training data not disclosed in enough detail
Notes
No separate paper; data disclosed in the HF model card. GIFT-Eval's results config for timesfm_2_0_500m sets "testdata_leakage": "Yes". The 1.0 paper's held-out statements (ETT, Monash, Darts) are not restated for 2.0, and the added LOTSA subset contains names that resemble Monash datasets (london_smart_meters_with_missing, solar_power, wind_power, kaggle_web_traffic_weekly) and CDC FluView (cdc_fluview_who_nrevss) — mapping left to the lead.
Sources
- google/timesfm-2.0-500m-pytorch model card · model card · retrieved 2026-10-15
- google/timesfm-2.0-500m-jax model card (same data section) · model card · retrieved 2026-10-15
- google-research/timesfm v1/README.md · readme · retrieved 2026-10-15
- A decoder-only foundation model for time-series forecasting (arXiv v4) · paper · retrieved 2026-10-15
- GIFT-Eval results config for timesfm_2_0_500m · leaderboard · retrieved 2026-10-15