-
Notifications
You must be signed in to change notification settings - Fork 665
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Migrate additional examples from xgboost-operator (#1461)
Signed-off-by: terrytangyuan <terrytangyuan@gmail.com>
- Loading branch information
1 parent
06f4325
commit c41956d
Showing
29 changed files
with
3,304 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,44 @@ | ||
FROM ubuntu:16.04 | ||
|
||
ARG CONDA_DIR=/opt/conda | ||
ENV PATH $CONDA_DIR/bin:$PATH | ||
|
||
RUN apt-get update && \ | ||
apt-get install -y --no-install-recommends \ | ||
ca-certificates \ | ||
cmake \ | ||
build-essential \ | ||
gcc \ | ||
g++ \ | ||
git \ | ||
curl && \ | ||
# python environment | ||
curl -sL https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -o conda.sh && \ | ||
/bin/bash conda.sh -f -b -p $CONDA_DIR && \ | ||
export PATH="$CONDA_DIR/bin:$PATH" && \ | ||
conda config --set always_yes yes --set changeps1 no && \ | ||
# lightgbm | ||
conda install -q -y numpy==1.20.3 scipy==1.6.2 scikit-learn==0.24.2 pandas==1.3.0 && \ | ||
git clone --recursive --branch stable --depth 1 https://github.com/Microsoft/LightGBM && \ | ||
mkdir LightGBM/build && \ | ||
cd LightGBM/build && \ | ||
cmake .. && \ | ||
make -j4 && \ | ||
make install && \ | ||
cd ../python-package && \ | ||
python setup.py install_lib && \ | ||
# clean | ||
apt-get autoremove -y && apt-get clean && \ | ||
conda clean -a -y && \ | ||
rm -rf /usr/local/src/* && \ | ||
rm -rf /LightGBM | ||
|
||
WORKDIR /app | ||
|
||
# Download the example data | ||
RUN mkdir data | ||
ADD https://raw.githubusercontent.com/microsoft/LightGBM/stable/examples/parallel_learning/binary.train data/. | ||
ADD https://raw.githubusercontent.com/microsoft/LightGBM/stable/examples/parallel_learning/binary.test data/. | ||
COPY *.py ./ | ||
|
||
ENTRYPOINT [ "python", "/app/main.py" ] |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,210 @@ | ||
### Distributed Lightgbm Job train | ||
|
||
This folder containers Dockerfile and Python scripts to run a distributed Lightgbm training using the XGBoost operator. | ||
The code is based in this [example](https://github.com/microsoft/LightGBM/tree/master/examples/parallel_learning) in the official github repository of the library. | ||
|
||
|
||
**Build image** | ||
The default image name and tag is `kubeflow/lightgbm-dist-py-test:1.0` respectiveily. | ||
|
||
```shell | ||
docker build -f Dockerfile -t kubeflow/lightgbm-dist-py-test:1.0 ./ | ||
``` | ||
|
||
**Start the training** | ||
|
||
``` | ||
kubectl create -f xgboostjob_v1_lightgbm_dist_training.yaml | ||
``` | ||
|
||
**Look at the job status** | ||
``` | ||
kubectl get -o yaml XGBoostJob/lightgbm-dist-train-test | ||
``` | ||
Here is sample output when the job is running. The output result like this | ||
|
||
``` | ||
apiVersion: xgboostjob.kubeflow.org/v1 | ||
kind: XGBoostJob | ||
metadata: | ||
annotations: | ||
kubectl.kubernetes.io/last-applied-configuration: | | ||
{"apiVersion":"xgboostjob.kubeflow.org/v1","kind":"XGBoostJob","metadata":{"annotations":{},"name":"lightgbm-dist-train-test","namespace":"default"},"spec":{"xgbReplicaSpecs":{"Master":{"replicas":1,"restartPolicy":"Never","template":{"apiVersion":"v1","kind":"Pod","spec":{"containers":[{"args":["--job_type=Train","--boosting_type=gbdt","--objective=binary","--metric=binary_logloss,auc","--metric_freq=1","--is_training_metric=true","--max_bin=255","--data=data/binary.train","--valid_data=data/binary.test","--num_trees=100","--learning_rate=01","--num_leaves=63","--tree_learner=feature","--feature_fraction=0.8","--bagging_freq=5","--bagging_fraction=0.8","--min_data_in_leaf=50","--min_sum_hessian_in_leaf=50","--is_enable_sparse=true","--use_two_round_loading=false","--is_save_binary_file=false"],"image":"kubeflow/lightgbm-dist-py-test:1.0","imagePullPolicy":"Never","name":"xgboostjob","ports":[{"containerPort":9991,"name":"xgboostjob-port"}]}]}}},"Worker":{"replicas":2,"restartPolicy":"ExitCode","template":{"apiVersion":"v1","kind":"Pod","spec":{"containers":[{"args":["--job_type=Train","--boosting_type=gbdt","--objective=binary","--metric=binary_logloss,auc","--metric_freq=1","--is_training_metric=true","--max_bin=255","--data=data/binary.train","--valid_data=data/binary.test","--num_trees=100","--learning_rate=01","--num_leaves=63","--tree_learner=feature","--feature_fraction=0.8","--bagging_freq=5","--bagging_fraction=0.8","--min_data_in_leaf=50","--min_sum_hessian_in_leaf=50","--is_enable_sparse=true","--use_two_round_loading=false","--is_save_binary_file=false"],"image":"kubeflow/lightgbm-dist-py-test:1.0","imagePullPolicy":"Never","name":"xgboostjob","ports":[{"containerPort":9991,"name":"xgboostjob-port"}]}]}}}}}} | ||
