Remove the max_nnz argument again, which cna just be set in the mat…

…ching pursuit method.

src/KSVD.jl

@@ -59,7 +59,6 @@ Dictionary vectors will be normalized such that ~all(norm.(eachcol(D), 2) .≈ 1
 # Arguments
 - `Y::AbstractMatrix{T}`: Input data matrix of size (num_features x num_samples)
 - `n_atoms::Int`: Number of atoms (columns) in the dictionary
-- `max_nnz=max(n_atoms÷10, 1)`: Maximum number of non-zero coefficients in each sparse representation
 
 # Keyword Arguments
 - `ksvd_update_method`: Method used for updating the dictionary (default: BatchedParallelKSVD)
@@ -84,7 +83,7 @@ A named tuple containing:
 # Notes
 To enable timing outputs, run `TimerOutputs.enable_debug_timings(KSVD)`.
 """
-function ksvd(Y::AbstractMatrix{T}, n_atoms::Int, max_nnz=n_atoms÷10;
+function ksvd(Y::AbstractMatrix{T}, n_atoms::Int;
               ksvd_update_method = BatchedParallelKSVD{false, T}(; shuffle_indices=true, batch_size_per_thread=1),
               sparse_coding_method = ParallelMatchingPursuit(; max_nnz, rtol=5e-2),
               minibatch_size=nothing,