from clipper_admin import ClipperConnection, DockerContainerManager clipper_conn = ClipperConnection(DockerContainerManager()) clipper_conn.connect() clipper_conn.get_clipper_logs()
from clipper_admin import ClipperConnection, DockerContainerManager from clipper_admin.deployers.pytorch import deploy_pytorch_model from torch import nn clipper_conn = ClipperConnection(DockerContainerManager()) clipper_conn.connect() log_files = clipper_conn.get_clipper_logs() logs = [open(filename).read() for filename in log_files] # for log in logs: # print(log) print(log_files)
import sys sys.path.insert(0, '/Users/xunzhang/Desktop/2018/github/clipper/clipper_admin/') import tensorflow as tf from clipper_admin import ClipperConnection, DockerContainerManager from clipper_admin.deployers.tensorflow import deploy_tensorflow_model cur_dir = os.path.dirname(os.path.abspath(__file__)) app_name = "tf-lr-app" model_name = "tf-lr-model" if __name__ == "__main__": clipper_conn = ClipperConnection(DockerContainerManager()) clipper_conn.stop_all() clipper_conn.start_clipper() clipper_conn.register_application(name=app_name, input_type="integers", default_output="rabbit", slo_micros=100000) print(os.path.abspath("data")) clipper_conn.build_and_deploy_model( name=model_name, version=1, input_type="integers", model_data_path=os.path.abspath("data"), base_image="xunzhang/tf_lr_container:latest", num_replicas=1) clipper_conn.link_model_to_app(app_name, model_name) print(clipper_conn.get_clipper_logs())