# OpenTSLM/OpenTSLM

OpenTSLM: Time-Series Language Models for Reasoning over Multivariate Medical Text- and Time-Series Data

Repository: https://github.com/OpenTSLM/OpenTSLM
Canonical: https://ross.abutalabs.com/products/opentslm
Homepage: https://www.opentslm.com
Language: Python
License: MIT
License Family: permissive
Last push: 2026-07-21T05:24:22+00:00

## Health v2 (maintenance only)
Score: 61/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 93, release rhythm 35, longevity 34
- inputs: {"age_days": 476, "days_push": 43, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1211, forks 113 (observed 2026-08-28T04:04:00.119395+00:00)

## What it is
OpenTSLM is a family of Time-Series Language Models that integrate time series as a native modality into pretrained LLMs (Llama, Gemma), enabling natural-language reasoning over multivariate medical text and time-series data. It is a Python library available on PyPI, released under the MIT license by researchers from Stanford, ETH Zurich, and collaborators.

## Use cases
- reason over ECG time series with natural language questions
- sleep staging from EEG readings using an LLM
- human activity recognition from accelerometer data
- generate captions and rationales for time series
- build patient-facing digital health applications with temporal reasoning
- fine-tune a language model on multivariate time-series data

## When to choose
- you need an LLM that natively understands time-series data alongside text
- you are doing research on multimodal temporal reasoning in medicine
- you want an open-source, MIT-licensed TSLM with pretrained checkpoints
- your tasks include ECG QA, sleep staging, HAR, or time-series captioning

## When to avoid
- you only need classical time-series forecasting without language reasoning
- you cannot access gated Hugging Face models like Llama or Gemma
- you need a production-ready clinical system rather than a research library
- you lack GPU resources for training or inference

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, llm-training, llm-inference, nlp
- domain: machine-learning, large-language-models, healthcare, time-series, artificial-intelligence
- platform: python, cross-platform
- tags: time-series-language-model, multimodal, medical-ai, ecg, eeg, huggingface, research, time-series-analysis, gpu

## Member repositories
- OpenTSLM/OpenTSLM (main) score 61

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:00.119395+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:18:02.571446+00:00, confidence not recorded.
  - readme: https://github.com/OpenTSLM/OpenTSLM (fetched 2026-08-28T04:04:00.119395+00:00, sha 06709942a449)
  - homepage: https://www.opentslm.com (fetched 2026-08-29T12:26:04.384764+00:00, sha 868e475a2c81)
- Data as of 2026-08-30T08:39:29.467469+00:00.