How severe does this issue affect your experience of using Ray?
- High: It blocks me to complete my task.
My code:
import math
import ray
import time
import pandas as pd
from sklearn import metrics
from sklearn.neighbors import KNeighborsClassifier
from sklearn.model_selection import train_test_split
from sklearn.feature_selection import SelectKBest, chi2
@ray.remote(num_cpus=1, max_restarts = 5)
class Actor(object):
def __init__(self):
filename = "Dataset - HIKARI-2021.csv"
df = pd.read_csv(filename, delimiter=',')
df = df.drop(columns=["uid", "originh", "responh"])
headers = list(df.columns)
X = df[headers[:-2]]
Y = df["traffic_category"]
X = X[:len(X)//8]
Y = Y[:len(Y)//8]
self.X_train, self.X_test, self.Y_train, self.Y_test = train_test_split(X, Y, test_size=0.2, random_state=1)
def compute(self,start, end):
print(str(start) + " - " + str(end))
aux = start
tmp = []
while aux <= end and aux <= 10:
fs = SelectKBest(score_func=chi2, k=aux)
fs.fit(self.X_train, self.Y_train)
X_train_fs = fs.transform(self.X_train)
X_test_fs = fs.transform(self.X_test)
model = KNeighborsClassifier(n_jobs=1)
model.fit(X_train_fs, self.Y_train)
Y_predict = model.predict(X_test_fs)
accuracy = metrics.accuracy_score(self.Y_test, Y_predict)
tmp.append([aux, accuracy])
aux += 1
"""filename = "results.txt" #str(start) + "-" + str(end) + ".txt"
f = open(filename, "a")
for item in tmp:
f.write("%s\n" % item)
f.close()
#self.result.append(tmp)
#self.result += 1
#return self.result"""
return 1
cpus = 4
actors = [Actor.remote() for _ in range(cpus)]
batch = math.ceil(10 / cpus)
res = ray.get([a.compute.remote(idx*batch+1,idx*batch+batch) for idx, a in enumerate(actors)])
print(res)
I am using Ray 1.12.1