fixed pipeline, remoted a step (kubeflow#768)

pipelines/azurepipeline/code/deploy/deploy.sh

@@ -17,4 +17,4 @@ while getopts "m:n:i:d:s:p:u:r:w:t:b:" option;
     esac
 done
 az login --service-principal --username ${SERVICE_PRINCIPAL_ID} --password ${SERVICE_PRINCIPAL_PASSWORD} -t $TENANT_ID
-az ml model deploy -n $MODEL_NAME -m ${MODEL}:1 --ic $INFERENCE_CONFIG --pi ${BASE_PATH}/myprofileresult.json --dc $DEPLOYMENTCONFIG -w $WORKSPACE -g $RESOURCE_GROUP --overwrite -v
+az ml model deploy -n $MODEL_NAME -m ${MODEL}:1 --ic $INFERENCE_CONFIG --dc $DEPLOYMENTCONFIG -w $WORKSPACE -g $RESOURCE_GROUP --overwrite -v

pipelines/azurepipeline/code/pipeline.py

@@ -33,7 +33,7 @@ def tacosandburritos_train(
   # preprocess data
   operations['preprocess'] = dsl.ContainerOp(
     name='preprocess',
-    image='insert your image here',
+    image='insert image name:tag',
     command=['python'],
     arguments=[
       '/scripts/data.py',
@@ -48,7 +48,7 @@ def tacosandburritos_train(
   # train
   operations['training'] = dsl.ContainerOp(
     name='training',
-    image='insert your image here',
+    image='insert image name:tag',
     command=['python'],
     arguments=[
       '/scripts/train.py',
@@ -67,7 +67,7 @@ def tacosandburritos_train(
   # register model
   operations['register'] = dsl.ContainerOp(
     name='register',
-    image='insert your image here',
+    image='insert image name:tag',
     command=['python'],
     arguments=[
       '/scripts/register.py',
@@ -84,30 +84,9 @@ def tacosandburritos_train(
   )
   operations['register'].after(operations['training'])
 
-  operations['profile'] = dsl.ContainerOp(
-    name='profile',
-    image='insert your image here',
-    command=['sh'],
-    arguments=[
-      '/scripts/profile.sh',
-      '-n', profile_name,
-      '-m', model_name,
-      '-i', '/scripts/inferenceconfig.json',
-      '-d', '{"image":"https://www.exploreveg.org/files/2015/05/sofritas-burrito.jpeg"}',
-      '-t', tenant_id,
-      '-r', resource_group,
-      '-w', workspace,
-      '-s', service_principal_id,
-      '-p', service_principal_password,
-      '-u', subscription_id,
-      '-b', persistent_volume_path
-    ]
-  )
-  operations['profile'].after(operations['register'])
-
   operations['deploy'] = dsl.ContainerOp(
     name='deploy',
-    image='insert your image here',
+    image='insert image name:tag',
     command=['sh'],
     arguments=[
       '/scripts/deploy.sh',
@@ -124,7 +103,7 @@ def tacosandburritos_train(
       '-b', persistent_volume_path
     ]
   )
-  operations['deploy'].after(operations['profile'])
+  operations['deploy'].after(operations['register'])
   for _, op_1 in operations.items():
     op_1.container.set_image_pull_policy("Always")
     op_1.add_volume(

pipelines/azurepipeline/code/preprocess/Dockerfile

@@ -1,4 +1,4 @@
-FROM tensorflow/tensorflow:2.0.0a0-gpu-py3
+FROM tensorflow/tensorflow:2.0.0a0-py3
 
 # pip install
 COPY requirements.txt /scripts/requirements.txt

pipelines/azurepipeline/code/register/register.py

@@ -82,7 +82,7 @@ def run(mdl_path, model_name, ws, tgs):
     if i == 'service_principal_password':
       print('{} => **********'.format(i))
     else:
-      print('{} => {}'.format(i, rgs[i]))
+      print('{} => {}'.format(i, wsrgs[i]))
 
   with(open(str(params_path), 'r')) as f:
     tags = json.load(f)
@@ -91,7 +91,7 @@ def run(mdl_path, model_name, ws, tgs):
   for tag in tags:
     print('{} => {}'.format(tag, tags[tag]))
 
-  rgs['tags'] = tags
+  rgs['tgs'] = tags
 
   workspc = get_ws(**wsrgs)
   rgs['ws'] = workspc

pipelines/azurepipeline/code/training/Dockerfile

@@ -1,12 +1,11 @@
-FROM tensorflow/tensorflow:2.0.0a0-gpu-py3
+FROM tensorflow/tensorflow:2.0.0a0-py3
 
 # pip install
 COPY requirements.txt /scripts/requirements.txt
 RUN pip install -r /scripts/requirements.txt
 
 COPY train.py /scripts/train.py
 
-# python train.py -d data/PetImages -e 1 -b 32 -l 0.0001 -o model -f dataset.txt
 # will be overwritten by kf pipeline
 ENTRYPOINT [ "python", \
             "/scripts/train.py", \

pipelines/azurepipeline/code/training/train.py

@@ -82,7 +82,8 @@ def run(
   # training data
   train_data, train_labels = zip(*train)
   train_ds = Dataset.zip((Dataset.from_tensor_slices(list(train_data)),
-                          Dataset.from_tensor_slices(list(train_labels)), img_size))
+                          Dataset.from_tensor_slices(list(train_labels)),
+                          Dataset.from_tensor_slices([img_size]*len(train_data))))
 
   train_ds = train_ds.map(map_func=process_image,
                           num_parallel_calls=5)