Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[quant] Quantized flip dispatch #46235

Closed
wants to merge 4 commits into from
Closed

[quant] Quantized flip dispatch #46235

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
a4fa0a1
[quant] Quantized flip dispatch
z-a-f Oct 13, 2020
a0bae8d
Update on "[quant] Quantized flip dispatch"
z-a-f Oct 13, 2020
cd6ee10
Update on "[quant] Quantized flip dispatch"
z-a-f Oct 15, 2020
3fc2721
Update on "[quant] Quantized flip dispatch"
z-a-f Nov 2, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
33 changes: 24 additions & 9 deletions aten/src/ATen/native/TensorTransformations.cpp
View file
Open in desktop
Expand Up @@ -61,15 +61,30 @@ Tensor flip_cpu(const Tensor& self, IntArrayRef dims) {
}
}

AT_DISPATCH_ALL_TYPES_AND_COMPLEX_AND(at::ScalarType::Bool, in_tensor.scalar_type(), "flip_cpu", [&] {
flip_cpu_kernel<scalar_t>(
total_dims,
stride_contiguous_v,
flip_dims_b,
in_tensor,
out_tensor
);
});
if (in_tensor.is_quantized()) {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

if this pattern becomes common, then we probably want to create a new macro that include both AT_DISPATCH_ALL_TYPES_AND_COMPLEX_AND and AT_DISPATCH_QINT_AND_SUB_BYTE_TYPES

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree. Alternatively, I can copy-paste this code into the quantized folder, and register the quantized kernel there.

AT_DISPATCH_QINT_AND_SUB_BYTE_TYPES(in_tensor.scalar_type(),
"flip_quantized_cpu", [&] {
flip_cpu_kernel<scalar_t>(
total_dims,
stride_contiguous_v,
flip_dims_b,
in_tensor,
out_tensor
);
});
} else {
AT_DISPATCH_ALL_TYPES_AND_COMPLEX_AND(at::ScalarType::Bool,
in_tensor.scalar_type(),
"flip_cpu", [&] {
flip_cpu_kernel<scalar_t>(
total_dims,
stride_contiguous_v,
flip_dims_b,
in_tensor,
out_tensor
);
});
}

return out_tensor;
}
Expand Down
2 changes: 1 addition & 1 deletion aten/src/ATen/native/native_functions.yaml
View file
Open in desktop
Expand Up @@ -3533,7 +3533,7 @@
use_c10_dispatcher: full
variants: function, method
dispatch:
CPU: flip_cpu
CPU, QuantizedCPU: flip_cpu
CUDA: flip_cuda

- func: fliplr(Tensor self) -> Tensor
Expand Down