Overview
all-MiniLM-L6-v2 maps sentences to 384-dimensional vectors. Its speed and small size made it a default choice for prototypes and CPU-only search.
Highlights
- About 22M parameters
- 384-dimensional embeddings
- Very fast on CPUs
- Apache 2.0 license
Typical uses
Prototyping, small search engines and clustering on modest hardware.
Things to consider
- Lower retrieval quality than modern embedding models.
- Inputs truncated at 256 word pieces by default.
- Download sizes refer to the default 4-bit quantized versions in the Ollama library; full-precision weights are larger.
Specifications are taken from the model card on Hugging Face, the Ollama library and the provider’s announcement. Benchmark results depend on evaluation settings; see how AI benchmarks work.