Skip to content

Commit ea66c1a

Browse files
author
Yehudit Kerido
committed
sdk tests with papermill
Signed-off-by: Yehudit Kerido <[email protected]>
1 parent 8ef3732 commit ea66c1a

File tree

3 files changed

+66
-174
lines changed

3 files changed

+66
-174
lines changed

‎examples/v1beta1/sdk/.ipynb_checkpoints/tune-train-from-func-checkpoint.ipynb

+29-15
Original file line numberDiff line numberDiff line change
@@ -36,9 +36,16 @@
3636
{
3737
"cell_type": "code",
3838
"execution_count": null,
39-
"id": "1d71674a-baca-4a89-adf0-4f702ac27904",
39+
"id": "1fedadf1",
4040
"metadata": {
41+
"collapsed": false,
4142
"editable": true,
43+
"jupyter": {
44+
"outputs_hidden": false
45+
},
46+
"pycharm": {
47+
"name": "#%%\n"
48+
},
4249
"slideshow": {
4350
"slide_type": ""
4451
},
@@ -48,22 +55,22 @@
4855
},
4956
"outputs": [],
5057
"source": [
51-
"katib_python_sdk = \"kubeflow-katib\""
58+
"katib_python_sdk = \"kubeflow-katib\"\n",
59+
"namespace = \"default\" "
5260
]
5361
},
5462
{
5563
"cell_type": "code",
5664
"execution_count": null,
57-
"id": "5de885ca-e96a-4d59-9e78-75f6fc6f5ce7",
65+
"id": "518cfe2d",
5866
"metadata": {
59-
"editable": true,
67+
"collapsed": false,
68+
"jupyter": {
69+
"outputs_hidden": false
70+
},
6071
"pycharm": {
6172
"name": "#%%\n"
62-
},
63-
"slideshow": {
64-
"slide_type": ""
65-
},
66-
"tags": []
73+
}
6774
},
6875
"outputs": [],
6976
"source": [
@@ -75,32 +82,39 @@
7582
{
7683
"cell_type": "code",
7784
"execution_count": null,
78-
"id": "15ea90f2-ad99-4bf4-ae07-cc35bcbe6884",
85+
"id": "7211d9e4",
7986
"metadata": {
87+
"collapsed": false,
8088
"editable": true,
89+
"jupyter": {
90+
"outputs_hidden": false
91+
},
8192
"pycharm": {
8293
"name": "#%%\n"
8394
},
8495
"slideshow": {
8596
"slide_type": ""
8697
},
87-
"tags": [
88-
"parameters"
89-
]
98+
"tags": []
9099
},
91100
"outputs": [],
92101
"source": [
93-
"# Experiment namespace\n",
94-
"namespace = \"default\" "
102+
"!pip install {katib_python_sdk}\n",
103+
"!pip install tensorflow\n",
104+
"!pip install kubeflow-training"
95105
]
96106
},
97107
{
98108
"cell_type": "markdown",
99109
"id": "881aae2f-d08e-4439-bef9-1684ff87556d",
100110
"metadata": {
111+
"editable": true,
101112
"pycharm": {
102113
"name": "#%% md\n"
103114
},
115+
"slideshow": {
116+
"slide_type": ""
117+
},
104118
"tags": []
105119
},
106120
"source": [

