UMAP#

class cuml.dask.manifold.UMAP(*, model, client=None, **kwargs)[source]#

Uniform Manifold Approximation and Projection

Finds a low dimensional embedding of the data that approximates an underlying manifold.

Important: This Dask wrapper is designed exclusively for distributed inference; you must first train a cuml.UMAP model on a single GPU and then provide the trained model to this wrapper for distributed transform operations. Distributed training is not supported.

Parameters:

modelcuml.UMAP, required: A fitted single-GPU UMAP model instance. The model must be trained before passing it to this wrapper.
clientdask.distributed.Client, optional: Dask client to use
Adapted from https://github.com/lmcinnes/umap/blob/master/umap/umap_.py

Methods

transform(X[, convert_dtype])

Transform X into the existing embedded space and return that transformed output.

Notes

The single-GPU cuml.UMAP module is heavily based on Leland McInnes’ reference UMAP package [1].

References

[1]

Leland McInnes, John Healy, James Melville UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction.

Examples

>>> from dask_cuda import LocalCUDACluster
>>> from dask.distributed import Client
>>> import dask.array as da
>>> from cuml.datasets import make_blobs
>>> from cuml.manifold import UMAP
>>> from cuml.dask.manifold import UMAP as MNMG_UMAP
>>> import numpy as np

>>> cluster = LocalCUDACluster(threads_per_worker=1)
>>> client = Client(cluster)

>>> X, y = make_blobs(1000, 10, centers=42, cluster_std=0.1,
...                   dtype=np.float32, random_state=10)

>>> local_model = UMAP(random_state=10, verbose=0)

>>> selection = np.random.RandomState(10).choice(1000, 100)
>>> X_train = X[selection]
>>> y_train = y[selection]
>>> local_model.fit(X_train, y=y_train)
UMAP()

>>> distributed_model = MNMG_UMAP(model=local_model)
>>> distributed_X = da.from_array(X, chunks=(500, -1))
>>> embedding = distributed_model.transform(distributed_X)
>>> result = embedding.compute()
>>> print(result)
[[  4.1684933    4.1890593 ]
[  5.0110254   -5.2143383 ]
[  1.7776365  -17.665699  ]
...
[ -6.6378727   -0.15353012]
[ -3.1891193   -0.83906937]
[ -0.5042019    2.1454725 ]]
>>> client.close()
>>> cluster.close()

transform(X, convert_dtype=True)[source]#

Transform X into the existing embedded space and return that transformed output.

Please refer to the reference UMAP implementation for information on the differences between fit_transform() and running fit() transform().

Specifically, the transform() function is stochastic: https://github.com/lmcinnes/umap/issues/158

Parameters:

Xarray-like (device or host) shape = (n_samples, n_features): New data to be transformed. Acceptable formats: dask cuDF, dask CuPy/NumPy/Numba Array

Returns:

X_newarray, shape (n_samples, n_components): Embedding of the new data in low-dimensional space.