creationTimestamp: "2020-10-14T15:31:23Z" | ||
generation: 7 | ||
managedFields: | ||
- apiVersion: xgboostjob.kubeflow.org/v1 | ||
fieldsType: FieldsV1 | ||
fieldsV1: | ||
f:metadata: | ||
f:annotations: | ||
.: {} | ||
f:kubectl.kubernetes.io/last-applied-configuration: {} | ||
f:spec: | ||
.: {} | ||
f:xgbReplicaSpecs: | ||
.: {} | ||
f:Master: | ||
.: {} | ||
f:replicas: {} | ||
f:restartPolicy: {} | ||
f:template: | ||
.: {} | ||
f:spec: {} | ||
f:Worker: | ||
.: {} | ||
f:replicas: {} | ||
f:restartPolicy: {} | ||
f:template: | ||
.: {} | ||
f:spec: {} | ||
manager: kubectl-client-side-apply | ||
operation: Update | ||
time: "2020-10-14T15:31:23Z" | ||
- apiVersion: xgboostjob.kubeflow.org/v1 | ||
fieldsType: FieldsV1 | ||
fieldsV1: | ||
f:spec: | ||
f:RunPolicy: | ||
.: {} | ||
f:cleanPodPolicy: {} | ||
f:xgbReplicaSpecs: | ||
f:Master: | ||
f:template: | ||
f:metadata: | ||
.: {} | ||
f:creationTimestamp: {} | ||
f:spec: | ||
f:containers: {} | ||
f:Worker: | ||
f:template: | ||
f:metadata: | ||
.: {} | ||
f:creationTimestamp: {} | ||
f:spec: | ||
f:containers: {} | ||
f:status: | ||
.: {} | ||
f:completionTime: {} | ||
f:conditions: {} | ||
f:replicaStatuses: | ||
.: {} | ||
f:Master: | ||
.: {} | ||
f:succeeded: {} | ||
f:Worker: | ||
.: {} | ||
f:succeeded: {} | ||
manager: main | ||
operation: Update | ||
time: "2020-10-14T15:34:44Z" | ||
name: lightgbm-dist-train-test | ||
namespace: default | ||
resourceVersion: "38923" | ||
selfLink: /apis/xgboostjob.kubeflow.org/v1/namespaces/default/xgboostjobs/lightgbm-dist-train-test | ||
uid: b2b887d0-445b-498b-8852-26c8edc98dc7 | ||
spec: | ||
RunPolicy: | ||
cleanPodPolicy: None | ||
xgbReplicaSpecs: | ||
Master: | ||
replicas: 1 | ||
restartPolicy: Never | ||
template: | ||
metadata: | ||
creationTimestamp: null | ||
spec: | ||
containers: | ||
- args: | ||
- --job_type=Train | ||
- --boosting_type=gbdt | ||
- --objective=binary | ||
- --metric=binary_logloss,auc | ||
- --metric_freq=1 | ||
- --is_training_metric=true | ||
- --max_bin=255 | ||
- --data=data/binary.train | ||
- --valid_data=data/binary.test | ||
- --num_trees=100 | ||
- --learning_rate=01 | ||
- --num_leaves=63 | ||
- --tree_learner=feature | ||
- --feature_fraction=0.8 | ||
- --bagging_freq=5 | ||
- --bagging_fraction=0.8 | ||
- --min_data_in_leaf=50 | ||
- --min_sum_hessian_in_leaf=50 | ||
- --is_enable_sparse=true | ||
- --use_two_round_loading=false | ||
- --is_save_binary_file=false | ||
image: kubeflow/lightgbm-dist-py-test:1.0 | ||
imagePullPolicy: Never | ||
name: xgboostjob | ||
ports: | ||
- containerPort: 9991 | ||
name: xgboostjob-port | ||
resources: {} | ||
Worker: | ||
replicas: 2 | ||
restartPolicy: ExitCode | ||
template: | ||
metadata: | ||
creationTimestamp: null | ||
spec: | ||
containers: | ||
- args: | ||
- --job_type=Train | ||
- --boosting_type=gbdt | ||
- --objective=binary | ||
- --metric=binary_logloss,auc | ||
- --metric_freq=1 | ||
- --is_training_metric=true | ||
- --max_bin=255 | ||
- --data=data/binary.train | ||
- --valid_data=data/binary.test | ||
- --num_trees=100 | ||
- --learning_rate=01 | ||
- --num_leaves=63 | ||
- --tree_learner=feature | ||
- --feature_fraction=0.8 | ||
- --bagging_freq=5 | ||
- --bagging_fraction=0.8 | ||
- --min_data_in_leaf=50 | ||
- --min_sum_hessian_in_leaf=50 | ||
- --is_enable_sparse=true | ||
- --use_two_round_loading=false | ||
- --is_save_binary_file=false | ||
image: kubeflow/lightgbm-dist-py-test:1.0 | ||
imagePullPolicy: Never | ||
name: xgboostjob | ||
ports: | ||
- containerPort: 9991 | ||