‎examples/v1beta1/sdk/tune-train-from-func.ipynb

+36-158
Original file line numberDiff line numberDiff line change
@@ -36,46 +36,41 @@
3636
{
3737
"cell_type": "code",
3838
"execution_count": null,
39-
"outputs": [],
40-
"source": [
41-
"katib_python_sdk = \"kubeflow-katib\""
42-
],
39+
"id": "1fedadf1",
4340
"metadata": {
4441
"collapsed": false,
42+
"editable": true,
43+
"jupyter": {
44+
"outputs_hidden": false
45+
},
4546
"pycharm": {
4647
"name": "#%%\n"
47-
}
48-
}
49-
},
50-
{
51-
"cell_type": "code",
52-
"execution_count": null,
48+
},
49+
"slideshow": {
50+
"slide_type": ""
51+
},
52+
"tags": [
53+
"parameters"
54+
]
55+
},
5356
"outputs": [],
5457
"source": [
55-
"!pip install {katib_python_sdk}\n",
56-
"!pip install tensorflow\n",
57-
"!pip install kubeflow-training"
58-
],
59-
"metadata": {
60-
"collapsed": false,
61-
"pycharm": {
62-
"name": "#%%\n"
63-
}
64-
}
58+
"katib_python_sdk = \"kubeflow-katib\"\n",
59+
"namespace = \"default\" "
60+
]
6561
},
6662
{
6763
"cell_type": "code",
6864
"execution_count": null,
69-
"id": "5de885ca-e96a-4d59-9e78-75f6fc6f5ce7",
65+
"id": "518cfe2d",
7066
"metadata": {
71-
"editable": true,
67+
"collapsed": false,
68+
"jupyter": {
69+
"outputs_hidden": false
70+
},
7271
"pycharm": {
7372
"name": "#%%\n"
74-
},
75-
"slideshow": {
76-
"slide_type": ""
77-
},
78-
"tags": []
73+
}
7974
},
8075
"outputs": [],
8176
"source": [
@@ -87,158 +82,41 @@
8782
{
8883
"cell_type": "code",
8984
"execution_count": null,
90-
"id": "15ea90f2-ad99-4bf4-ae07-cc35bcbe6884",
85+
"id": "7211d9e4",
9186
"metadata": {
87+
"collapsed": false,
9288
"editable": true,
89+
"jupyter": {
90+
"outputs_hidden": false
91+
},
9392
"pycharm": {
9493
"name": "#%%\n"
9594
},
9695
"slideshow": {
9796
"slide_type": ""
9897
},
99-
"tags": [
100-
"parameters"
101-
]
98+
"tags": []
10299
},
103100
"outputs": [],
104101
"source": [
105-
"# Experiment namespace\n",
106-
"namespace = \"default\" "
102+
"!pip install {katib_python_sdk}\n",
103+
"!pip install tensorflow\n",
104+
"!pip install kubeflow-training"
107105
]
108106
},
109107
{
110108
"cell_type": "markdown",
111-
"id": "881aae2f-d08e-4439-bef9-1684ff87556d",
112-
"metadata": {
113-
"pycharm": {
114-
"name": "#%% md\n"
115-
},
116-
"tags": []
117-
},
118109
"source": [
119110
"## Create Train Script for CNN Model\n",
120111
"\n",
121112
"This is simple **Convolutional Neural Network (CNN)** model for recognizing hand-written digits using [MNIST Dataset](http://yann.lecun.com/exdb/mnist/). "
122-
]
123-
},
124-
{
125-
"cell_type": "code",
126-
"execution_count": 1,
127-
"id": "98a77484-cee4-4d77-bd76-915c469a242b",
113+
],
128114
"metadata": {
129-
"editable": true,
130-
"execution": {
131-
"iopub.execute_input": "2022-09-12T18:46:59.051290Z",
132-
"iopub.status.busy": "2022-09-12T18:46:59.050544Z",
133-
"iopub.status.idle": "2022-09-12T18:46:59.249456Z",
134-
"shell.execute_reply": "2022-09-12T18:46:59.248292Z",
135-
"shell.execute_reply.started": "2022-09-12T18:46:59.051211Z"
136-
},
115+
"collapsed": false,
137116
"pycharm": {
138-
"name": "#%%\n"
139-
},
140-
"slideshow": {
141-
"slide_type": ""
142-
},
143-
"tags": []
144-
},
145-
"outputs": [],
146-
"source": [
147-
"def train_mnist_model(parameters):\n",
148-
" import tensorflow as tf\n",
149-
" import numpy as np\n",
150-
" import logging\n",
151-
"\n",
152-
" logging.basicConfig(\n",
153-
" format=\"%(asctime)s %(levelname)-8s %(message)s\",\n",
154-
" datefmt=\"%Y-%m-%dT%H:%M:%SZ\",\n",
155-
" level=logging.INFO,\n",
156-
" )\n",
157-
" logging.info(\"--------------------------------------------------------------------------------------\")\n",
158-
" logging.info(f\"Input Parameters: {parameters}\")\n",
159-
" logging.info(\"--------------------------------------------------------------------------------------\\n\\n\")\n",
160-
"\n",
161-
