Partitioned matrix multiplication

I have two lists with tensors. Tensors in first list have shapes (3, 200), (5, 200), (11, 200). Tensors in second list have shapes (2, 200), (4, 200), (7, 200), (13, 200), (14, 200). I want to get a matrix of shape (3, 5) - score of each tensor in first list with each tensor in second list. I tried to concat tensors in first list, tensors in second list and get tensors of shapes (19, 200) and (40, 200), then multiply them, but there is no way to get tensor of shape (3, 5) at the end. How can I do that without padding