Fix CSV loading and train_test_split label variable

[harshitrajput2802]

Summary

This pull request fixes two issues that prevented the project from running correctly.

Changes

• Removed index_col=0 when reading mnist_mapped.csv in NEURAL2-TORCH.py.

  • This prevents the first pixel column from being interpreted as the DataFrame index, ensuring all 784 input features are loaded.

• Fixed the variable name in mnist_test.ipynb.

  • Changed label to labels in the train_test_split() call to match the variable defined earlier in the notebook.

Result

These changes resolve runtime errors encountered during dataset loading and preprocessing, allowing the project to run successfully with the generated mnist_mapped.csv.

Testing

• Generated mnist_mapped.csv using mnist_test.ipynb.
• Successfully ran NEURAL2-TORCH.py.
• Verified that the visualization window opens and training starts correctly.