CoolFace
Modelpublic

second-state/All-MiniLM-L6-v2-Embedding-GGUF

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
25likes123kdownloads
Model Card

<!-- header start --> <!-- 200823 --> <div style="width: auto; margin-left: auto; margin-right: auto"> <img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;"> </div> <hr style="margin-top: 1.0em; margin-bottom: 1.0em;"> <!-- header end -->

All-MiniLM-L6-v2-GGUF

Original Model

sentence-transformers/all-MiniLM-L6-v2

Run with LlamaEdge

  • —LlamaEdge version: v0.8.2 and above
  • —Context size: 384
  • —Vector size: 256
  • —Run as LlamaEdge service
bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:all-MiniLM-L6-v2-ggml-model-f16.gguf \
    llama-api-server.wasm \
    --prompt-template llama-2-chat \
    --ctx-size 256 \
    --model-name all-MiniLM-L6-v2

Quantized GGUF Models

NameQuant methodBitsSizeUse case
all-MiniLM-L6-v2-Q2_K.ggufQ2_K219.2 MBsmallest, significant quality loss - not recommended for most purposes
all-MiniLM-L6-v2-Q3_K_L.ggufQ3KL320.5 MBsmall, substantial quality loss
all-MiniLM-L6-v2-Q3_K_M.ggufQ3KM319.9 MBvery small, high quality loss
all-MiniLM-L6-v2-Q3_K_S.ggufQ3KS319.2 MBvery small, high quality loss
all-MiniLM-L6-v2-Q4_0.ggufQ4_0419.7 MBlegacy; small, very high quality loss - prefer using Q3KM
all-MiniLM-L6-v2-Q4_K_M.ggufQ4KM421 MBmedium, balanced quality - recommended
all-MiniLM-L6-v2-Q4_K_S.ggufQ4KS420.7 MBsmall, greater quality loss
all-MiniLM-L6-v2-Q5_0.ggufQ5_0521 MBlegacy; medium, balanced quality - prefer using Q4KM
all-MiniLM-L6-v2-Q5_K_M.ggufQ5KM521.7 MBlarge, very low quality loss - recommended
all-MiniLM-L6-v2-Q5_K_S.ggufQ5KS521.5 MBlarge, low quality loss - recommended
all-MiniLM-L6-v2-Q6_K.ggufQ6_K624.2 MBvery large, extremely low quality loss
all-MiniLM-L6-v2-Q8_0.ggufQ8_0825 MBvery large, extremely low quality loss - not recommended
all-MiniLM-L6-v2-ggml-model-f16.ggufQ8_0845.9 MBvery large, extremely low quality loss - not recommended

Quantized with llama.cpp b2334