Update README.md

Author: junwha

README.md

@@ -8,15 +8,28 @@ An efficient Docker image transfer system that extracts and merges necessary ima
 
 **Step 2**: You can use DockerSlice in two ways:
 
-1. **Share a Common Base Image:**
+A. **Share a Common Base Image:**
 
-- Push the [base images](#docker-base-images) to both servers.
-- Build custom images on top of the shared base images (e.g., `FROM junwha/dslice-base-torch:py3.10-torch2.4.1` at the top of the Dockerfile).
-- Save the above layers as a tar.gz file (`dslice save <image tag> `=> <image tag>.tar.gz will be created) 
-- Transfer only the new layers above the base, minimizing data transfer.
-- Load the tar.gz file at the offline server (`dslice load <image.tag>.tar.gz`)
+1. Push the [base images](#docker-base-images) to both servers.
 
-2. **Partial Updates:**
+    `dslice push junwha/dslice-base-torch:py3.10-torch2.4.1`
+
+3. Build custom images on top of the shared base images.
+   
+    Start with `FROM junwha/dslice-base-torch:py3.10-torch2.4.1` at the top of the Dockerfile
+   
+    `docker build -t <TAG> ./`
+   
+4. Save the above layers as a tar.gz file
+   
+    `dslice save <TAG>` (`<TAG>.tar.gz` will be created)
+
+5. Transfer only the new layers (`<TAG>.tar.gz`) above the base, minimizing data transfer.
+6. Load the tar.gz file at the offline server
+
+     `dslice load <TAG>.tar.gz`
+
+B. **Partial Updates:**
 
 - When a Docker image on the offline server needs an update, make an update on the image online (e.g., install a package).
 - Then, move only the updated layers to the offline server.