Hallo All
Have following situation:
- Creating of Tensor with values from 1 to 9 in a 5-digit number range
a = torch.arange(1, 10)
comb = torch.combinations(a, r=5)
a = comb.size()
a = comb.size()
data = torch.zeros([a[0], a[1]])
data = comb
it can be also other arange values
Example:
data([[1, 2, 3, 4, 5],
[1, 2, 3, 4, 6],
[1, 2, 3, 4, 7],
[1, 2, 3, 4, 8],
.............
[5, 6, 7, 8, 9])
- the Task: Grouping the Columns by first value and first- digit number,so that we have, create a dictionary
Example:
Number 1 at column 0 can be found at index 0 to 69 in the data[idx]
Nr_1_0: = > index: 0 - 69
Number 2 at column 0 can be found at index 70 to 104
Nr_2_0: => index: 70 - 104
Number 2 at column 1 can be found at index 0 to 34
Nr_2_1: => index 0 - 34
Number 3 at column 0 can be found at index 119 to 105
Nr_3_0: => index 105 - 119
Number 3 at column 1 can be found at index 35 to 54
Nr_3_1: => index 35 - 54
Number 3 at column 2 can be found at index 0 to 14
Nr_3_2: => index 0 - 14
and so one, until number 9 and column 4
more Number
So the Question is : how we can do that without any loops (optimization) ?
Thank you in Advance