Update iris fastapi and mms examples

Author: chiragjn

deploy-model-with-fastapi/README.md

@@ -1,46 +1,26 @@
-# getting-started-examples
-Examples to get started with using TrueFoundry
+# Deploy Scikit-Learn Iris flower classification model with FastAPI
 
-Deployment
 ---
-This example runs a simple iris app for inferring using a iris classifier.
-Mainly this example shows how to deploy to TrueFoundry using a Pythonfile and TrueFoundry Python SDK.
 
-## Run Locally
+### Install requirements
 
 1. Install requirements
 
 ```shell
 python -m pip install -r requirements.txt
 ```
 
-2. Start the iris app
+### Start the server
 
 ```shell
 export MODEL_DIR="$(pwd)"
 gunicorn -k uvicorn.workers.UvicornWorker --bind 0.0.0.0:8000 server:app
 ```
 
-## Deploy with TrueFoundry
-
-1. Install `truefoundry`
-
-```shell
-python -m pip install -U "truefoundry>=0.10.0,<0.11.0"
-```
-
-2. Login
-
-```shell
-tfy login --host "<Host name of TrueFoundry UI. e.g. https://company.truefoundry.cloud>"
-```
-
-3. Deploy!
-
-> Please refer to following docs
-> - [Getting workspace FQN](https://docs.truefoundry.com/docs/key-concepts#get-workspace-fqn)
-> - [Get host and path for deploying applications](https://docs.truefoundry.com/docs/define-ports-and-domains#identifying-available-domains)
+### Example Inference Call
 
 ```shell
-python deploy.py --name iris --workspace-fqn <Workspace FQN> --host <Ingress Host for the cluster> --path <optional path>
+curl -X 'POST' \
+  'http://0.0.0.0:8000/predict?sepal_length=1&sepal_width=1&petal_length=1&petal_width=1' \
+  -H 'accept: application/json'
 ```

deploy-model-with-fastapi/deploy.py

@@ -1,2 +1,3 @@
 model/
 logs/
+model_store/

deploy-model-with-mms/.tfyignore deploy-model-with-mms/.dockerignoredeploy-model-with-mms/.tfyignore renamed to deploy-model-with-mms/.dockerignore

@@ -0,0 +1 @@
+logs/

deploy-model-with-mms/.gitignore

@@ -33,10 +33,6 @@ WORKDIR /home/model-server
 COPY requirements.txt requirements.txt
 RUN python -m pip install --no-cache-dir -r requirements.txt
 
-# Here we are copying the models to the image. But for large models it is recommended to cache the models in a volume.
-# See: https://docs.truefoundry.com/docs/download-and-cache-models#download-models-and-artifacts
-COPY model_store/ model_store/
-
 COPY config.properties config.properties
 
 ENTRYPOINT ["multi-model-server"]

deploy-model-with-mms/Dockerfile

@@ -1,20 +1,28 @@
 # Deploy MNIST Model with MMS
+
 ---
 
+### Install requirements
+
+```bash
+python -m pip install -r requirements.txt
+```
+
 ### Package the model
 
 ```bash
 model-archiver --model-name mnist --model-path model/ --handler mnist_handler.py:handle --export-path model_store/ --runtime python --force
 ```
 
-### Deploy
+### Start the server
 
-```shell
-python deploy.py --workspace-fqn ... --host ... --path ...
+```bash
+export MODEL_DIR="$(pwd)/model_store"
+multi-model-server --foreground --model-store $MODEL_DIR --start --mms-config config.properties
 ```
 
 ### Example Inference Call
 
 ```bash
-curl -X POST -H "Content-Type: application/json" https://<endpoint>/predictions/mnist -T 0.png
+curl -X POST -H "Content-Type: application/json" http://0.0.0.0:8080/predictions/mnist -T 0.png
 ```

deploy-model-with-mms/README.md

@@ -11,5 +11,4 @@ load_models=ALL
 management_address=http://0.0.0.0:8081
 max_request_size=6553500
 max_response_size=6553500
-model_store=/home/model-server/model_store
 number_of_netty_threads=8