1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

//! N-way tensor broadcast
use crate::model::TVec;
use crate::dim::DimLike;

/// Computes a shape, if any, to which all shapes can be broadcasted.
pub fn multi_broadcast<D>(shapes: &[impl AsRef<[D]>]) -> Option<TVec<D>>
where
    D: DimLike
{
    let one = D::one();
    let len = shapes.iter().map(|shape| shape.as_ref().len()).max()?;
    let mut shape: TVec<D> = tvec!();
    for i in 0..len {
        let mut wanted_size = D::one();
        for shape in shapes {
            let len = shape.as_ref().len();
            let dim = if i < len { &shape.as_ref()[len - i - 1] } else { &one };
            if dim != &D::one() {
                if wanted_size != D::one() && dim != &wanted_size {
                    return None;
                }
                wanted_size = dim.clone();
            }
        }
        shape.push(wanted_size)
    }
    shape.reverse();
    Some(shape)
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn onnx_1() {
        assert_eq!(multi_broadcast(&tvec![tvec![2, 3, 4, 5], tvec![]]), Some(tvec![2, 3, 4, 5]))
    }

    #[test]
    fn onnx_2() {
        assert_eq!(multi_broadcast(&tvec![tvec![2, 3, 4, 5], tvec![5]]), Some(tvec![2, 3, 4, 5]))
    }

    #[test]
    fn onnx_3() {
        assert_eq!(multi_broadcast(&tvec![tvec![4, 5], tvec![2, 3, 4, 5]]), Some(tvec![2, 3, 4, 5]))
    }

    #[test]
    fn onnx_4() {
        assert_eq!(
            multi_broadcast(&tvec![tvec![1, 4, 5], tvec![2, 3, 4, 1]]),
            Some(tvec![2, 3, 4, 5])
        )
    }

    #[test]
    fn onnx_5() {
        assert_eq!(
            multi_broadcast(&tvec![tvec![3, 4, 5], tvec![2, 1, 1, 1]]),
            Some(tvec![2, 3, 4, 5])
        )
    }
}