main #
vtl_vandermont
Builds a 5×5 Vandermonde matrix, then demonstrates tensor slicing and in-place assignment.
What it demonstrates
- Creating a 2-D tensor from a nested V array with
vtl.from_2d - Slicing a sub-tensor by row/column range with
.slice_hilo() - Inspecting tensor shape with
.shape - Assigning into a slice with
.assign()(modifies the underlying tensor)
How to run
v run main.vExpected output
[[ 1, 1, 1, 1, 1],
[ 2, 4, 8, 16, 32],
[ 3, 9, 27, 81, 243],
[ 4, 16, 64, 256, 1024],
[ 5, 25, 125, 625, 3125]]
[5, 5]
slice:
[[ 27, 81],
[ 64, 256]]
[2, 2]
span slice:
[[ 4, 16, 64, 256, 1024],
[ 5, 25, 125, 625, 3125]]
[2, 5]
slice until:
[[1, 1, 1, 1, 1],
[2, 4, 8, 16, 32],
[3, 9, 27, 81, 243]]
[3, 5]
assign:
[[ 1, 1, 1, 1, 1],
[ 2, 4, 8, 16, 32],
[ 3, 9, 999, 999, 243],
[ 4, 16, 999, 999, 1024],
[ 5, 25, 125, 625, 3125]]
Key API
| Function / method | Description |
|---|---|
vtl.from_2d(arr) |
Create a 2-D tensor from a nested V array |
t.slice_hilo(rows, cols) |
Slice by [start, end) row and column ranges |
t.shape |
[]int with the size of each dimension |
slice.assign(t) |
Write values from t into the slice's memory |
Related examples
vtl_basic_usage— 1-D tensors and basic operationsautograd_backprop— automatic differentiation