name: xgboostjob-port | ||
resources: {} | ||
status: | ||
completionTime: "2020-10-14T15:34:44Z" | ||
conditions: | ||
- lastTransitionTime: "2020-10-14T15:31:23Z" | ||
lastUpdateTime: "2020-10-14T15:31:23Z" | ||
message: xgboostJob lightgbm-dist-train-test is created. | ||
reason: XGBoostJobCreated | ||
status: "True" | ||
type: Created | ||
- lastTransitionTime: "2020-10-14T15:31:23Z" | ||
lastUpdateTime: "2020-10-14T15:31:23Z" | ||
message: XGBoostJob lightgbm-dist-train-test is running. | ||
reason: XGBoostJobRunning | ||
status: "False" | ||
type: Running | ||
- lastTransitionTime: "2020-10-14T15:34:44Z" | ||
lastUpdateTime: "2020-10-14T15:34:44Z" | ||
message: XGBoostJob lightgbm-dist-train-test is successfully completed. | ||
reason: XGBoostJobSucceeded | ||
status: "True" | ||
type: Succeeded | ||
replicaStatuses: | ||
Master: | ||
succeeded: 1 | ||
Worker: | ||
succeeded: 2 | ||
``` |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,80 @@ | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
import os | ||
import logging | ||
import argparse | ||
|
||
from train import train | ||
|
||
from utils import generate_machine_list_file, generate_train_conf_file | ||
|
||
|
||
logger = logging.getLogger(__name__) | ||
|
||
|
||
def main(args, extra_args): | ||
|
||
master_addr = os.environ["MASTER_ADDR"] | ||
master_port = os.environ["MASTER_PORT"] | ||
worker_addrs = os.environ["WORKER_ADDRS"] | ||
worker_port = os.environ["WORKER_PORT"] | ||
world_size = int(os.environ["WORLD_SIZE"]) | ||
rank = int(os.environ["RANK"]) | ||
|
||
logger.info( | ||
"extract cluster info from env variables \n" | ||
f"master_addr: {master_addr} \n" | ||
f"master_port: {master_port} \n" | ||
f"worker_addrs: {worker_addrs} \n" | ||
f"worker_port: {worker_port} \n" | ||
f"world_size: {world_size} \n" | ||
f"rank: {rank} \n" | ||
) | ||
|
||
if args.job_type == "Predict": | ||
logging.info("starting the predict job") | ||
|
||
elif args.job_type == "Train": | ||
logging.info("starting the train job") | ||
logging.info(f"extra args:\n {extra_args}") | ||
machine_list_filepath = generate_machine_list_file( | ||
master_addr, master_port, worker_addrs, worker_port | ||
) | ||
logging.info(f"machine list generated in: {machine_list_filepath}") | ||
local_port = worker_port if rank else master_port | ||
config_file = generate_train_conf_file( | ||
machine_list_file=machine_list_filepath, | ||
world_size=world_size, | ||
output_model="model.txt", | ||
local_port=local_port, | ||
extra_args=extra_args, | ||
) | ||
logging.info(f"config generated in: {config_file}") | ||
train(config_file) | ||
logging.info("Finish distributed job") | ||
|
||
|
||
if __name__ == "__main__": | ||
parser = argparse.ArgumentParser() | ||
|
||
parser.add_argument( | ||
"--job_type", | ||
help="Job type to execute", | ||
choices=["Train", "Predict"], | ||
required=True, | ||
) | ||
|
||
logging.basicConfig(format="%(message)s") | ||
logging.getLogger().setLevel(logging.INFO) | ||
args, extra_args = parser.parse_known_args() | ||
main(args, extra_args) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,26 @@ | ||
# Licensed under the Apache License, Version 2.0 (the "License"); | ||
# you may not use this file except in compliance with the License. | ||
# You may obtain a copy of the License at | ||
# | ||
# http://www.apache.org/licenses/LICENSE-2.0 | ||
# | ||
# Unless required by applicable law or agreed to in writing, software | ||
# distributed under the License is distributed on an "AS IS" BASIS, | ||
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | ||
# See the License for the specific language governing permissions and | ||
# limitations under the License. | ||
|
||
|
||
import logging | ||
import subprocess | ||
|
||
logger = logging.getLogger(__name__) | ||
|
||
|
||
def train(train_config_filepath: str): | ||
cmd = ["lightgbm", f"config={train_config_filepath}"] | ||
proc = subprocess.Popen(cmd, stdout=subprocess.PIPE) | ||
line = proc.stdout.readline() | ||
while line: | ||
logger.info((line.decode("utf-8").strip())) | ||
line = proc.stdout.readline() |
Oops, something went wrong.