# github/CodeSearchNet

Datasets, tools, and benchmarks for representation learning of code.

Repository: https://github.com/github/CodeSearchNet
Canonical: https://ross.abutalabs.com/products/codesearchnet
Homepage: https://arxiv.org/abs/1909.09436
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: deep-learning, natural-language-processing, programming-language-theory, machine-learning, tensorflow, datasets, data-science, machine-learning-on-source-code, representation-learning, neural-networks, python, ml, open-data, data, cnn, rnn, self-attention, bert, nlp, nlp-machine-learning
Archived: true
Last push: 2022-01-31T09:25:07+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2743, "days_push": 1675, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2443, forks 407 (observed 2026-08-28T04:06:52.520777+00:00)

## What it is
CodeSearchNet is a dataset and benchmark suite from GitHub for semantic code search, containing about 6 million functions across six programming languages with natural language documentation, plus expert relevance annotations. It includes baseline models and training scripts for representation learning of code.

## Use cases
- train models to search code from natural language queries
- learn code embeddings for functions
- benchmark semantic code search models
- pretrain models on a large code corpus
- evaluate code retrieval with human relevance annotations
- research neural networks on source code

## When to choose
- you need a large labeled corpus of code and docstrings for research
- you want to benchmark semantic code search or code embedding models
- you need training data spanning Go, Java, JavaScript, PHP, Python, and Ruby

## When to avoid
- you need an actively maintained tool or recent framework support
- you want a production code search engine rather than research data
- you need languages outside the six supported ones

## Facets
- artifact type: dataset
- maturity: maintenance
- function: machine-learning, deep-learning, nlp, search-engine, benchmarking, data-science
- domain: machine-learning, programming-languages, big-data
- platform: python
- tags: code-search, semantic-code-search, representation-learning, code-embeddings, benchmark, corpus, tensorflow, bert, natural-language-processing, search, docker, gpu, linux

## Member repositories
- github/CodeSearchNet (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:52.520777+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 2026-08-30T02:30:36.145281+00:00, confidence not recorded.
  - readme: https://github.com/github/CodeSearchNet (fetched 2026-08-28T04:06:52.520777+00:00, sha 6482859d0c83)
  - homepage: https://arxiv.org/abs/1909.09436 (fetched 2026-08-29T10:12:08.160593+00:00, sha e0782dfeee27)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T10:12:08.169967+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T10:12:08.173450+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T10:12:08.175242+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T10:12:08.171807+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.