archive model

.gitignore

@@ -65,4 +65,4 @@ dev.csv
 .venv/
 model.h5
 evaluation.png
-mlruns/*
+mlruns*

Dockerfile

@@ -10,3 +10,4 @@ COPY ["Zadanie 1.py", "."]
 COPY ["stats.py", "."]
 COPY ["train.py", "."]
 COPY ["evaluate.py", "."]
+COPY ["MLProject", "."]

Jenkinsfile-train

@@ -1,6 +1,6 @@
 pipeline {
     agent {
-        docker { image 'adnovac/ium_s434760:1.2' }
+        docker { image 'adnovac/ium_s434760:3' }
     }
     parameters{
         buildSelector(
@@ -39,6 +39,7 @@ pipeline {
             stage('archive artifacts') {
                 steps {
                     archiveArtifacts 'model.h5'
+                    archiveArtifacts 'fifa_overall/**/*'
                 }
             }
             stage('send email') {

MLProject

@@ -1,11 +1,11 @@
-name: Fifa Players
+name: fifa_players
+
 docker_env:
-   image:  docker.io/adnovac/ium_s434760:2.0 
+   image: adnovac/ium_s434760
+
 entry_points:
   train:
     parameters:
       batch_size: {type: int, default: 15}
       epochs: {type: int, default: 16}
-    command: "python train.py {batch_size} {epochs}"
-  evaluate:
-    command: "python evaluate.py"
+    command: "python3.8 train.py {batch_size} {epochs}"

fifa_overall/MLmodel

@@ -0,0 +1,20 @@
+flavors:
+  keras:
+    data: data
+    keras_module: tensorflow.keras
+    keras_version: 2.4.0
+    save_format: tf
+  python_function:
+    data: data
+    env: conda.yaml
+    loader_module: mlflow.keras
+    python_version: 3.8.10
+saved_input_example_info:
+  artifact_path: input_example.json
+  pandas_orient: split
+  type: dataframe
+signature:
+  inputs: '[{"type": "tensor", "tensor-spec": {"dtype": "float64", "shape": [-1, 4]}}]'
+  outputs: '[{"type": "tensor", "tensor-spec": {"dtype": "float32", "shape": [-1,
+    1]}}]'
+utc_time_created: '2021-05-23 19:48:30.385262'

fifa_overall/conda.yaml

@@ -0,0 +1,10 @@
+channels:
+- defaults
+- conda-forge
+dependencies:
+- python=3.8.10
+- pip
+- pip:
+  - mlflow
+  - tensorflow==2.4.1
+name: mlflow-env

fifa_overall/input_example.json

@@ -0,0 +1 @@
+{"columns": ["Age", "Nationality", "Position", "Club"], "data": [[0.38, 0.175, 1, 0.91846154]]}

train.py

@@ -37,5 +37,12 @@ history = model.fit(
     batch_size=batch_size,
     epochs=epochs,
 )
-
 model.save(model_name)
+signature = mlflow.models.signature.infer_signature(X, model.predict(X))
+input_example = {
+    "Age": 0.38,
+    "Nationality": 0.175,
+    "Position": 1,
+    "Club" :0.91846154
+}
+mlflow.keras.save_model(model, "fifa_overall", signature=signature, input_example = input_example)