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
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161

use crate::internal::*;
use crate::ops::matmul::*;

/// The binary op. It will declutter to MatMulUnary if either A or B is constant.
///
/// TODO: implemnent TypedOp fully to play nice with optimizer.
/// TODO: codegen fails if A and B are variable inputs.
#[derive(Debug, Clone, Default, Hash)]
pub struct MatMul {
    pub axes: MatMulAxes,
}

impl_dyn_hash!(MatMul);

impl Op for MatMul {
    fn name(&self) -> Cow<str> {
        "MatMul".into()
    }

    op_as_typed_op!();
}

impl EvalOp for MatMul {
    fn is_stateless(&self) -> bool {
        true
    }

    fn eval(&self, inputs: TVec<TValue>) -> TractResult<TVec<TValue>> {
        if inputs[0].rank() != inputs[1].rank() {
            bail!("Rank mismatch {:?} vs {:?}", inputs[0], inputs[1]);
        }
        Ok(tvec!(eval(&inputs[0], &inputs[1], self.axes)?.into()))
    }
}

impl TypedOp for MatMul {
    fn output_facts(&self, inputs: &[&TypedFact]) -> TractResult<TVec<TypedFact>> {
        if inputs[0].rank() != inputs[1].rank() {
            bail!(
                "Inconsistent matmul between {:?} and {:?} (rank mismatch)",
                inputs[0],
                inputs[1]
            );
        }
        let (_m, _k, _n, c_shape) = compute_shape(&inputs[0].shape, &inputs[1].shape, self.axes)?;
        Ok(tvec!(output_type(inputs[0].datum_type).fact(c_shape)))
    }

    fn declutter(
        &self,
        model: &TypedModel,
        node: &TypedNode,
    ) -> TractResult<Option<TypedModelPatch>> {
        let a_fact = model.outlet_fact(node.inputs[0])?;
        let b_fact = model.outlet_fact(node.inputs[1])?;
        let konst_ix = if a_fact.konst.is_some() {
            0
        } else if b_fact.konst.is_some() {
            1
        } else {
            return Ok(None);
        };

        let var_ix = 1 - konst_ix;
        let flip = konst_ix == 1;

        let axes = if flip {
            MatMulAxes {
                a_m: self.axes.b_n,
                a_k: self.axes.b_k,
                b_n: self.axes.a_m,
                b_k: self.axes.a_k,
                c_m: self.axes.c_n,
                c_n: self.axes.c_m,
            }
        } else {
            self.axes
        };

        let konst = model.outlet_fact(node.inputs[konst_ix])?.konst.clone().unwrap();
        TypedModelPatch::replace_single_op(
            model,
            node,
            &node.inputs[var_ix..][..1],
            MatMulUnary::new(konst, axes),
        )
        .map(Some)
    }

    fn cost(&self, inputs: &[&TypedFact]) -> TractResult<TVec<(Cost, TDim)>> {
        super::cost(
            &inputs[0].shape.to_tvec(),
            &inputs[1].shape.to_tvec(),
            inputs[0].datum_type,
            self.axes,
        )
    }

    as_op!();
}

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

    #[test]
    fn bin() {
        //           0
        //           1
        //           2
        //
        // 0 1 2     5
        // 3 4 5    14
        let a = tensor2(&[[0f32, 1.0, 2.0], [3.0, 4.0, 5.0]]);
        let b = tensor2(&[[0f32], [1.0], [2.0]]);
        let c = tensor2(&[[5f32], [14.0]]);
        let op = MatMul::default();
        let c_found = op.eval(tvec!(a.into(), b.into())).unwrap().pop().unwrap();
        c.close_enough(&c_found, true).unwrap();
    }

    #[test]
    fn bin_transpose() {
        let a = tensor2(&[[0f32, 1.0, 2.0], [3.0, 4.0, 5.0]]);
        let b = tensor2(&[[0f32], [1.0], [2.0]]);
        let c = tensor2(&[[5f32], [14.0]]);
        let op = MatMul { axes: MatMulAxes::default().transposing(true, true, true) };
        let c_found = op.eval(tvec!(b.into(), a.into())).unwrap().pop().unwrap();

        c.close_enough(&c_found, true).unwrap();
    }

    #[test]
    fn batch_input() -> TractResult<()> {
        crate::setup_test_logger();
        let (batch, len, ci, co) = (2, 3, 4, 5);
        let mut model = TypedModel::default();
        let input_shape = tvec!(batch, len, ci);
        let mut wire = tvec!(model.add_source("s", f32::fact(&*input_shape))?);
        let mut a = Tensor::zero::<f32>(&[1, ci, co])?;
        a.as_slice_mut::<f32>().unwrap()[0] = 1.0;
        let a = a.into_arc_tensor();
        wire = model.wire_node(
            "m",
            MatMulUnary { a, axes: MatMulAxes::default_for_rank(3).transposing(true, true, true) },
            &wire,
        )?;
        let mut b = Tensor::zero::<f32>(&[1, 1, co])?;
        b.as_slice_mut::<f32>().unwrap()[0] = 1.0;
        let b = b.into_arc_tensor();
        let b = model.add_const("b", b)?;
        wire = model.wire_node("a", crate::ops::math::add(), &[wire[0], b])?;
        model.set_output_outlets(&wire)?;
        let input = Tensor::zero::<f32>(&input_shape)?.into_tvalue();
        trace!("running mir");
        model.clone().into_runnable()?.run(tvec!(input.clone()))?;
        trace!("running optimized");
        model.into_decluttered()?.into_optimized()?.into_runnable()?.run(tvec!(input))?;
        Ok(())
    }
}