from __future__ import print_function, with_statement, absolute_import
import shutil
import torch
import logging
import re
import os
import json
import sys
from ..version import __version__
from ..clipper_admin import ClipperException
from .deployer_utils import save_python_function, serialize_object
logger = logging.getLogger(__name__)
PYTORCH_WEIGHTS_RELATIVE_PATH = "pytorch_weights.pkl"
PYTORCH_MODEL_RELATIVE_PATH = "pytorch_model.pkl"
[docs]def create_endpoint(clipper_conn,
name,
input_type,
func,
pytorch_model,
default_output="None",
version=1,
slo_micros=3000000,
labels=None,
registry=None,
base_image="default",
num_replicas=1,
batch_size=-1,
pkgs_to_install=None):
"""Registers an app and deploys the provided predict function with PyTorch model as
a Clipper model.
Parameters
----------
clipper_conn : :py:meth:`clipper_admin.ClipperConnection`
A ``ClipperConnection`` object connected to a running Clipper cluster.
name : str
The name to be assigned to both the registered application and deployed model.
input_type : str
The input_type to be associated with the registered app and deployed model.
One of "integers", "floats", "doubles", "bytes", or "strings".
func : function
The prediction function. Any state associated with the function will be
captured via closure capture and pickled with Cloudpickle.
pytorch_model : pytorch model object
The PyTorch model to save.
default_output : str, optional
The default output for the application. The default output will be returned whenever
an application is unable to receive a response from a model within the specified
query latency SLO (service level objective). The reason the default output was returned
is always provided as part of the prediction response object. Defaults to "None".
version : str, optional
The version to assign this model. Versions must be unique on a per-model
basis, but may be re-used across different models.
slo_micros : int, optional
The query latency objective for the application in microseconds.
This is the processing latency between Clipper receiving a request
and sending a response. It does not account for network latencies
before a request is received or after a response is sent.
If Clipper cannot process a query within the latency objective,
the default output is returned. Therefore, it is recommended that
the SLO not be set aggressively low unless absolutely necessary.
100000 (100ms) is a good starting value, but the optimal latency objective
will vary depending on the application.
labels : list(str), optional
A list of strings annotating the model. These are ignored by Clipper
and used purely for user annotations.
registry : str, optional
The Docker container registry to push the freshly built model to. Note
that if you are running Clipper on Kubernetes, this registry must be accesible
to the Kubernetes cluster in order to fetch the container from the registry.
base_image : str, optional
The base Docker image to build the new model image from. This
image should contain all code necessary to run a Clipper model
container RPC client.
num_replicas : int, optional
The number of replicas of the model to create. The number of replicas
for a model can be changed at any time with
:py:meth:`clipper.ClipperConnection.set_num_replicas`.
batch_size : int, optional
The user-defined query batch size for the model. Replicas of the model will attempt
to process at most `batch_size` queries simultaneously. They may process smaller
batches if `batch_size` queries are not immediately available.
If the default value of -1 is used, Clipper will adaptively calculate the batch size for individual
replicas of this model.
pkgs_to_install : list (of strings), optional
A list of the names of packages to install, using pip, in the container.
The names must be strings.
"""
clipper_conn.register_application(name, input_type, default_output,
slo_micros)
deploy_pytorch_model(clipper_conn, name, version, input_type, func,
pytorch_model, base_image, labels, registry,
num_replicas, batch_size, pkgs_to_install)
clipper_conn.link_model_to_app(name, name)
[docs]def deploy_pytorch_model(clipper_conn,
name,
version,
input_type,
func,
pytorch_model,
base_image="default",
labels=None,
registry=None,
num_replicas=1,
batch_size=-1,
pkgs_to_install=None):
"""Deploy a Python function with a PyTorch model.
Parameters
----------
clipper_conn : :py:meth:`clipper_admin.ClipperConnection`
A ``ClipperConnection`` object connected to a running Clipper cluster.
name : str
The name to be assigned to both the registered application and deployed model.
version : str
The version to assign this model. Versions must be unique on a per-model
basis, but may be re-used across different models.
input_type : str
The input_type to be associated with the registered app and deployed model.