"\n",
162-
" # Get HyperParameters from the input params dict.\n",
163-
" lr = float(parameters[\"lr\"])\n",
164-
" num_epoch = int(parameters[\"num_epoch\"])\n",
165-
"\n",
166-
" # Set dist parameters and strategy.\n",
167-
" is_dist = parameters[\"is_dist\"]\n",
168-
" num_workers = parameters[\"num_workers\"]\n",
169-
" batch_size_per_worker = 64\n",
170-
" batch_size_global = batch_size_per_worker * num_workers\n",
171-
" strategy = tf.distribute.MultiWorkerMirroredStrategy(\n",
172-
" communication_options=tf.distribute.experimental.CommunicationOptions(\n",
173-
" implementation=tf.distribute.experimental.CollectiveCommunication.RING\n",
174-
" )\n",
175-
" )\n",
176-
"\n",
177-
" # Callback class for logging training.\n",
178-
" # Katib parses metrics in this format: <metric-name>=<metric-value>.\n",
179-
" class CustomCallback(tf.keras.callbacks.Callback):\n",
180-
" def on_epoch_end(self, epoch, logs=None):\n",
181-
" logging.info(\n",
182-
" \"Epoch {}/{}. accuracy={:.4f} - loss={:.4f}\".format(\n",
183-
" epoch+1, num_epoch, logs[\"accuracy\"], logs[\"loss\"]\n",
184-
" )\n",
185-
" )\n",
186-
"\n",
187-
" # Prepare MNIST Dataset.\n",
188-
" def mnist_dataset(batch_size):\n",
189-
" (x_train, y_train), _ = tf.keras.datasets.mnist.load_data()\n",
190-
" x_train = x_train / np.float32(255)\n",
191-
" y_train = y_train.astype(np.int64)\n",
192-
" train_dataset = (\n",
193-
" tf.data.Dataset.from_tensor_slices((x_train, y_train))\n",
194-
" .shuffle(60000)\n",
195-
" .repeat()\n",
196-
" .batch(batch_size)\n",
197-
" )\n",
198-
" return train_dataset\n",
199-
"\n",
200-
" # Build and compile CNN Model.\n",
201-
" def build_and_compile_cnn_model():\n",
202-
" model = tf.keras.Sequential(\n",
203-
" [\n",
204-
" tf.keras.layers.InputLayer(input_shape=(28, 28)),\n",
205-
" tf.keras.layers.Reshape(target_shape=(28, 28, 1)),\n",
206-
" tf.keras.layers.Conv2D(32, 3, activation=\"relu\"),\n",
207-
" tf.keras.layers.Flatten(),\n",
208-
" tf.keras.layers.Dense(128, activation=\"relu\"),\n",
209-
" tf.keras.layers.Dense(10),\n",
210-
" ]\n",
211-
" )\n",
212-
" model.compile(\n",
213-
" loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
214-
" optimizer=tf.keras.optimizers.SGD(learning_rate=lr),\n",
215-
" metrics=[\"accuracy\"],\n",
216-
" )\n",
217-
" return model\n",
218-
" \n",
219-
" # Download Dataset.\n",
220-
" dataset = mnist_dataset(batch_size_global)\n",
221-
"\n",
222-
" # For dist strategy we should build model under scope().\n",
223-
" if is_dist:\n",
224-
" logging.info(\"Running Distributed Training\")\n",
225-
" logging.info(\"--------------------------------------------------------------------------------------\\n\\n\")\n",
226-
" with strategy.scope():\n",
227-
" model = build_and_compile_cnn_model()\n",
228-
" else:\n",
229-
" logging.info(\"Running Single Worker Training\")\n",
230-
" logging.info(\"--------------------------------------------------------------------------------------\\n\\n\")\n",
231-
" model = build_and_compile_cnn_model()\n",
232-
" \n",
233-
" # Start Training.\n",
234-
" model.fit(\n",
235-
" dataset,\n",
236-
" epochs=num_epoch,\n",
237-
" steps_per_epoch=70,\n",
238-
" callbacks=[CustomCallback()],\n",
239-
" verbose=0,\n",
240-
" )"
241-
]
117+
"name": "#%% md\n"
118+
}
119+
}
242120
},
243121
{
244122
"cell_type": "markdown",

‎test/e2e/v1beta1/scripts/gh-actions/run-notebook.sh

+1-1
Original file line numberDiff line numberDiff line change
@@ -23,7 +23,7 @@ set -o pipefail
2323
NOTEBOOK_INPUT=""
2424
NOTEBOOK_OUTPUT="-" # outputs to console
2525
NAMESPACE="default"
26-
KATIB_PYTHON_SDK="../../../sdk/python"
26+
KATIB_PYTHON_SDK="./sdk/python"
2727

2828
usage() {
2929
echo "Usage: $0 -i <input_notebook> -o <output_notebook> [-p \"<param> <value>\"...] [-y <params.yaml>]"

0 commit comments

Comments
 (0)