burn-onnx 0.21.0-pre.3

Library for importing ONNX models into the Burn framework
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

use super::prelude::*;
use onnx_ir::ir::ArgType;
use quote::ToTokens as _;

impl NodeCodegen for onnx_ir::node::eye_like::EyeLikeNode {
    fn inputs(&self) -> &[Argument] {
        &self.inputs
    }

    fn outputs(&self) -> &[Argument] {
        &self.outputs
    }

    fn forward(&self, scope: &mut ScopeAtPosition<'_>) -> TokenStream {
        let input = scope.arg(self.inputs.first().unwrap());
        let output = arg_to_ident(self.outputs.first().unwrap());
        let k_offset = self.config.k.to_token_stream();

        // Convert mask to appropriate type based on output tensor kind
        let output_ty = &self.outputs.first().unwrap().ty;
        let conversion = match output_ty {
            ArgType::Tensor(t) => {
                let dtype_tokens = t.dtype.to_tokens();
                match &t.dtype {
                    dtype if dtype.is_int() || dtype.is_uint() => {
                        quote! { .int().cast(#dtype_tokens) }
                    }
                    dtype if dtype.is_float() => quote! { .float().cast(#dtype_tokens) },
                    dtype if dtype.is_bool() => quote! {},
                    _ => panic!("Unsupported EyeLike output dtype: {:?}", t.dtype),
                }
            }
            _ => panic!("EyeLike output must be a tensor"),
        };

        // Use diag_mask to create the diagonal matrix, then invert it
        // diag_mask returns false on diagonal, true off-diagonal
        // EyeLike needs true on diagonal, false off-diagonal
        quote! {
            let #output = Tensor::diag_mask(#input.shape(), #k_offset, &self.device).bool_not()#conversion;
        }
    }
}

#[cfg(test)]
mod tests {
    use super::super::test_helpers::*;
    use burn::tensor::{BoolStore, DType};
    use insta::assert_snapshot;
    use onnx_ir::node::eye_like::{EyeLikeConfig, EyeLikeNodeBuilder};

    #[test]
    fn test_eye_like_float() {
        let config = EyeLikeConfig::new(None, 0);
        let node = EyeLikeNodeBuilder::new("eye1")
            .input_tensor("input", 2, DType::F32)
            .output_tensor("output", 2, DType::F32)
            .config(config)
            .build();
        let code = codegen_forward_default(&node);
        assert_snapshot!(code, @r"
        pub fn forward(&self, input: Tensor<B, 2>) -> Tensor<B, 2> {
            let output = Tensor::diag_mask(input.shape(), 0i64, &self.device)
                .bool_not()
                .float()
                .cast(burn::tensor::DType::F32);
            output
        }
        ");
    }

    #[test]
    fn test_eye_like_int() {
        let config = EyeLikeConfig::new(None, 1);
        let node = EyeLikeNodeBuilder::new("eye2")
            .input_tensor("input", 2, DType::I32)
            .output_tensor("output", 2, DType::I32)
            .config(config)
            .build();
        let code = codegen_forward_default(&node);
        assert_snapshot!(code, @r"
        pub fn forward(&self, input: Tensor<B, 2, Int>) -> Tensor<B, 2, Int> {
            let output = Tensor::diag_mask(input.shape(), 1i64, &self.device)
                .bool_not()
                .int()
                .cast(burn::tensor::DType::I32);
            output
        }
        ");
    }

    #[test]
    fn test_eye_like_bool() {
        let config = EyeLikeConfig::new(None, 0);
        let node = EyeLikeNodeBuilder::new("eye3")
            .input_tensor("input", 2, DType::Bool(BoolStore::Native))
            .output_tensor("output", 2, DType::Bool(BoolStore::Native))
            .config(config)
            .build();
        let code = codegen_forward_default(&node);
        assert_snapshot!(code, @r"
        pub fn forward(&self, input: Tensor<B, 2, Bool>) -> Tensor<B, 2, Bool> {
            let output = Tensor::diag_mask(input.shape(), 0i64, &self.device).bool_not();
            output
        }
        ");
    }
}