Description:
There is a potential issue in CeLEry/CeLEry_package/CeLEry/fit_functions.py at line 109 where the code attempts to convert layer_weights to a NumPy array before creating a PyTorch tensor:
In the tutorial.ipynb, layer_weights is already a tensor type, which does not have a to_numpy() method, causing an AttributeError. This suggests that the handling of layer_weights should be more robust to accommodate different data types.
Potential Optimization:
# If there's no need for a new tensor copy, consider using:
# layer_weights = layer_weights
# This avoids unnecessary data conversion if the only goal is to ensure it's a tensor.
Proposed Solution:
To prevent errors and make the code more robust to different input types, I recommend adding a type check before processing layer_weights. Here is the suggested modification:
if isinstance(layer_weights, torch.Tensor):
# If layer_weights is already a Tensor, convert to NumPy array and back to Tensor (if necessary)
layer_weights = torch.tensor(layer_weights.numpy())
elif isinstance(layer_weights, pd.DataFrame):
# If layer_weights is a DataFrame, convert to NumPy array first
layer_weights = torch.tensor(layer_weights.to_numpy())
else:
# If the input type is neither Tensor nor DataFrame, raise an error
raise TypeError("layer_weights must be either a pandas DataFrame or a torch Tensor.")
Additional Context:
This change will ensure that layer_weights can be correctly processed whether it comes as a PyTorch tensor or a pandas DataFrame, thereby enhancing the robustness and reliability of the Fit_layer function.
Description: There is a potential issue in
CeLEry/CeLEry_package/CeLEry/fit_functions.pyat line 109 where the code attempts to convertlayer_weightsto a NumPy array before creating a PyTorch tensor:In the
tutorial.ipynb,layer_weightsis already a tensor type, which does not have ato_numpy()method, causing anAttributeError. This suggests that the handling oflayer_weightsshould be more robust to accommodate different data types.Potential Optimization:
Proposed Solution: To prevent errors and make the code more robust to different input types, I recommend adding a type check before processing
layer_weights. Here is the suggested modification:Additional Context: This change will ensure that
layer_weightscan be correctly processed whether it comes as a PyTorch tensor or a pandas DataFrame, thereby enhancing the robustness and reliability of theFit_layerfunction.