-
Notifications
You must be signed in to change notification settings - Fork 231
/
Copy pathlltm_cuda.cpp
83 lines (70 loc) · 2.05 KB
/
lltm_cuda.cpp
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
#include <torch/extension.h>
#include <vector>
// CUDA forward declarations
std::vector<torch::Tensor> lltm_cuda_forward(
torch::Tensor input,
torch::Tensor weights,
torch::Tensor bias,
torch::Tensor old_h,
torch::Tensor old_cell);
std::vector<torch::Tensor> lltm_cuda_backward(
torch::Tensor grad_h,
torch::Tensor grad_cell,
torch::Tensor new_cell,
torch::Tensor input_gate,
torch::Tensor output_gate,
torch::Tensor candidate_cell,
torch::Tensor X,
torch::Tensor gate_weights,
torch::Tensor weights);
// C++ interface
// NOTE: AT_ASSERT has become AT_CHECK on master after 0.4.
#define CHECK_CUDA(x) AT_ASSERTM(x.type().is_cuda(), #x " must be a CUDA tensor")
#define CHECK_CONTIGUOUS(x) AT_ASSERTM(x.is_contiguous(), #x " must be contiguous")
#define CHECK_INPUT(x) CHECK_CUDA(x); CHECK_CONTIGUOUS(x)
std::vector<torch::Tensor> lltm_forward(
torch::Tensor input,
torch::Tensor weights,
torch::Tensor bias,
torch::Tensor old_h,
torch::Tensor old_cell) {
CHECK_INPUT(input);
CHECK_INPUT(weights);
CHECK_INPUT(bias);
CHECK_INPUT(old_h);
CHECK_INPUT(old_cell);
return lltm_cuda_forward(input, weights, bias, old_h, old_cell);
}
std::vector<torch::Tensor> lltm_backward(
torch::Tensor grad_h,
torch::Tensor grad_cell,
torch::Tensor new_cell,
torch::Tensor input_gate,
torch::Tensor output_gate,
torch::Tensor candidate_cell,
torch::Tensor X,
torch::Tensor gate_weights,
torch::Tensor weights) {
CHECK_INPUT(grad_h);
CHECK_INPUT(grad_cell);
CHECK_INPUT(input_gate);
CHECK_INPUT(output_gate);
CHECK_INPUT(candidate_cell);
CHECK_INPUT(X);
CHECK_INPUT(gate_weights);
CHECK_INPUT(weights);
return lltm_cuda_backward(
grad_h,
grad_cell,
new_cell,
input_gate,
output_gate,
candidate_cell,
X,
gate_weights,
weights);
}
TORCH_LIBRARY_IMPL(myops, CUDA, m){
m.impl(TORCH_SELECTIVE_NAME("lltm"), TORCH_FN(lltm_forward));
m.impl(TORCH_SELECTIVE_NAME("lltm.backward"), TORCH_FN(lltm_backward));
}