update getting started guide

docs/getting_started.md

@@ -3,26 +3,14 @@
 Prerequisites:
 
 - Kuda installed & cluster setup ([Installation Guide](install.md))
-- A local copy of the [`hello-gpu-flask`](/examples/hello-gpu-flask) example:
-  ```
-  git clone github.com/cyrildiagne/kuda
-  cd kuda/examples/hello-gpu-flask
-  ```
 
-## 1 - Initialize
-
-Generate the configuration files using `kuda init`:
-
-```bash
-kuda init \
-   -d gcr.io/your-gcp-project/hello-gpu \
-   hello-gpu
+First, let's get a copy of the [`hello-gpu-flask`](/examples/hello-gpu-flask) example:
+```
+git clone github.com/cyrildiagne/kuda
+cd kuda/examples/hello-gpu-flask
 ```
 
-Replace `gcr.io/your-gcp-project/hello-gpu` with a docker registry you have write
-access to.
-
-## 2 - Dev
+## 1 - Dev
 
 Run the API remotely in dev mode using:
 
@@ -34,7 +22,7 @@ Depending on your configuration, the whole process could take a while.
 But once the image has been built, pushed, provisioned, deployed & started,
 you should start seeing the startup logs from the Flask debug server.
 
-## 3 - Test
+## 2 - Test
 
 You can then call and test your dev API, for example using cURL (replace `your-domain.com` by your domain):
 
@@ -58,7 +46,7 @@ curl http://hello-gpu-dev.default.your-domain.com
 is running and the remote API should automatically synchronize & reload
 with the new changes.
 
-## 4 - Deploy
+## 3 - Deploy
 
 Once you're happy with your API, you can deploy the production build using:
 
@@ -69,5 +57,5 @@ kuda deploy
 And call the production API, for example using cURL:
 
 ```bash
-curl http://hello-gpu.default.$cluster_ip.xip.io
+curl http://hello-gpu.default.your-domain.com
 ```