numpy.linalg.linalg.LinAlgError: Singular matrix

[nabidam]

hello, I'm using NMF_ANLS_BLOCKPIVOT algorithm on a 200x990 matrix with this code
W1, H1, info1 = NMF_ANLS_BLOCKPIVOT().run(X1, 56, max_iter=500, verbose=1)
but after running for about 15 minutes, it shows an error for singular matrix:

File "nnmf.py", line 120, in
W1, H1, info1 = nmf().run(X1, rank, max_iter=500, verbose=1)
File "/home/navid/anaconda3/envs/fr/lib/python3.6/site-packages/nonnegfac/nmf.py", line 97, in run
(W, H) = self.iter_solver(A, W, H, k, i)
File "/home/navid/anaconda3/envs/fr/lib/python3.6/site-packages/nonnegfac/nmf.py", line 195, in iter_solver
Sol, info = nnlsm_blockpivot(W, A, init=H.T)
File "/home/navid/anaconda3/envs/fr/lib/python3.6/site-packages/nonnegfac/nnls.py", line 120, in nnlsm_blockpivot
AtA, AtB[:, not_opt_cols], PassSet[:, not_opt_cols])
File "/home/navid/anaconda3/envs/fr/lib/python3.6/site-packages/nonnegfac/nnls.py", line 322, in normal_eq_comb
Z[ix1] = nla.solve(AtA[ix2], AtB[ix1])
File "/home/navid/anaconda3/envs/fr/lib/python3.6/site-packages/numpy/linalg/linalg.py", line 394, in solve
r = gufunc(a, b, signature=signature, extobj=extobj)
File "/home/navid/anaconda3/envs/fr/lib/python3.6/site-packages/numpy/linalg/linalg.py", line 89, in _raise_linalgerror_singular
raise LinAlgError("Singular matrix")
numpy.linalg.linalg.LinAlgError: Singular matrix

I searched the singular matrix error and I got this error happens if a numpy matrix isn't invertible.
now what should I do?
sorry for my bad english and I thank you.