def train_stop(peer):
model_inference(peer, one_batch=True)
acceptance_rate = round(peer.params.n_accept / peer.params.exchanges * 100,
2)
peer.params.ar = acceptance_rate
log(
'info',
f"{peer} Acceptance rate for sigma=({peer.params.sigma}) DMedian( {np.median(peer.params.D)}): {acceptance_rate} %"
)
peer.stop()
return
def fit(self, inference=True):
# train the model
history = model_fit(self)
# set local model variable
self.local_model = self.model
# evaluate against a one batch or the whole inference dataset
# history = None
if inference:
model_inference(self, one_batch=False)
return history
def train_stop(peer: Node):
model_inference(peer, one_batch=True)
peer.stop()
def train_stop(peer: Node, args):
if peer.id == args.server_id:
model_inference(peer, one_batch=True)
peer.stop()
def train_stop(peer):
model_inference(peer, one_batch=True)
# acceptance_rate = peer.params.n_accept / peer.params.exchanges * 100
# log('info', f"{peer} Acceptance rate for alpha_max=({peer.params.alpha_max}): {acceptance_rate} %")
return