E5
Text embedding model family trained on weakly supervised pairs.
Overview
E5 is an embedding model family trained on large collections of weakly supervised text pairs, with the task described in the input so one model serves queries, passages and classification. It solves the problem that an embedding model trained on one kind of pair transfers poorly to another. It exists because describing the task in the text is a cheap way to make one model cover several retrieval settings.
Technology Graph
How E5 connects to the rest of the stack.
Alternatives
-
BGE
model
An alternative to E5 · source
-
Cohere Embed
model
An alternative to E5 · source
-
GTE
model
An alternative to E5 · source
-
Instructor
model
An alternative to E5 · source
-
Jina Embeddings
model
An alternative to E5 · source
-
Nomic Embed
model
An alternative to E5 · source
-
OpenAI Text Embedding 3
model
An alternative to E5 · source
Comparisons
Side-by-side breakdowns with the alternatives above.