Skip to content

Update hmi client to use new API #237

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
syandroo merged 2 commits into from
Mar 3, 2022
Merged

Update hmi client to use new API #237

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d9008c1
update to use new stuff
syandroo Mar 2, 2022
efa83d7
reformat
syandroo Mar 3, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
60 changes: 31 additions & 29 deletions nucleus/deploy/client.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -83,6 +83,8 @@ def create_model_bundle(
self,
model_bundle_name: str,
load_predict_fn: Callable[[DeployModel_T], Callable[[Any], Any]],
env_params: Dict[str, str],
requirements: Optional[List[str]] = None,
model: Optional[DeployModel_T] = None,
load_model_fn: Optional[Callable[[], DeployModel_T]] = None,
bundle_url: Optional[str] = None,
Expand All @@ -102,6 +104,17 @@ def create_model_bundle(
load_model_fn: Function that when run, loads a model, e.g. a Pytorch module
load_predict_fn: Function that when called with model, returns a function that carries out inference
bundle_url: Only for self-hosted mode. Desired location of bundle.
requirements: A list of python package requirements, e.g.
["tensorflow==2.3.0", "tensorflow-hub==0.11.0"]. If no list has been passed, will default to the currently
imported list of packages.
env_params: A dictionary that dictates environment information e.g.
the use of pytorch or tensorflow, which cuda/cudnn versions to use.
Specifically, the dictionary should contain the following keys:
"framework_type": either "tensorflow" or "pytorch".
"pytorch_version": Version of pytorch, e.g. "1.5.1", "1.7.0", etc. Only applicable if framework_type is pytorch
"cuda_version": Version of cuda used, e.g. "11.0".
"cudnn_version" Version of cudnn used, e.g. "cudnn8-devel".
"tensorflow_version": Version of tensorflow, e.g. "2.3.0". Only applicable if framework_type is tensorflow
"""

if (model is not None and load_model_fn is not None) or (
Expand All @@ -112,6 +125,18 @@ def create_model_bundle(
)
# TODO should we try to catch when people intentionally pass both model and load_model_fn as None?

if requirements is None:
requirements_inferred = find_packages_from_imports(globals())
requirements = [
f"{key}=={value}"
for key, value in requirements_inferred.items()
]
logger.info(
"Using \n%s\n for model bundle %s",
requirements,
model_bundle_name,
)

bundle_metadata = {}
# Create bundle
if model is not None:
Expand All @@ -126,7 +151,8 @@ def create_model_bundle(
bundle_metadata["load_predict_fn"] = inspect.getsource(
load_predict_fn
)
bundle_metadata["load_model_fn"] = inspect.getsource(load_model_fn) # type: ignore
bundle_metadata["load_model_fn"] = inspect.getsource(
load_model_fn) # type: ignore
serialized_bundle = cloudpickle.dumps(bundle)

if self.is_self_hosted:
Expand All @@ -150,9 +176,11 @@ def create_model_bundle(

self.connection.post(
payload=dict(
id=model_bundle_name,
bundle_name=model_bundle_name,
location=raw_bundle_url,
bundle_metadata=bundle_metadata,
requirements=requirements,
env_params=env_params,
),
route="model_bundle",
) # TODO use return value somehow
Expand All @@ -170,8 +198,6 @@ def create_model_endpoint(
min_workers: int,
max_workers: int,
per_worker: int,
env_params: Dict[str, str],
requirements: Optional[List[str]] = None,
gpu_type: Optional[str] = None,
overwrite_existing_endpoint: bool = False,
endpoint_type: str = "async",
Expand All @@ -190,17 +216,6 @@ def create_model_endpoint(
max_workers: Maximum number of workers for model endpoint
per_worker: An autoscaling parameter. Use this to make a tradeoff between latency and costs,
a lower per_worker will mean more workers are created for a given workload
requirements: A list of python package requirements, e.g.
["tensorflow==2.3.0", "tensorflow-hub==0.11.0"]. If no list has been passed, will default to the currently
imported list of packages.
env_params: A dictionary that dictates environment information e.g.
the use of pytorch or tensorflow, which cuda/cudnn versions to use.
Specifically, the dictionary should contain the following keys:
"framework_type": either "tensorflow" or "pytorch".
"pytorch_version": Version of pytorch, e.g. "1.5.1", "1.7.0", etc. Only applicable if framework_type is pytorch
"cuda_version": Version of cuda used, e.g. "11.0".
"cudnn_version" Version of cudnn used, e.g. "cudnn8-devel".
"tensorflow_version": Version of tensorflow, e.g. "2.3.0". Only applicable if framework_type is tensorflow
gpu_type: If specifying a non-zero number of gpus, this controls the type of gpu requested. Current options are
"nvidia-tesla-t4" for NVIDIA T4s, or "nvidia-tesla-v100" for NVIDIA V100s.
overwrite_existing_endpoint: Whether or not we should overwrite existing endpoints
Expand All @@ -210,21 +225,9 @@ def create_model_endpoint(
A ModelEndpoint object that can be used to make requests to the endpoint.

"""
if requirements is None:
requirements_inferred = find_packages_from_imports(globals())
requirements = [
f"{key}=={value}"
for key, value in requirements_inferred.items()
]
logger.info(
"Using \n%s\n for model endpoint %s",
requirements,
endpoint_name,
)
# TODO test
# TODO test
payload = dict(
endpoint_name=endpoint_name,
env_params=env_params,
bundle_name=model_bundle.name,
cpus=cpus,
memory=memory,
Expand All @@ -233,7 +236,6 @@ def create_model_endpoint(
min_workers=min_workers,
max_workers=max_workers,
per_worker=per_worker,
requirements=requirements,
endpoint_type=endpoint_type,
)
if gpus == 0:
Expand Down