Dataset · sentence-transformers
Millions of comment-code pairs for embeddings
sentence-transformers/codesearchnet
1 to 10 million pairs of text descriptions and code snippets across various programming languages.
What's inside
A large text corpus from the CodeSearchNet collection containing 1 to 10 million records. It includes pairs of string values: an English text comment (comment column) and its corresponding code snippet in various programming languages (code column). The dataset is prepared for direct use with the Sentence Transformers library.
What you can do with it
- Training embedding models for semantic codebase search
- Training models to match text queries with code snippets
- Tasks
- Embeddings, Sentence similarity
- Languages
- English
- Size
- 1M–10M rows
- Downloads this month
- 587
- Likes
- 20
Models trained on it
- ig1/Qwen3-Coder-30B-A3B-Instruct-NVFP4Text generation 828
- isuruwijesiri/all-MiniLM-L6-v2-code-search-512Sentence similarity 298
- Kervyn-AI/Kervyn-Rerank-100M-PreviewText ranking 81
- Kervyn-AI/Kervyn-Embeddings-100M-PreviewSentence similarity 79
- thealper2/t5-base-code-summarizationText generation 65
- sarthak1/codemaltEmbeddings 60
- ctrltokyo/ColBERT-Zero-6L-CodeSearchSentence similarity 38
- isuruwijesiri/all-MiniLM-L12-v2-code-search-512Sentence similarity 35
How to load it
from datasets import load_dataset
ds = load_dataset("sentence-transformers/codesearchnet")