evaluate result not nan
All checks were successful
s444498-evaluation/pipeline/head This commit looks good
s444498-training/pipeline/head This commit looks good

This commit is contained in:
Wirusik 2022-05-08 23:38:16 +02:00
parent 8780f35c2f
commit 9b71dc20d6
2 changed files with 32 additions and 27 deletions

37
init.py
View File

@ -5,14 +5,17 @@ import numpy as np
from sklearn.model_selection import train_test_split
import matplotlib
from pathlib import Path
import math
# Inicjalizacja danych
file_exists = exists('./df_atp.csv')
file_exists = exists("./df_atp.csv")
if not file_exists:
subprocess.run(["kaggle", "datasets", "download", "-d", "hakeem/atp-and-wta-tennis-data"])
subprocess.run(
["kaggle", "datasets", "download", "-d", "hakeem/atp-and-wta-tennis-data"]
)
subprocess.run(["unzip", "-o", "atp-and-wta-tennis-data.zip"])
atp_data = pd.read_csv('df_atp.csv')
print(atp_data)
atp_data = pd.read_csv("df_atp.csv")
# Średnia ilość gemów w pierwszym secie zwycięzców meczu
print(atp_data[["Winner", "W1"]].mean())
@ -30,25 +33,25 @@ print(atp_data[["Winner", "W1"]].std())
print(atp_data[["Winner", "W1"]].median())
# Zmiana nazwy nienazwanej kolumny
atp_data.rename(columns={'Unnamed: 0':'ID'}, inplace=True)
atp_data.rename(columns={"Unnamed: 0": "ID"}, inplace=True)
# Jak często kto był zwycięzcą
print(atp_data.groupby("Winner")["ID"].nunique())
# Normalizacja rund -1: Finał, -2: Półfinał, -3: Ćwiartka, -4: Każdy z każdym
# 1: pierwsza runda, 2: druga runda, 3: trzecia runda, 4: czwarta runda
atp_data.loc[atp_data["Round"] == 'The Final', "Round"] = -1
atp_data.loc[atp_data["Round"] == 'Semifinals', "Round"] = -2
atp_data.loc[atp_data["Round"] == 'Quarterfinals', "Round"] = -3
atp_data.loc[atp_data["Round"] == 'Round Robin', "Round"] = -4
atp_data.loc[atp_data["Round"] == '1st Round', "Round"] = 1
atp_data.loc[atp_data["Round"] == '2nd Round', "Round"] = 2
atp_data.loc[atp_data["Round"] == '3rd Round', "Round"] = 3
atp_data.loc[atp_data["Round"] == '4th Round', "Round"] = 4
atp_data.loc[atp_data["Round"] == "The Final", "Round"] = -1
atp_data.loc[atp_data["Round"] == "Semifinals", "Round"] = -2
atp_data.loc[atp_data["Round"] == "Quarterfinals", "Round"] = -3
atp_data.loc[atp_data["Round"] == "Round Robin", "Round"] = -4
atp_data.loc[atp_data["Round"] == "1st Round", "Round"] = 1
atp_data.loc[atp_data["Round"] == "2nd Round", "Round"] = 2
atp_data.loc[atp_data["Round"] == "3rd Round", "Round"] = 3
atp_data.loc[atp_data["Round"] == "4th Round", "Round"] = 4
print(atp_data["Round"])
# Czyszczenie: W polu z datą zamienimy ######## na pustego stringa
atp_data.loc[atp_data["Date"] == '########', "Date"] = ''
atp_data.loc[atp_data["Date"] == "########", "Date"] = ""
print(atp_data["Date"])
# Podział na podzbiory: trenujący, testowy, walidujący w proporcjach 6:2:2
@ -62,6 +65,6 @@ print("\nElements of dev set: " + str(len(atp_dev)))
print("\nElements of train set: " + str(len(atp_train)))
# Stworzenie plików z danymi trenującymi i testowymi
atp_test.to_csv('atp_test.csv', encoding="utf-8", index=False)
atp_dev.to_csv('atp_dev.csv', encoding="utf-8", index=False)
atp_train.to_csv('atp_train.csv', encoding="utf-8", index=False)
atp_test.to_csv("atp_test.csv", encoding="utf-8", index=False)
atp_dev.to_csv("atp_dev.csv", encoding="utf-8", index=False)
atp_train.to_csv("atp_train.csv", encoding="utf-8", index=False)

View File

@ -11,13 +11,15 @@ default_epochs = 4
device = "cuda" if torch.cuda.is_available() else "cpu"
class AtpDataset(Dataset):
def __init__(self, file_name):
df = pd.read_csv(file_name)
df = pd.read_csv(file_name, usecols=["AvgL", "AvgW"])
df = df.dropna()
# Loser avg and Winner avg
x = df.iloc[:, 4].values
y = df.iloc[:, 3].values
x = df.iloc[:, 1].values
y = df.iloc[:, 0].values
self.x_train = torch.from_numpy(x)
self.y_train = torch.from_numpy(y)
@ -76,9 +78,9 @@ def test(dataloader, model, loss_fn):
def setup_args():
args_parser = argparse.ArgumentParser(prefix_chars='-')
args_parser.add_argument('-b', '--batchSize', type=int, default=default_batch_size)
args_parser.add_argument('-e', '--epochs', type=int, default=default_epochs)
args_parser = argparse.ArgumentParser(prefix_chars="-")
args_parser.add_argument("-b", "--batchSize", type=int, default=default_batch_size)
args_parser.add_argument("-e", "--epochs", type=int, default=default_epochs)
return args_parser.parse_args()
@ -87,8 +89,8 @@ print(f"Using {device} device")
args = setup_args()
batch_size = args.batchSize
plant_test = AtpDataset('atp_test.csv')
plant_train = AtpDataset('atp_train.csv')
plant_test = AtpDataset("atp_test.csv")
plant_train = AtpDataset("atp_train.csv")
train_dataloader = DataLoader(plant_train, batch_size=batch_size)
test_dataloader = DataLoader(plant_test, batch_size=batch_size)
@ -111,5 +113,5 @@ for t in range(epochs):
test(test_dataloader, model, loss_fn)
print("Finish!")
torch.save(model.state_dict(), './model.zip')
print("Model saved in ./model.zip file.")
torch.save(model.state_dict(), "./model.zip")
print("Model saved in ./model.zip file.")