Pythonでの実行エラー

[tmsk]

初めて投稿させていただきます。
C言語ではないのですが、もしわかる方がいらっしゃれば教えていただきたく思います。

Pythonで入力されたデータセットを0と1に分類し、その入力されたデータセットと共に結果を出力するというプログラムを作成中です。

しかしターミナルで実行したところ以下のようなエラーが出てしまいます。どのように解決すれば良いのか分からない状態です。

また、入力されたデータの数値を一緒にcsvファイルに書き込むプログラムの書き方がわかりません。
もしご助言をいただけるなら幸いです。よろしくお願いいたします。

該当するプログラムのコード（classification.py）

コード:

import time
import torch
import pandas as pd
import numpy as np
import csv

# データサイズ(I行J列)
DATA_SIZE_I = 11
DATA_SIZE_J = 11

# ファイルパス
TRAIN_CSV_PATH = r"C:\Users\Tokyo\.vscode\su\train_data_noise.csv"
VALIDATION_CSV_PATH = r"C:\Users\Tokyo\.vscode\su\validation_data_noise.csv"
TEST_CSV_PATH = r"C:\Users\Tokyo\.vscode\su\test_data_noise.csv"
OUTPUT_CSV_PATH = 'result_noise.csv'
LOSS_CSV_PATH = '.result_losses.csv'

# ハイパーパラメータ
EPOCHS = 10
BATCH = 100
lr = 1.0e-4
NODE_NUM = DATA_SIZE_I*DATA_SIZE_J


class BGS(torch.utils.data.Dataset):

    def __init__(self, csv_path):
        df = pd.read_csv(csv_path, header=0, dtype='float')
        data = df.iloc[:, 1:].values
        data = data_transform(data)
        self.data = torch.tensor(data, dtype=torch.float32)
        labels = df.iloc[:, 0].values
        self.labels = torch.tensor(labels, dtype=torch.long)

    def __len__(self):
        return len(self.data)

    def __getitem__(self, index):
        return self.data[index], self.labels[index]


class Net(torch.nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.fc1 = torch.nn.Linear(DATA_SIZE_I*DATA_SIZE_J, NODE_NUM*5)
        self.fc2 = torch.nn.Linear(NODE_NUM*5, NODE_NUM)
        self.fc3 = torch.nn.Linear(NODE_NUM, 2)

    def forward(self, x):
        x = self.fc1(x)
        x = torch.nn.functional.relu(x)
        x = self.fc2(x)
        x = torch.nn.functional.relu(x)
        x = self.fc3(x)
        x = torch.nn.functional.softmax(x, dim=1)
        return x


def data_transform(x):
    return (x - np.mean(x, axis=1, keepdims=True)) / np.std(x, axis=1, ddof=1, keepdims=True)


# 学習用関数
def train(train_loader, model, optimizer, criterion, device, num_epochs, epoch, losses):
    model.train()  # モデルを学習モードに変更
    train_loss = 0
    train_acc = 0
    i = 0
    for i, (data, target) in enumerate(train_loader):
        data, target = data.to(device), target.to(device)
        optimizer.zero_grad()  # 勾配を初期化
        output = model(data)  # 順伝播の計算
        loss = criterion(output, target)  # 誤差を計算
        train_loss += float(loss.item())
        preds = torch.max(output, 1)[1]
        train_acc += torch.sum(preds == target).item() / len(target)
        loss.backward()  # 誤差を逆伝播させる
        optimizer.step()  # 重みを更新する
    losses.append(train_loss)
    print('#train_epoch: [{}/{}], Loss: {}, acc: {}'.format(epoch + 1, num_epochs, train_loss / (i+1), train_acc / (i+1)), end=' ')


def validation(validation_loader, model, criterion, device, num_epochs, epoch, losses):
    model.eval()  # モデルを推論モードに変更
    val_loss = 0
    val_acc = 0
    i = 0
    with torch.no_grad():
        for i, (data, target) in enumerate(validation_loader):
            data, target = data.to(device), target.to(device)
            output = model(data)
            loss = criterion(output, target)  # 誤差を計算
            val_loss += float(loss.item())
            preds = torch.max(output, 1)[1]
            val_acc += torch.sum(preds == target).item() / len(target)
        losses.append(val_loss)
        print('#val_epoch: [{}/{}], Loss: {}, acc: {}'.format(epoch + 1, num_epochs, val_loss / (i+1), val_acc / (i+1)))


def test(test_loader, model, device):
    model.eval()  # モデルを推論モードに変更
    with torch.no_grad():
        for i, (data, target) in enumerate(test_loader):
            data, target = data.to(device), target.to(device)
            output = model(data)
            preds = torch.max(output, 1)[1]
            accuracy = torch.sum(preds == target).item() / len(target)
            target, preds = target.cpu().numpy(), preds.cpu().numpy()
            target, preds = pd.DataFrame(target), pd.DataFrame(preds)
            df = pd.concat([target, preds], axis=1)
            with open(OUTPUT_CSV_PATH, 'w', newline='') as file:
                writer = csv.writer(file)
                writer.writerow(['accuracy', '{}'.format(accuracy)])
                writer.writerow(['labels', 'predict'])
            df.to_csv(OUTPUT_CSV_PATH, mode='a', header=False, index=False)


def main():
    # 1. GPUの設定（PyTorchでは明示的に指定する必要がある）
    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
    print(device)

