def to_cugraph(edge_index: Tensor,
edge_weight: Optional[Tensor] = None,
relabel_nodes: bool = True):
r"""Converts a graph given by :obj:`edge_index` and optional
:obj:`edge_weight` into a :obj:`cugraph` graph object.
Args:
relabel_nodes (bool, optional): If set to :obj:`True`,
:obj:`cugraph` will remove any isolated nodes, leading to a
relabeling of nodes. (default: :obj:`True`)
"""
import cudf
import cugraph
df = cudf.from_dlpack(to_dlpack(edge_index.t()))
if edge_weight is not None:
assert edge_weight.dim() == 1
df[2] = cudf.from_dlpack(to_dlpack(edge_weight))
return cugraph.from_cudf_edgelist(
df,
source=0,
destination=1,
edge_attr=2 if edge_weight is not None else None,
renumber=relabel_nodes)
def test_from_dlpack_cupy_2d(data_2d):
cupy_array = cupy.array(data_2d, order="F")
cupy_host_array = cupy_array.get().flatten()
dlt = cupy_array.toDlpack()
gdf = cudf.from_dlpack(dlt)
cudf_host_array = np.array(gdf.to_pandas()).flatten()
assert_eq(cudf_host_array, cupy_host_array)
def test_from_dlpack_cupy_1d(data_1d):
cupy_array = cupy.array(data_1d)
cupy_host_array = cupy_array.get()
dlt = cupy_array.toDlpack()
gs = cudf.from_dlpack(dlt)
cudf_host_array = gs.to_array(fillna="pandas")
assert_eq(cudf_host_array, cupy_host_array)
def test_from_dlpack_cupy_1d(data_1d):
cupy_array = cupy.array(data_1d)
cupy_host_array = cupy_array.get()
dlt = cupy_array.toDlpack()
gs = cudf.from_dlpack(dlt)
cudf_host_array = gs.to_numpy(na_value=np.nan)
assert_eq(cudf_host_array, cupy_host_array)
def KMeans(data, cluster):
warnings.filterwarnings('ignore')
data_pack = torch.utils.dlpack.to_dlpack(data)
data_df = cudf.from_dlpack(data_pack)
model = cuml.KMeans(n_clusters=cluster)
result = model.fit(data_df)
labels = torch.utils.dlpack.from_dlpack(result.labels_.to_dlpack())
warnings.filterwarnings('once')
return labels
def test_from_dlpack_cupy_2d(data_2d):
expectation = data_size_expectation_builder(data_2d)
with expectation:
cupy_array = cupy.array(data_2d, order="F")
cupy_host_array = cupy_array.get().flatten()
dlt = cupy_array.toDlpack()
gdf = cudf.from_dlpack(dlt)
cudf_host_array = np.array(gdf.to_pandas()).flatten()
assert_eq(cudf_host_array, cupy_host_array)
def test_from_dlpack_cupy_1d(data_1d):
expectation = data_size_expectation_builder(data_1d)
with expectation:
cupy_array = cupy.array(data_1d)
cupy_host_array = cupy_array.get()
dlt = cupy_array.toDlpack()
gs = cudf.from_dlpack(dlt)
cudf_host_array = gs.to_array(fillna="pandas")
assert_eq(cudf_host_array, cupy_host_array)
vertex = cp.fromDlpack(vertex.to_dlpack())
partition = cp.fromDlpack(partition.to_dlpack())
vertex = cp.reshape(vertex, XYZ_C.shape[0])
labelRE = cp.reshape(partition, XYZ_C.shape[0])
index = cp.argsort(vertex)
vertex = cp.take_along_axis(vertex, index, axis=0)
labelRE = cp.take_along_axis(labelRE, index, axis=0)
print(result)
print(vertex)
print(labelRE)
print(index)
louvain_endtime = datetime.datetime.now()
print(louvain_endtime - louvain_starttime)
print(labelRE)
labelRE = cp.asfortranarray(labelRE)
labelRE = cd.from_dlpack(labelRE.toDlpack())
#labelRE = cd.DataFrame(labelRE)
print(labelRE)
df = cd.DataFrame({'label': labelRE})
df.to_csv('HW_AI/HW_Final/gpu.csv')
'''
plt.figure(1)
ax = plt.axes(projection='3d')
z = XYZ_C[:,2]
x = XYZ_C[:,0]
y = XYZ_C[:,1]
c = labelRE
ax.scatter(x, y, z, c = c, cmap = plt.get_cmap('jet'))
plt.title('Cluster result by modularity')
ax.view_init(80, 0)
def f(x):
return cudf.from_dlpack(x.toDlpack())
