tract-tensorflow 0.22.1

Tiny, no-nonsense, self contained, TensorFlow and ONNX inference
Documentation 
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

use tract_hir::internal::*;
use tract_ndarray::prelude::*;

use tract_hir::tract_core::ops::cnn::{Conv, PoolSpec};

#[derive(Debug, Copy, Clone, Hash)]
pub enum PaddingStrat {
    FlexFixed(usize),
    FixedFlex(usize),
    FixedFixed(usize, usize),
}

#[derive(Debug, Clone, new, Hash)]
pub struct SpaceToBatchUnary {
    pub datum_type: DatumType,
    pub space_shape: TVec<TDim>,
    pub batch_shape: TVec<TDim>,
    pub block_shape: Array1<i32>,
    pub pad: TVec<PaddingStrat>,
}

impl Op for SpaceToBatchUnary {
    fn name(&self) -> StaticName {
        "SpaceToBatchUnary".into()
    }

    op_as_typed_op!();
}

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

    fn eval(&self, inputs: TVec<TValue>) -> TractResult<TVec<TValue>> {
        let input = args_1!(inputs);
        let mut paddings = Array2::zeros((self.block_shape.len(), 2));
        for (ax, &strat) in self.pad.iter().enumerate() {
            let spread = (self.batch_shape[2 + ax].clone() * self.block_shape[ax]
                - &self.space_shape[2 + ax])
                .to_usize()?;
            let (bef, aft) = match strat {
                PaddingStrat::FlexFixed(f) => (spread - f, f),
                PaddingStrat::FixedFlex(f) => (f, spread - f),
                PaddingStrat::FixedFixed(a, b) => (a, b),
            };
            paddings[(ax, 0)] = bef as i32;
            paddings[(ax, 1)] = aft as i32;
        }
        let r = dispatch_numbers!(super::space_to_batch(input.datum_type())(
            input,
            &self.block_shape.view(),
            &paddings.view()
        ))?;
        Ok(tvec!(r))
    }
}

impl TypedOp for SpaceToBatchUnary {
    fn output_facts(&self, inputs: &[&TypedFact]) -> TractResult<TVec<TypedFact>> {
        Ok(tvec!(inputs[0].datum_type.fact(&self.batch_shape)))
    }

    fn declutter(
        &self,
        model: &TypedModel,
        node: &TypedNode,
    ) -> TractResult<Option<TypedModelPatch>> {
        let [succ] = &*model.node(node.id).outputs[0].successors else { return Ok(None) };
        let conv_node = model.node(succ.node);
        let Some(conv_op) = conv_node.op_as::<Conv>() else { return Ok(None) };
        let [succ] = &*conv_node.outputs[0].successors else { return Ok(None) };
        let b2s_node = model.node(succ.node);
        let Some(_bs2_op) = b2s_node.op_as::<BatchToSpaceUnary>() else { return Ok(None) };
        let op = Conv {
            pool_spec: PoolSpec {
                dilations: Some(self.block_shape.iter().map(|&i| i as usize).collect()),
                ..conv_op.pool_spec.clone()
            },
            ..conv_op.clone()
        };
        let mut patch = TypedModelPatch::default();
        let taps_s2b = patch.taps(model, &node.inputs)?;
        let mut taps_conv = patch.taps(model, &conv_node.inputs)?;
        taps_conv[0] = taps_s2b[0];
        let out = patch.model.wire_node(&*conv_node.name, op, &taps_conv)?[0];
        patch.shunt_outside(model, OutletId::new(b2s_node.id, 0), out)?;
        Ok(Some(patch))
    }

    as_op!();
}

#[derive(Debug, Clone, new, Hash)]
pub struct BatchToSpaceUnary {
    datum_type: DatumType,
    batch_shape: TVec<TDim>,
    space_shape: TVec<TDim>,
    block_shape: Array1<i32>,
    pad: Vec<PaddingStrat>,
}

impl Op for BatchToSpaceUnary {
    fn name(&self) -> StaticName {
        "BatchToSpaceUnary".into()
    }

    op_as_typed_op!();
}

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

    fn eval(&self, inputs: TVec<TValue>) -> TractResult<TVec<TValue>> {
        let input = args_1!(inputs);
        let mut paddings = Array2::zeros((self.block_shape.len(), 2));
        for (ax, &strat) in self.pad.iter().enumerate() {
            let spread = (self.batch_shape[2 + ax].clone() * self.block_shape[ax]
                - &self.space_shape[2 + ax])
                .to_usize()?;
            let (bef, aft) = match strat {
                PaddingStrat::FlexFixed(f) => (spread - f, f),
                PaddingStrat::FixedFlex(f) => (f, spread - f),
                PaddingStrat::FixedFixed(a, b) => (a, b),
            };
            paddings[(ax, 0)] = bef as i32;
            paddings[(ax, 1)] = aft as i32;
        }
        let r = dispatch_numbers!(super::batch_to_space(input.datum_type())(
            input,
            &self.block_shape.view(),
            &paddings.view()
        ))?;
        Ok(tvec!(r))
    }
}

impl TypedOp for BatchToSpaceUnary {
    as_op!();

    fn output_facts(&self, inputs: &[&TypedFact]) -> TractResult<TVec<TypedFact>> {
        Ok(tvec!(inputs[0].datum_type.fact(&self.space_shape)))
    }
}