Multiple driver error

module multiplier(A ,B , P);

parameter BIT_WIDTH = 4;

input [BIT_WIDTH-1 : 0] A;
input [BIT_WIDTH-1 : 0] B;
output  [2*BIT_WIDTH-1 : 0] P;

wire [BIT_WIDTH-1 : 0] int_sum  [BIT_WIDTH-1 : 0];
wire [BIT_WIDTH-1 : 0] int_cout [BIT_WIDTH-1 : 0];

genvar i,j;
generate
   for(i=0;i<=BIT_WIDTH-1;i=i+1)				// B corresponds to i
		for(j=0;j<=BIT_WIDTH-1;j=j+1)			// A corespond to j
			if(i==0)
				if(j <= BIT_WIDTH-1 )
					begin:FIRST_ROW
							mul_unit(
								.in_x(A[j]), 
								.in_y(B[0]), 
								.cin(), 
								.sin(), 
								.cout(int_cout[i][0]), 
								.sout(int_sum[i][0])
										);
					end
			
			else
					if(i == BIT_WIDTH-1)
						begin:LAST_ROW
								mul_unit(
									.in_x(), 
									.in_y(B[i]), 
									.cin(int_cout[i-1][j]), 
									.sin(int_sum[i-1][j+1]), 
									.cout(int_cout[i][j]), 
									.sout(int_sum[i][j])
											);
						end
					else
							begin:OTHER_ROW
							mul_unit(
								.in_x(), 
								.in_y(B[i]), 
								.cin(int_cout[i-1][j]), 
								.sin(int_sum[i-1][j+1]), 
								.cout(int_cout[i][j]), 
								.sout(int_sum[i][j])
								);
							end
endgenerate;

	for(i=0;i<=BIT_WIDTH-1;i=i+1)			// B corresponds to i
		for(j=0;j<=BIT_WIDTH-1;j=j+1)		// A corespond to j
			if(i< BIT_WIDTH -1 )
				assign P[i] = int_sum[i][0]	;
			else
				if(j< BIT_WIDTH-1)
					assign P[BIT_WIDTH+j] = int_sum[i][j];
				else
					assign P[BIT_WIDTH+j] = int_cout[i][j]; 
		

endmodule

generate
   for(i=0;i<=BIT_WIDTH-1;i=i+1)				// B corresponds to i
		for(j=0;j<=BIT_WIDTH-1;j=j+1)			// A corespond to j
			if(i==0)
				if(j <= BIT_WIDTH-1 ) // <== are you sure about this? (*)
					begin:FIRST_ROW
							mul_unit(
								.in_x(A[j]), 
								.in_y(B[0]), 
								.cin(), 
								.sin(), 
								.cout(int_cout[i][0]), 
								.sout(int_sum[i][0]) // <== That one!
										);