fix: use numpy reshape instead of keras.ops.reshape in SequenceEstimator

[jrosenfeld13]

Fixes #5 - torch backend LSTMRegressor fit fails with CUDA tensors.

The issue was that ops.reshape() converts numpy arrays to backend tensors. With torch backend + CUDA GPU, this creates CUDA tensors that numpy.asarray() cannot convert back, causing the error: "can't convert cuda:0 device type tensor to numpy"

The fix uses numpy's native reshape() method instead, keeping data as numpy arrays until model.fit() where Keras handles the conversion internally. This matches the pattern used by other estimators (dense.py, autoencoder.py, tree.py) which don't use keras.ops for preprocessing.

---

Note

Replaces Keras ops.reshape with NumPy reshape in SequenceEstimator to keep data as NumPy arrays during preprocessing and avoid unintended backend tensor conversion (e.g., CUDA tensors with Torch backend).

• Update _reshape to call X.reshape(...) and similarly for validation data; return unchanged API
• Remove unused keras.ops import
• Keeps prediction path intact; only reshaping logic changed prior to model.fit()

^{Written by Cursor Bugbot for commit fa223a9. This will update automatically on new commits. Configure here.}