def test_sandbox():
sy.create_sandbox(globals(), download_data=True)
assert alice == alice # noqa: F821
assert isinstance(alice, sy.VirtualWorker) # noqa: F821
assert andy == andy # noqa: F821
assert isinstance(andy, sy.VirtualWorker) # noqa: F821
assert bob == bob # noqa: F821
assert isinstance(bob, sy.VirtualWorker) # noqa: F821
assert jason == jason # noqa: F821
assert isinstance(jason, sy.VirtualWorker) # noqa: F821
assert jon == jon # noqa: F821
assert isinstance(jon, sy.VirtualWorker) # noqa: F821
assert theo == theo # noqa: F821
assert isinstance(theo, sy.VirtualWorker) # noqa: F821
assert hook == hook # noqa: F821
assert isinstance(hook, TorchHook) # noqa: F821
assert grid == grid # noqa: F821
assert isinstance(grid, PrivateGridNetwork) # noqa: F821
assert workers == [bob, theo, jason, alice, andy, jon] # noqa: F821
assert bob.search(["#boston"]) # noqa: F821
assert bob.search(["#diabetes"]) # noqa: F821
assert bob.search(["#breast_cancer_dataset"]) # noqa: F821
assert bob.search(["#digits_dataset"]) # noqa: F821
assert bob.search(["#iris_dataset"]) # noqa: F821
assert bob.search(["#wine_dataset"]) # noqa: F821
assert bob.search(["#linnerrud_dataset"]) # noqa: F821
try:
from tensorflow.keras import datasets # noqa: F401
assert bob.search(["#fashion_mnist"]) # noqa: F821
assert bob.search(["#cifar10"]) # noqa: F821
except ImportError: # pragma: no cover
pass # pragma: no cover
def test_sandbox():
sy.create_sandbox(globals(), download_data=False)
assert alice == alice # noqa: F821
assert isinstance(alice, sy.VirtualWorker) # noqa: F821
assert andy == andy # noqa: F821
assert isinstance(andy, sy.VirtualWorker) # noqa: F821
assert bob == bob # noqa: F821
assert isinstance(bob, sy.VirtualWorker) # noqa: F821
assert jason == jason # noqa: F821
assert isinstance(jason, sy.VirtualWorker) # noqa: F821
assert jon == jon # noqa: F821
assert isinstance(jon, sy.VirtualWorker) # noqa: F821
assert theo == theo # noqa: F821
assert isinstance(theo, sy.VirtualWorker) # noqa: F821
assert hook == hook # noqa: F821
assert isinstance(hook, TorchHook) # noqa: F821
assert grid == grid # noqa: F821
assert isinstance(grid, PrivateGridNetwork) # noqa: F821
assert workers == [bob, theo, jason, alice, andy, jon] # noqa: F821
def test_sandbox():
sy.create_sandbox(globals(), download_data=False)
# check to make sure global variable gets set for alice
assert alice == alice # noqa: F821
assert isinstance(alice, sy.VirtualWorker) # noqa: F821
import syft as sy
import torch as th
sy.create_sandbox(globals(), verbose=False)
device = th.device("cuda")
"""Then search for a dataset"""
boston_data = grid.search("#boston", "#data")
boston_target = grid.search("#boston", "#target")
"""We load a model and an optimizer"""
n_features = boston_data['alice'][0].shape[1]
n_targets = 1
model = th.nn.Linear(n_features, n_targets).to(device)
"""Here we cast the data fetched in a `FederatedDataset`. See the workers which hold part of the data."""
print("A total of {} workers".format(len(boston_data.keys())))
# Cast the result in BaseDatasets
datasets = []
for worker in boston_data.keys():
dataset = sy.BaseDataset(boston_data[worker][0], boston_target[worker][0])
datasets.append(dataset)
# Build the FederatedDataset object
dataset = sy.FederatedDataset(datasets)
print(dataset.workers)
optimizers = {}
for worker in dataset.workers:
optimizers[worker] = th.optim.Adam(params=model.parameters(), lr=1e-2)
# optimizers[worker] = th.optim.SGD(model.parameters(), lr=0.05)
from flask import Flask
from flask import request
import torch as th
import syft as sy
sy.create_sandbox(globals())
app = Flask(__name__)
# Iniitalize A Toy Model
model = th.zeros([2, 1])
ptr = None
@app.route("/get_model", methods=["GET"])
def get_model():
global model
global ptr
ptr = model.create_pointer()
return model.ser()
@app.route("/send_data", methods=["POST"])
def send_data():
global ptr
ptr = sy.serde.deserialize(request.data)
return model.ser()
"_protocol": str(protocolPb),
"jason": str(simplifiedJasonPlan),
"_jason": str(jasonPlanPb),
"andy": str(simplifiedAndyPlan),
"_andy": str(andyPlanPb),
}
def serializeToBase64Pb(worker, obj):
pb = protobuf.serde._bufferize(worker, obj)
bin = pb.SerializeToString()
return pb, base64.b64encode(bin).decode('utf-8')
# Replace ID_PROVIDER so we have same nice IDs every time
sy.ID_PROVIDER = [ int(1e10) + i for i in reversed(range(100)) ]
sy.create_sandbox(globals(), download_data=False)
hook.local_worker.is_client_worker = False
hook.local_worker.framework = None
me = hook.local_worker
first = generateThreeWayProtocol(me)
second = generateTwoWayProtocol(me)
data = { "three-way": first, "two-way": second }
print(data)
print("\n----------\n\nSeed file created successfully!")
with open('./seed/data.json', 'w', encoding='utf-8') as f:
json.dump(data, f)