Skip to content

beehive-lab/gpullama3-quarkus-langchain4j-demo

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Summarization Service Quarkus application

inspired by : https://docs.quarkiverse.io/quarkus-langchain4j/dev/quickstart-summarization.html#

Environment

  1. Java 25
sdk use java 25.0.2-open
  1. TornadoVM
sdk use tornadovm 3.0.0-jdk25-ptx
  1. Manual Clone and Build Quarkus-langchain4j to enable jdk25 support:
  • Note: As of 2/4/2026 Quarkus-langchain4j artifacts are not compiled with jdk25; hence we need to build from source.
git clone https://github.com/quarkiverse/quarkus-langchain4j.git
cd ~/quarkus-langchain4j/model-provivers/gpu-llama3
mvn clean install -DskipTests -DTornado

Build

Build all demos:

cd ~/Quarkus-Langchain4j-GPULlama3-Demos
mvn clean install

Run:

Run chat-demo:

java @$TORNADOVM_HOME/tornado-argfile \
    --add-modules jdk.incubator.vector \
    -Dtornado.device.memory=15GB \
    -jar demos/chat-demo/target/quarkus-app/quarkus-run.jar
  • with batched prefill-decode
java @$TORNADOVM_HOME/tornado-argfile \
    --add-modules jdk.incubator.vector \
    -Dtornado.device.memory=15GB \
    -Dllama.batchedPrefill=true \
    -Dllama.prefillBatchSize=32 \
    -jar demos/chat-demo/target/quarkus-app/quarkus-run.jar

Run streaming-demo:

java @$TORNADOVM_HOME/tornado-argfile \
    --add-modules jdk.incubator.vector \
    -Dtornado.device.memory=15GB \
    -jar demos/streaming-demo/target/quarkus-app/quarkus-run.jar

About

Demos for GPULlama3 integration with Quarkus-Langchain4j extension

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors