def benchmark(args, dataset_folder, dataset):
data = prepare_dataset(dataset_folder, dataset, args.nrows)
results = {}
# "all" runs all algorithms
if args.algorithm == "all":
args.algorithm = "xgb-gpu,xgb-cpu,lgbm-cpu,lgbm-gpu,cat-cpu,cat-gpu"
for alg in args.algorithm.split(","):
print("Running '%s' ..." % alg)
runner = algorithms.Algorithm.create(alg)
with runner:
train_time = runner.fit(data, args)
pred = runner.test(data)
results[alg] = {
"train_time": train_time,
"accuracy": get_metrics(data, pred),
}
return results
import cudf
import numpy as np
import pandas as pd
import pickle
from datasets import prepare_dataset
from sklearn.metrics import accuracy_score
from cuml.ensemble import RandomForestClassifier as GPURandomForestClassifier
import ray
from ray import tune
from ray.tune.utils import pin_in_object_store, get_pinned_object
data = prepare_dataset("/data", "airline", None)
X_train, X_test, y_train, y_test = data.X_train, data.X_test, data.y_train, data.y_test
y_train = y_train.astype(np.int32)
y_test = y_test.astype(np.int32)
QUARTER = len(X_train) // 3
X_train = X_train[QUARTER:]
y_train = y_train[QUARTER:]
# ray.init()
# data_id = pin_in_object_store([X_train, X_test, y_train, y_test])
import os
from filelock import FileLock
class CUMLTrainable(tune.Trainable):
def _setup(self, config):
except:
start_epoch, epoch_iter = 1, 0
print('Resuming from epoch %d at iteration %d' % (start_epoch, epoch_iter))
else:
start_epoch, epoch_iter = 1, 0
if opt.debug:
opt.display_freq = 1
opt.print_freq = 1
opt.niter = 1
opt.niter_decay = 0
opt.max_dataset_size = 10
dataurl = opt.dataurl
opt.dataroot = str(prepare_dataset(dataurl, opt.dataroot) / 'subject4' / 'train')
data_loader = CreateDataLoader(opt)
dataset = data_loader.load_data()
dataset_size = len(data_loader)
print('#training images = %d' % dataset_size)
""" new residual model """
model = create_model_fullts(opt)
visualizer = Visualizer(opt)
total_steps = (start_epoch-1) * dataset_size + epoch_iter
for epoch in range(start_epoch, opt.niter + opt.niter_decay + 1):
epoch_start_time = time.time()
if epoch != start_epoch:
from datasets import prepare_dataset
import os
os.makedirs(os.path.expanduser("~/data"), exist_ok=True)
prepare_dataset("~/data", "airline", None)