Pickling cuML Models for Persistence¶
This notebook demonstrates simple pickling of both single-GPU and multi-GPU cuML models for persistence
[1]:
import warnings
warnings.filterwarnings("ignore", category=FutureWarning)
Single GPU Model Pickling¶
All single-GPU estimators are pickleable. The following example demonstrates the creation of a synthetic dataset, training, and pickling of the resulting model for storage. Trained single-GPU models can also be used to distribute the inference on a Dask cluster, which the Distributed Model Pickling section below demonstrates.
[2]:
from cuml.datasets import make_blobs
X, y = make_blobs(n_samples=50,
n_features=10,
centers=5,
cluster_std=0.4,
random_state=0)
[3]:
from cuml.cluster import KMeans
model = KMeans(n_clusters=5)
model.fit(X)
[3]:
KMeans(handle=<cuml.raft.common.handle.Handle object at 0x7fe104514c30>, n_clusters=5, max_iter=300, tol=0.0001, verbose=4, random_state=1, init='scalable-k-means++', n_init=1, oversampling_factor=2.0, max_samples_per_batch=32768, output_type='input')
[4]:
import pickle
pickle.dump(model, open("kmeans_model.pkl", "wb"))
[5]:
model = pickle.load(open("kmeans_model.pkl", "rb"))
[6]:
model.cluster_centers_
[6]:
array([[-5.7684636 , 2.3276033 , -3.7457774 , -1.8541754 , -5.1695833 ,
7.667088 , 2.7118318 , 8.495609 , 1.7038484 , 1.1884269 ],
[ 4.647688 , 8.37788 , -9.070581 , 9.4593315 , 8.450423 ,
-1.0210547 , 3.3920872 , -7.8629856 , -0.7527662 , 0.48384118],
[-2.9414437 , 4.6401706 , -4.5027537 , 2.2855108 , 1.644645 ,
-2.4937892 , -5.2241607 , -1.5499196 , -8.063638 , 2.816936 ],
[-4.271077 , 5.561165 , -5.6640916 , -1.8229512 , -9.2925 ,
0.73028314, 4.4586773 , -2.8876226 , -5.1257744 , 9.694357 ],
[ 5.5837417 , -4.1515303 , 4.369667 , -3.0020504 , 3.638897 ,
-4.341912 , -3.318711 , 6.503671 , -6.865036 , -1.0266498 ]],
dtype=float32)
Distributed Model Pickling¶
The distributed estimator wrappers inside of the cuml.dask are not intended to be pickled directly. The Dask cuML estimators provide a function get_combined_model(), which returns the trained single-GPU model for pickling. The combined model can be used for inference on a single-GPU, and the ParallelPostFit wrapper from the Dask-ML library can be used to perform distributed inference on a Dask cluster.
[7]:
from dask.distributed import Client
from dask_cuda import LocalCUDACluster
cluster = LocalCUDACluster()
client = Client(cluster)
client
[7]:
Client
|
Cluster
|
[8]:
from cuml.dask.datasets import make_blobs
n_workers = len(client.scheduler_info()["workers"].keys())
X, y = make_blobs(n_samples=5000,
n_features=30,
centers=5,
cluster_std=0.4,
random_state=0,
n_parts=n_workers*5)
X = X.persist()
y = y.persist()
[9]:
from cuml.dask.cluster import KMeans
dist_model = KMeans(n_clusters=5)
[10]:
dist_model.fit(X)
[10]:
<cuml.dask.cluster.kmeans.KMeans at 0x7fe0fcad4b10>
[11]:
import pickle
single_gpu_model = dist_model.get_combined_model()
pickle.dump(single_gpu_model, open("kmeans_model.pkl", "wb"))
[12]:
single_gpu_model = pickle.load(open("kmeans_model.pkl", "rb"))
[13]:
single_gpu_model.cluster_centers_
[13]:
array([[ 4.809875 , 8.422671 , -9.239022 , 9.37914 , 8.499881 ,
-1.0592818 , 3.3437855 , -7.802612 , -0.5946332 , 0.264476 ,
5.5073957 , -4.10698 , 4.2890778 , -2.8172052 , 3.6150153 ,
-4.1613 , -3.6209643 , 6.2185297 , -6.9460473 , -1.0828307 ,
-5.82677 , 2.2258763 , -3.8601217 , -1.6974076 , -5.313418 ,
7.5795784 , 2.9187474 , 8.540423 , 1.5523206 , 1.0841804 ],
[-2.8941853 , 4.4741907 , -4.4475675 , 2.3820987 , 1.7478832 ,
-2.5046246 , -5.208331 , -1.6937687 , -8.134755 , 2.6468298 ,
-4.3163624 , 5.5655394 , -5.732198 , -1.7384952 , -9.344658 ,
0.7084658 , 4.4358397 , -2.9009 , -4.948638 , 9.695302 ,
8.366521 , -6.2474537 , -6.3494725 , 1.9546973 , 4.157616 ,
-9.167902 , 4.6070676 , 8.788584 , 6.864423 , 2.2319884 ],
[-4.665713 , -9.558958 , 6.657229 , 4.440131 , 2.1730306 ,
2.5904036 , 0.58000994, 6.2550354 , -8.829285 , -0.4139966 ,
9.831051 , 7.5897346 , 9.975543 , -5.8561754 , -1.2414306 ,
-2.5572667 , -1.0441563 , -5.24611 , -9.311467 , 4.636607 ,
-0.11776031, -3.929529 , 6.207367 , -7.399014 , 5.6740923 ,
-8.5403 , -7.5186524 , -5.5301213 , 4.8341303 , 2.569168 ],
[-6.9581094 , -9.760796 , -6.55061 , -0.41965044, 6.0687685 ,
3.7602885 , -3.9751325 , 6.1493387 , -1.8729935 , 5.025274 ,
-6.8340993 , 1.3383292 , 9.0016775 , -0.98648345, 9.65402 ,
9.790737 , -8.618677 , 5.995579 , 2.2099135 , -3.6309097 ,
7.0714087 , -7.394622 , -5.2996335 , -6.9737043 , -7.908465 ,
6.681064 , -5.575639 , 7.1313105 , 6.599619 , -8.309574 ],
[ 6.2617536 , 9.228769 , 8.35813 , 9.017298 , 7.704466 ,
-1.0047106 , -6.2457666 , 1.3951722 , -6.976181 , -5.9480596 ,
1.0575897 , -0.0107428 , 2.8210258 , 1.8389362 , -8.247101 ,
3.0498965 , -8.483243 , 9.72164 , -7.7502713 , 3.4655957 ,
-3.9312134 , -4.0965166 , 2.6586983 , 1.283246 , 1.0177817 ,
5.2571115 , -1.644438 , 6.1383214 , -6.8840537 , -9.663093 ]],
dtype=float32)