# huggingface/smollm

Everything about the SmolLM and SmolVLM family of models

Repository: https://github.com/huggingface/smollm
Canonical: https://ross.abutalabs.com/products/smollm
Homepage: https://huggingface.co/HuggingFaceTB
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-05-26T10:33:16+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 84, release rhythm 35, longevity 47
- inputs: {"age_days": 667, "days_push": 99, "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 3884, forks 310 (observed 2026-08-28T04:08:28.197474+00:00)

## What it is
Hugging Face's repository for the SmolLM and SmolVLM families of compact, fully open language and vision-language models, including training code, datasets, and local inference tools. It provides model weights, training recipes, and utilities for running small AI models on-device.

## Use cases
- run a small language model on-device
- train my own small LLM with open data
- do visual question answering with a compact vision-language model
- find a lightweight alternative to Llama 3.2 3B
- run multimodal image and text inference locally
- learn how a small LLM was trained end to end

## When to choose
- you need compact, fully open text or vision-language models with published training details
- you want on-device inference with small model sizes (135M to 3B)
- you want reproducible training recipes with public datasets

## When to avoid
- you need frontier-scale model quality
- you want a turnkey hosted API rather than self-managed model weights
- you need non-Python or non-transformers integration

## Facets
- artifact type: library
- maturity: active
- function: llm-inference, llm-training, machine-learning, deep-learning
- domain: large-language-models, machine-learning, artificial-intelligence, computer-vision
- platform: python, cross-platform
- tags: small-language-models, vision-language-models, on-device-ai, open-weights, hugging-face, multimodal, gpu

## Member repositories
- huggingface/smollm (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:28.197474+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 2026-08-29T18:25:41.069282+00:00, confidence not recorded.
  - readme: https://github.com/huggingface/smollm (fetched 2026-08-28T04:08:28.197474+00:00, sha 581c9d342fcb)
  - homepage: https://huggingface.co/HuggingFaceTB (fetched 2026-08-29T09:19:42.939778+00:00, sha d58ac4ae8566)
  - site_page: https://huggingface.co/docs (fetched 2026-08-29T09:19:42.948775+00:00, sha bdec26667b98)
  - site_page: https://huggingface.co/docs/hub/organizations-cards (fetched 2026-08-29T09:19:42.952279+00:00, sha 635040b5f704)
  - site_page: https://huggingface.co/pricing (fetched 2026-08-29T09:19:42.950507+00:00, sha de6b7a178be5)
  - site_page: https://huggingface.co/spaces/HuggingFaceTB/SmolVLM2 (fetched 2026-08-29T09:19:42.953791+00:00, sha 882d5822ec69)
  - site_page: https://huggingface.co/huggingface (fetched 2026-08-29T09:19:42.955286+00:00, sha dbcf43399b48)
- Data as of 2026-08-30T08:39:29.467469+00:00.