From ca2965fb9f8b6bf5bfa183b5cfc79992f4f6bdbd Mon Sep 17 00:00:00 2001
From: Ziteng Zhang <87107332+Jasonzzt@users.noreply.github.com>
Date: Tue, 24 Oct 2023 15:07:16 +0800
Subject: [PATCH] hosted k8s.png on readthedocs (#9258)

---
 docker/llm/finetune/lora/cpu/README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docker/llm/finetune/lora/cpu/README.md b/docker/llm/finetune/lora/cpu/README.md
index 8eb8e486..47ab9d24 100644
--- a/docker/llm/finetune/lora/cpu/README.md
+++ b/docker/llm/finetune/lora/cpu/README.md
@@ -6,7 +6,7 @@ By constract, BigDL here provides a CPU optimization to accelerate the lora fine
 
 The architecture is illustrated in the following:
 
-![image](https://github.com/Jasonzzt/BigDL/assets/60865256/b66416bc-ad07-49af-8cb0-8967dffb5f58)
+![image](https://llm-assets.readthedocs.io/en/latest/_images/llm-finetune-lora-cpu-k8s.png)
 
 As above, BigDL implements its MPI training with [Kubeflow MPI operator](https://github.com/kubeflow/mpi-operator/tree/master), which encapsulates the deployment as MPIJob CRD, and assists users to handle the construction of a MPI worker cluster on Kubernetes, such as public key distribution, SSH connection, and log collection.