One of "integers", "floats", "doubles", "bytes", or "strings".
func : function
The prediction function. Any state associated with the function will be
captured via closure capture and pickled with Cloudpickle.
pytorch_model : pytorch model object
The Pytorch model to save.
base_image : str, optional
The base Docker image to build the new model image from. This
image should contain all code necessary to run a Clipper model
container RPC client.
labels : list(str), optional
A list of strings annotating the model. These are ignored by Clipper
and used purely for user annotations.
registry : str, optional
The Docker container registry to push the freshly built model to. Note
that if you are running Clipper on Kubernetes, this registry must be accesible
to the Kubernetes cluster in order to fetch the container from the registry.
num_replicas : int, optional
The number of replicas of the model to create. The number of replicas
for a model can be changed at any time with
:py:meth:`clipper.ClipperConnection.set_num_replicas`.
batch_size : int, optional
The user-defined query batch size for the model. Replicas of the model will attempt
to process at most `batch_size` queries simultaneously. They may process smaller
batches if `batch_size` queries are not immediately available.
If the default value of -1 is used, Clipper will adaptively calculate the batch size for individual
replicas of this model.
pkgs_to_install : list (of strings), optional
A list of the names of packages to install, using pip, in the container.
The names must be strings.
Example
-------
Define a pytorch nn module and save the model::
from clipper_admin import ClipperConnection, DockerContainerManager
from clipper_admin.deployers.pytorch import deploy_pytorch_model
from torch import nn
clipper_conn = ClipperConnection(DockerContainerManager())
# Connect to an already-running Clipper cluster
clipper_conn.connect()
model = nn.Linear(1, 1)
# Define a shift function to normalize prediction inputs
def predict(model, inputs):
pred = model(shift(inputs))
pred = pred.data.numpy()
return [str(x) for x in pred]
deploy_pytorch_model(
clipper_conn,
name="example",
version=1,
input_type="doubles",
func=predict,
pytorch_model=model)
"""
serialization_dir = save_python_function(name, func)
# save Torch model
torch_weights_save_loc = os.path.join(serialization_dir,
PYTORCH_WEIGHTS_RELATIVE_PATH)
torch_model_save_loc = os.path.join(serialization_dir,
PYTORCH_MODEL_RELATIVE_PATH)
try:
torch.save(pytorch_model.state_dict(), torch_weights_save_loc)
serialized_model = serialize_object(pytorch_model)
with open(torch_model_save_loc, "wb") as serialized_model_file:
serialized_model_file.write(serialized_model)
logger.info("Torch model saved")
py_minor_version = (sys.version_info.major, sys.version_info.minor)
# Check if Python 2 or Python 3 image
if base_image == "default":
if py_minor_version < (3, 0):
logger.info("Using Python 2 base image")
base_image = "clipper/pytorch-container:{}".format(__version__)
elif py_minor_version == (3, 5):
logger.info("Using Python 3.5 base image")
base_image = "clipper/pytorch35-container:{}".format(
__version__)
elif py_minor_version == (3, 6):
logger.info("Using Python 3.6 base image")
base_image = "clipper/pytorch36-container:{}".format(
__version__)
else:
msg = (
"PyTorch deployer only supports Python 2.7, 3.5, and 3.6. "
"Detected {major}.{minor}").format(
major=sys.version_info.major,
minor=sys.version_info.minor)
logger.error(msg)
# Remove temp files
shutil.rmtree(serialization_dir)
raise ClipperException(msg)
# Deploy model
clipper_conn.build_and_deploy_model(
name, version, input_type, serialization_dir, base_image, labels,
registry, num_replicas, batch_size, pkgs_to_install)
except Exception as e:
raise ClipperException("Error saving torch model: %s" % e)
# Remove temp files
shutil.rmtree(serialization_dir)