Skip to main content

Module validation

Module validation 

Source
Expand description

Input validation utilities for Python bindings

Functionsยง

validate_betas
Validate beta parameters for Adam-like optimizers
validate_broadcast_shapes
Validate that dimensions are compatible for broadcasting
validate_conv_params
Validate convolution parameters
validate_dimension
Validate that a dimension index is valid for a tensor
validate_dropout_probability
Validate dropout probability is in valid range [0, 1]
validate_epsilon
Validate that epsilon is positive
validate_finite
Validate that a value is finite (not NaN or infinity)
validate_index
Validate that an index is within bounds for a given dimension
validate_kernel_size
Validate kernel size is positive
validate_learning_rate
Validate that a learning rate is positive
validate_momentum
Validate that momentum is in valid range [0, 1]
validate_num_features
Validate that number of features matches expected value
validate_parameters_not_empty
Validate that parameters list is not empty
validate_pooling_output_size
Validate pooling output size calculation
validate_range
Validate that a range is valid (start < end)
validate_shape
Validate that a shape is valid (all dimensions > 0)
validate_stride
Validate stride is positive
validate_tensor_min_ndim
Validate that input tensor has at least minimum number of dimensions
validate_tensor_ndim
Validate that input tensor has expected number of dimensions
validate_tensor_shapes_match
Validate that tensor dimensions match for operations
validate_weight_decay
Validate that weight decay is non-negative