README.md

metadata

pretty_name: Lapras Reasoning Datasets
license: other
language:
  - en
task_categories:
  - question-answering
  - text-generation
tags:
  - time-series
  - reasoning
  - chain-of-thought
  - latent-reasoning
  - ecg
  - sleep-staging
  - human-activity-recognition
size_categories:
  - 100K<n<1M
configs:
  - config_name: ecg
    data_files:
      - split: train
        path: ecg/train.jsonl
      - split: test
        path: ecg/test.jsonl
  - config_name: ecg_ts_tags
    data_files:
      - split: train
        path: ecg/train_with_ts_tags.jsonl
      - split: test
        path: ecg/test_with_ts_tags.jsonl
  - config_name: sleep
    data_files:
      - split: train
        path: sleep/train.jsonl
      - split: test
        path: sleep/test.jsonl
  - config_name: sleep_ts_tags
    data_files:
      - split: train
        path: sleep/train_with_ts_tags.jsonl
      - split: test
        path: sleep/test_with_ts_tags.jsonl
  - config_name: har
    data_files:
      - split: train
        path: har/train.jsonl
      - split: test
        path: har/test.jsonl
  - config_name: har_ts_tags
    data_files:
      - split: train
        path: har/train_with_ts_tags.jsonl
      - split: test
        path: har/test_with_ts_tags.jsonl
  - config_name: tsr
    data_files:
      - split: train
        path: tsr/train.jsonl
      - split: test
        path: tsr/test.jsonl
  - config_name: tsr_ts_tags
    data_files:
      - split: train
        path: tsr/train_with_ts_tags.jsonl
      - split: test
        path: tsr/test_with_ts_tags.jsonl
  - config_name: engine
    data_files:
      - split: train
        path: engine/train.jsonl
      - split: test
        path: engine/test.jsonl
  - config_name: engine_ts_tags
    data_files:
      - split: train
        path: engine/train_with_ts_tags.jsonl
      - split: test
        path: engine/test_with_ts_tags.jsonl

Dataset Card for Lapras Reasoning Datasets

Five time series question answering datasets with reference reasoning traces, used to train and evaluate Lapras (latent reasoning for time series language models) and its No-CoT, CoT, iCoT and COCONUT baselines.

Dataset Details

Dataset Description

Each example pairs a natural-language question with a multichannel time series and a reference reasoning trace that ends in Answer: <label>. The five tasks cover cardiological diagnosis (ECG), sleep-stage classification (Sleep), human activity recognition (HAR), counterfactual consequence prediction (TSR) and aero-engine fault diagnosis (Engine). Every task ships in two prompt variants with identical rows: a plain prompt, and a *_with_ts_tags prompt with one <name> (length L): <ts><ts/> placeholder line per channel.

• Curated by: the Lapras authors (Yuliang Chen et al., see Citation)
• Language(s) (NLP): English
• License: per subset (see Data Sources)

Links

• Repository: https://github.com/yuc0805/Lapras
• Paper: https://arxiv.org/abs/2610.11111
• Checkpoints: https://huggingface.co/leochen085/Lapras

Uses

Direct Use

• Supervised finetuning and evaluation of time series language models, with or without reasoning traces.
• Reproducing the Lapras experiments: download into the dataset/ folder of the code repository.

huggingface-cli download leochen085/Lapras-Reasoning-Datasets --repo-type dataset --include "*.jsonl" --local-dir dataset

Quick start

from datasets import load_dataset

ds = load_dataset("leochen085/Lapras-Reasoning-Datasets", "har")
print(ds)

DatasetDict({
    train: Dataset({
        features: ['input', 'timeseries', 'output'],
        num_rows: 68542
    })
    test: Dataset({
        features: ['input', 'timeseries', 'output'],
        num_rows: 8222
    })
})

Each row is a dict with input (str), timeseries (list of channels, each a list of floats) and output (str).

