CoolFace
Modelpublic

second-state/Falcon3-10B-Instruct-GGUF

sourceHugging Faceotherupdated 2y agoView on Hugging Face
1likes571downloads
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 -->

Falcon3-10B-Instruct-GGUF

Original Model

tiiuae/Falcon3-10B-Instruct

Run with LlamaEdge

  • —LlamaEdge version: v0.16.0 and above
  • —Prompt template
  • —Prompt type: falcon3
  • —Prompt string
text
    <|system|>
    You are a helpful friendly assistant Falcon3 from TII, try to follow instructions as much as possible.
    <|user|>
    {user_message}
    <|assistant|>
  • —Context size: 32000
  • —Run as LlamaEdge service
bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:Falcon3-10B-Instruct-Q5_K_M.gguf \
    llama-api-server.wasm \
    --model-name Falcon3-10B-Instruct \
    --prompt-template falcon3 \
    --ctx-size 32000
  • —Run as LlamaEdge command app
bash
  wasmedge --dir .:. --nn-preload default:GGML:AUTO:Falcon3-10B-Instruct-Q5_K_M.gguf \
    llama-chat.wasm \
    --prompt-template falcon3 \
    --ctx-size 32000

Quantized GGUF Models

NameQuant methodBitsSizeUse case
Falcon3-10B-Instruct-Q2_K.ggufQ2_K23.92 GBsmallest, significant quality loss - not recommended for most purposes
Falcon3-10B-Instruct-Q3_K_L.ggufQ3KL35.45 GBsmall, substantial quality loss
Falcon3-10B-Instruct-Q3_K_M.ggufQ3KM35.05 GBvery small, high quality loss
Falcon3-10B-Instruct-Q3_K_S.ggufQ3KS34.59 GBvery small, high quality loss
Falcon3-10B-Instruct-Q4_0.ggufQ4_045.91 GBlegacy; small, very high quality loss - prefer using Q3KM
Falcon3-10B-Instruct-Q4_K_M.ggufQ4KM46.29 GBmedium, balanced quality - recommended
Falcon3-10B-Instruct-Q4_K_S.ggufQ4KS45.95 GBsmall, greater quality loss
Falcon3-10B-Instruct-Q5_0.ggufQ5_057.14 GBlegacy; medium, balanced quality - prefer using Q4KM
Falcon3-10B-Instruct-Q5_K_M.ggufQ5KM57.34 GBlarge, very low quality loss - recommended
Falcon3-10B-Instruct-Q5_K_S.ggufQ5KS57.14 GBlarge, low quality loss - recommended
Falcon3-10B-Instruct-Q6_K.ggufQ6_K68.46 GBvery large, extremely low quality loss
Falcon3-10B-Instruct-Q8_0.ggufQ8_0811.0 GBvery large, extremely low quality loss - not recommended
Falcon3-10B-Instruct-f16.gguff161620.6 GB

Quantized with llama.cpp b4381