    # 3. データセットを取得
    train_data = BGS(TRAIN_CSV_PATH)
    validation_data = BGS(VALIDATION_CSV_PATH)
    test_data = BGS(TEST_CSV_PATH)

    # 4. DataLoaderの作成
    train_loader = torch.utils.data.DataLoader(train_data, batch_size=BATCH, shuffle=True, num_workers=2)
    validation_loader = torch.utils.data.DataLoader(validation_data, batch_size=BATCH, shuffle=False)
    test_loader = torch.utils.data.DataLoader(test_data, batch_size=test_data.__len__(), shuffle=False)

    # 5. モデル作成
    torch.manual_seed(0)
    model = Net().to(device)

    # 6. 損失関数を定義
    criterion = torch.nn.CrossEntropyLoss()

    # 7. 最適化手法を定義
    optimizer = torch.optim.Adam(model.parameters(), lr=lr)

    # 8. 出力用ファイル作成
    train_losses = []
    # train_accuracy = []
    validation_losses = []
    # validation_accuracy = []

    # 9. 学習（エポック終了時点ごとにテスト用データで評価）
    start = time.time()
    for epoch in range(EPOCHS):
        train(train_loader, model, optimizer, criterion, device, EPOCHS, epoch, train_losses)
        validation(validation_loader, model, criterion, device, EPOCHS, epoch, validation_losses)
    test(test_loader, model, device)
    process_time = time.time() - start
    print("process_time:{0}".format(process_time) + "[sec]")
    # train_losses, validation_losses = pd.DataFrame(train_losses), pd.DataFrame(validation_losses)
    # df_losses = pd.concat([train_losses/1000, validation_losses/100], axis=1)
    # df_losses.to_csv(LOSS_CSV_PATH, mode='a', header=False, index=False)


if __name__ == '__main__':
    main()

表示されたエラー

コード:

PS C:\Users\Tokyo\PycharmProjects\SuNN> python classification.py
cuda
Traceback (most recent call last):
  File "pandas\_libs\parsers.pyx", line 1050, in pandas._libs.parsers.TextReader._convert_tokens
TypeError: Cannot cast array data from dtype('O') to dtype('float64') according to the rule 'safe'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "classification.py", line 163, in <module>
    main()
  File "classification.py", line 124, in main
    train_data = BGS(TRAIN_CSV_PATH)
  File "classification.py", line 28, in __init__
    df = pd.read_csv(csv_path, header=0, dtype='float')
  File "C:\Users\Tokyo\anaconda3\envs\pytorch19\lib\site-packages\pandas\io\parsers.py", line 610, in read_csv
    return _read(filepath_or_buffer, kwds)
  File "C:\Users\Tokyo\anaconda3\envs\pytorch19\lib\site-packages\pandas\io\parsers.py", line 468, in _read
    return parser.read(nrows)
  File "C:\Users\Tokyo\anaconda3\envs\pytorch19\lib\site-packages\pandas\io\parsers.py", line 1057, in read
    index, columns, col_dict = self._engine.read(nrows)
  File "C:\Users\Tokyo\anaconda3\envs\pytorch19\lib\site-packages\pandas\io\parsers.py", line 2061, in read
    data = self._reader.read(nrows)
  File "pandas\_libs\parsers.pyx", line 756, in pandas._libs.parsers.TextReader.read
  File "pandas\_libs\parsers.pyx", line 771, in pandas._libs.parsers.TextReader._read_low_memory
  File "pandas\_libs\parsers.pyx", line 850, in pandas._libs.parsers.TextReader._read_rows
  File "pandas\_libs\parsers.pyx", line 982, in pandas._libs.parsers.TextReader._convert_column_data
  File "pandas\_libs\parsers.pyx", line 1056, in pandas._libs.parsers.TextReader._convert_tokens
ValueError: could not convert string to float: ' 0.5185022.336320e-001'

[あたっしゅ]

東上☆海美☆「

viewtopic.php?f=3&t=21318
Pythonでの実行エラー - ミクプラ(ja)

と、二重書き込みみみ。

」