diff --git a/.devops/llama-cli-cann.Dockerfile b/.devops/llama-cli-cann.Dockerfile
index db5ba2f25..45c0585b0 100644
--- a/.devops/llama-cli-cann.Dockerfile
+++ b/.devops/llama-cli-cann.Dockerfile
@@ -1,6 +1,6 @@
 ARG ASCEND_VERSION=8.0.rc2.alpha003-910b-openeuler22.03-py3.8
 
-FROM cosdt/cann:$ASCEND_VERSION AS build
+FROM ascendai/cann:$ASCEND_VERSION AS build
 
 WORKDIR /app
 
@@ -26,7 +26,7 @@ RUN echo "Building with static libs" && \
     cmake --build build --config Release --target llama-cli
 
 # TODO: use image with NNRT
-FROM cosdt/cann:$ASCEND_VERSION AS runtime
+FROM ascendai/cann:$ASCEND_VERSION AS runtime
 COPY --from=build /app/build/bin/llama-cli /llama-cli
 
 ENV LC_ALL=C.utf8
diff --git a/docs/build.md b/docs/build.md
index 95512415a..52de2b4e2 100644
--- a/docs/build.md
+++ b/docs/build.md
@@ -375,7 +375,7 @@ cmake --build build --config release
 
 You can test with:
 
-`./build/llama-cli -m PATH_TO_MODEL -p "Building a website can be done in 10 steps:" -ngl 32`
+`./build/bin/llama-cli -m PATH_TO_MODEL -p "Building a website can be done in 10 steps:" -ngl 32`
 
 If the fllowing info is output on screen, you are using `llama.cpp by CANN backend`:
 ```bash