Processor, collator and DataLoader

lapras_dataloader.py holds a standard multimodal processor and collator, as for vision-language models: the processor tokenizes the text and packs the time series into a tensor, and the collator builds training batches in which only the response is supervised.

from huggingface_hub import hf_hub_download
hf_hub_download("leochen085/Lapras-Reasoning-Datasets", "lapras_dataloader.py", repo_type="dataset", local_dir=".")

from torch.utils.data import DataLoader
from transformers import AutoTokenizer
from lapras_dataloader import TimeSeriesProcessor, TimeSeriesCollator

processor = TimeSeriesProcessor(AutoTokenizer.from_pretrained("Qwen/Qwen3-8B"))
loader = DataLoader(ds["train"], batch_size=4, shuffle=True, collate_fn=TimeSeriesCollator(processor))
batch = next(iter(loader))

input_ids        (4, 514)       torch.int64     # length varies by batch
attention_mask   (4, 514)       torch.int64
timeseries       (4, 3, 128)    torch.float32   # (batch, channels, length), zero-padded
timeseries_mask  (4, 3, 128)    torch.bool      # True on real samples
labels           (4, 514)       torch.int64     # -100 on the prompt and padding

Series of different lengths (TSR) are zero-padded to the longest in the batch and marked in timeseries_mask.

Decoding back

tok = processor.tokenizer

ids = batch["input_ids"][0][batch["attention_mask"][0].bool()]
tok.decode(ids)
# '<|im_start|>user\nYou are given accelerometer data in all three dimensions. ... <|eot|>\nAnswer: sitting<|im_end|>'

target = tok.decode(batch["labels"][0][batch["labels"][0] != -100], skip_special_tokens=True)
target                                   # the example's `output`, character for character
target.rsplit("Answer:", 1)[1].strip()   # e.g. 'sitting'

n_ch = batch["timeseries_mask"][0].any(-1).sum()
batch["timeseries"][0][batch["timeseries_mask"][0]].view(n_ch, -1).tolist()  # the example's `timeseries`

Dataset Structure

ecg/     train.jsonl  test.jsonl  train_with_ts_tags.jsonl  test_with_ts_tags.jsonl
sleep/   ...
har/     ...
tsr/     ...
engine/  ...

Config	Task	Train	Test	Channels	Length	Answer
ecg	cardiological diagnosis	5,800	643	12	1000	Yes / No
sleep	sleep-stage classification	7,413	923	1	1500	Wake, Non-REM stage 1, Non-REM stage 2, Non-REM stage 3, REM sleep
har	human activity recognition	68,542	8,222	3	128	biking, lying, running, sitting, standing, walking, walking_down, walking_up
tsr	counterfactual consequence prediction	22,566	4,094	1	128–1020	option letter + text (A–D)
engine	aero-engine fault diagnosis	2,305	193	33	600	option letter + text (A–D)

Data Fields

• input (string): the prompt (question, instructions and, for multiple choice, the options).
• timeseries (list of list of float): one list per channel. ECG and Sleep series are z-normalized per channel; HAR, TSR and Engine series keep the scale of their source release.
• output (string): the reference trace. Except for ECG, reasoning steps are delimited by <|bot|> and <|eot|>; HAR, Sleep and TSR traces also contain segment-selection calls <timeseries_selection_tool> [channel, start, end] </timeseries_selection_tool> (Sleep: [start, end]; inclusive sample indices). ECG traces are a single paragraph. Every trace ends with Answer: <label>.
• ecg plain files only: _uid (identifier of the source ECG-Reasoning-Benchmark sample, prefixed mimic_iv_ecg| or ptbxl|) and original_trace (the source's multi-stage annotation, JSON string).

The *_with_ts_tags files contain the same rows in the same order; only input differs, by the placeholder lines prepended to the prompt.

Data Sources

Subset	Source
ECG	ECG-Reasoning-Benchmark
Sleep	OpenTSLM SleepEDF-CoT
HAR	OpenTSLM HAR-CoT
TSR	Merrill et al. (2024)
Engine	EngineMT-QA

Citation

@article{chen2026lapras,
  title   = {Lapras: Latent Reasoning for Time Series Language Models},
  author  = {Chen, Yuliang and Wu, Yu Yvonne and Langer, Patrick and Pillai, Arvind and Regmi, Sudarshan and
             Maritsch, Martin and Liu, Juncheng and Jakob, Robert and Kaar, Thomas and Griffin, Tess Z. and
             Marsch, Lisa and Heinz, Michael V. and Jacobson, Nicholas C. and Campbell, Andrew},
  journal = {arXiv preprint arXiv:2610.11111},
  year    = {2026}
}