nove_tensor 0.1.2

An easy-to-use, lightweight deep learning library wrapped around Candle Tensor.
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
147
148
149
150
151

use nove::tensor::{DType, Device, Tensor};
use proptest::prelude::*;

mod util;

fn test_add_two_tensors<S>((tensor1, data1): (Tensor, Vec<S>), (tensor2, data2): (Tensor, Vec<S>))
where
    S: candle_core::WithDType + std::fmt::Debug,
{
    let tensor3 = tensor1.add(&tensor2);
    let data3 = tensor3.unwrap().to_vec::<S>().unwrap();
    assert_eq!(
        data3,
        data1
            .iter()
            .zip(data2.iter())
            .map(|(a, b)| *a + *b)
            .collect::<Vec<S>>()
    );
}

proptest! {
    #[test]
    fn test_add_two_tensors_with_same_shape(
        shape_seed in prop::num::u64::ANY,
        dtype in prop::sample::select(vec![DType::U8, DType::U32, DType::I64, DType::F32,DType::F64]),
        grad_enabled in prop::bool::ANY,
        data_seed in prop::num::u64::ANY,
    ) {
        let shape = util::generate_random_shape_with_appropriate_size(shape_seed);
        let device = Device::cpu();
        let mut data_seed = data_seed;

        match dtype {
            DType::F32 => {
                let (tensor1, data1) = util::generate_random_tensor_and_corresponding_data::<f32>(&shape, &dtype, &device, grad_enabled, data_seed);
                data_seed += 1;
                let (tensor2, data2) = util::generate_random_tensor_and_corresponding_data::<f32>(&shape, &dtype, &device, grad_enabled, data_seed);
                test_add_two_tensors((tensor1, data1), (tensor2, data2));
            }
            DType::F64 => {
                let (tensor1, data1) = util::generate_random_tensor_and_corresponding_data::<f64>(&shape, &dtype, &device, grad_enabled, data_seed);
                data_seed += 1;
                let (tensor2, data2) = util::generate_random_tensor_and_corresponding_data::<f64>(&shape, &dtype, &device, grad_enabled, data_seed);
                test_add_two_tensors((tensor1, data1), (tensor2, data2));
            }
            DType::I64 => {
                let (tensor1, data1) = util::generate_random_tensor_and_corresponding_data::<i64>(&shape, &dtype, &device, grad_enabled, data_seed);
                data_seed += 1;
                let (tensor2, data2) = util::generate_random_tensor_and_corresponding_data::<i64>(&shape, &dtype, &device, grad_enabled, data_seed);
                test_add_two_tensors((tensor1, data1), (tensor2, data2));
            }
            DType::U32 => {
                let (tensor1, data1) = util::generate_random_tensor_and_corresponding_data::<u32>(&shape, &dtype, &device, grad_enabled, data_seed);
                data_seed += 1;
                let (tensor2, data2) = util::generate_random_tensor_and_corresponding_data::<u32>(&shape, &dtype, &device, grad_enabled, data_seed);
                test_add_two_tensors((tensor1, data1), (tensor2, data2));
            }
            DType::U8 => {
                let (tensor1, data1) = util::generate_random_tensor_and_corresponding_data::<u8>(&shape, &dtype, &device, grad_enabled, data_seed);
                data_seed += 1;
                let (tensor2, data2) = util::generate_random_tensor_and_corresponding_data::<u8>(&shape, &dtype, &device, grad_enabled, data_seed);
                test_add_two_tensors((tensor1, data1), (tensor2, data2));
            }

            _ => panic!("unexpected dtype {:?}", dtype),
        }
    }
}

fn test_stack_tensors<S>(datas: Vec<(Tensor, Vec<S>)>)
where
    S: candle_core::WithDType + std::fmt::Debug,
{
    let mut stacked_data: Vec<S> = Vec::with_capacity(
        datas.len() * datas[0].0.shape().unwrap().dims().iter().product::<usize>(),
    );
    let mut stacked_tensor = Vec::with_capacity(datas.len());

    datas.iter().for_each(|(tensor, data)| {
        stacked_tensor.push(tensor.clone());
        stacked_data.extend(data);
    });

    let stacked_tensor = Tensor::stack(&stacked_tensor, 0).unwrap();
    assert_eq!(stacked_tensor.to_vec::<S>().unwrap(), stacked_data);
}

proptest! {
    #[test]
    fn test_stack_tensors_on_axis_0_with_same_shape(
        num_tensors in 2..10,
        shape_seed in prop::num::u64::ANY,
        dtype in prop::sample::select(vec![DType::U8, DType::U32, DType::I64, DType::F32,DType::F64]),
        grad_enabled in prop::bool::ANY,
        data_seed in prop::num::u64::ANY,
    ) {
        let shape = util::generate_random_shape_with_appropriate_size(shape_seed);
        let device = Device::cpu();
        let mut data_seed= data_seed;

        match dtype {
            DType::F32 => {
                let datas = (0..num_tensors)
                    .map(|_| {
                        data_seed += 1;
                        util::generate_random_tensor_and_corresponding_data::<f32>(&shape, &dtype, &device, grad_enabled, data_seed)
                    })
                    .collect::<Vec<_>>();
                test_stack_tensors(datas);
            }
            DType::F64 => {
                let datas = (0..num_tensors)
                    .map(|_| {
                        data_seed += 1;
                        util::generate_random_tensor_and_corresponding_data::<f64>(&shape, &dtype, &device, grad_enabled, data_seed)
                    })
                    .collect::<Vec<_>>();
                test_stack_tensors(datas);
            }
            DType::I64 => {
                let datas = (0..num_tensors)
                    .map(|_| {
                        data_seed += 1;
                        util::generate_random_tensor_and_corresponding_data::<i64>(&shape, &dtype, &device, grad_enabled, data_seed)
                    })
                    .collect::<Vec<_>>();
                test_stack_tensors(datas);
            }
            DType::U32 => {
                let datas = (0..num_tensors)
                    .map(|_| {
                        data_seed += 1;
                        util::generate_random_tensor_and_corresponding_data::<u32>(&shape, &dtype, &device, grad_enabled, data_seed)
                    })
                    .collect::<Vec<_>>();
                test_stack_tensors(datas);
            }
            DType::U8 => {
                let datas = (0..num_tensors)
                    .map(|_| {
                        data_seed += 1;
                        util::generate_random_tensor_and_corresponding_data::<u8>(&shape, &dtype, &device, grad_enabled, data_seed)
                    })
                    .collect::<Vec<_>>();
                test_stack_tensors(datas);
            }
            _ => panic!("unexpected dtype {:?}", dtype),
        }
    }
}