triton-gpu-oke

adding an entire walkthrough to deploy Triton inference server with TensorRT-LLM Backend on a GPU node within the OKE service.

Author: bruno-garbaccio

cloud-infrastructure/ai-infra-gpu/GPU/triton-gpu-oke/Dockerfile

@@ -0,0 +1,8 @@
+FROM <region-key>.ocir.io/<tenancy-namespace>/triton_llm:triton_trt_llm_23.12_manual_build
+
+COPY  output_llama_hf /app/output_llama_hf
+
+COPY model_repo /app/model_repo
+
+RUN mkdir /app/cache
+RUN chmod 777 /app/cache