Lexicographic Sorting

Mirek keeps several files of integers and had to sort the integers in each file in ascending order. He found a command line tool whose name was easy to guess and ran it. The tool treated every integer as a string and sorted the numbers lexicographically. Even so, the files came out in ascending numerical order, and that surprised him.

Say that the lexicographic and numerical orders of a set of integers agree when sorting the decimal representations of its elements as strings gives the same listing as sorting the elements by value. Equivalently, every two elements $x < y$ of the set satisfy that the string of $x$ comes before the string of $y$ lexicographically.

Strings are compared one character at a time, and when one string is a prefix of the other, the shorter one comes first. So 1 comes before 10, while 98 comes after 100.

Given the range $[A, B]$, count the subsets of $\{A, A+1, \ldots, B\}$ whose lexicographic and numerical orders agree. The empty set and every subset with a single element satisfy the condition.

The first and only line contains two integers $A$ and $B$ ($1 \le A \le B \le 10^{18}$, $B - A \le 10^5$).

Print one integer $M$ on a single line, the number of subsets of $\{A, A+1, \ldots, B\}$ that satisfy the condition. The count can be very large, so print it modulo $10^9 + 7$.

For $A = 98$ and $B = 101$ the subsets that satisfy the condition are $\emptyset$, $\{98\}$, $\{99\}$, $\{100\}$, $\{101\}$, $\{98, 99\}$ and $\{100, 101\}$.