),
# "static-second-layer-varying-sparsity": dict(
# model="DSCNN",
# network="gsc_sparse_dscnn",
# prune_methods=["none", "static"],
# sparsity=tune.grid_search([0.98, 0.99, 0.999]),
# ),
}
exp_configs = (
[(name, new_experiment(base_exp_config, c)) for name, c in experiments.items()]
if experiments
else [(experiment_name, base_exp_config)]
)
# Download dataset.
download_dataset(base_exp_config)
# Register serializers.
ray.init()
for t in [
torch.FloatTensor,
torch.DoubleTensor,
torch.HalfTensor,
torch.ByteTensor,
torch.CharTensor,
torch.ShortTensor,
torch.IntTensor,
torch.LongTensor,
torch.Tensor,
]:
ray.register_custom_serializer(t, serializer=serializer, deserializer=deserializer)
boost_strength_factor=0.7,
test_noise=True,
noise_level=0.1,
kwinners=tune.grid_search([True, False]), # moved to a parameter
)
tune_config = dict(
name="SET_DSNN_BoostingEval",
num_samples=1,
local_dir=os.path.expanduser("~/nta/results"),
config=exp_config,
checkpoint_freq=0,
checkpoint_at_end=False,
stop={"training_iteration": 300},
resources_per_trial={
"cpu": 1,
"gpu": 1
},
loggers=DEFAULT_LOGGERS,
verbose=1,
)
# override when running local for test
if not torch.cuda.is_available():
exp_config["device"] = "cpu"
tune_config["resources_per_trial"] = {"cpu": 1}
download_dataset(exp_config)
ray.init()
tune.run(Trainable, **tune_config)
