Skip to content

[STABLE ABI] Port forced_align #4079

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

Open
wants to merge 5 commits into
base: main
Choose a base branch
from
Open

[STABLE ABI] Port forced_align #4079

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
ed1d81d
[STABLE ABI] Port forced_align
pearu Sep 1, 2025
c4478d2
Collect a torch::stable wishlist in src/libtorchaudio/stable
pearu Sep 18, 2025
e246029
Collect a torch::stable wishlist in src/libtorchaudio/stable
pearu Sep 18, 2025
bbf6f38
Apply pre-commit
pearu Sep 18, 2025
b308882
Add toString(ScalarType)
pearu Sep 18, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
136 changes: 78 additions & 58 deletions src/libtorchaudio/forced_align/cpu/compute.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,26 +1,27 @@
#include <torch/csrc/inductor/aoti_torch/c/shim.h>
#include <libtorchaudio/stable/dispatch.h>
#include <libtorchaudio/stable/ops.h>
#include <libtorchaudio/utils.h>
#include <torch/csrc/stable/library.h>
#include <torch/csrc/stable/ops.h>
#include <torch/csrc/stable/tensor.h>
#include <torch/script.h>
#include <torch/torch.h>

using namespace std;

namespace torchaudio {
namespace alignment {
namespace cpu {

using torch::headeronly::ScalarType;
using torch::stable::Tensor;

// Inspired from
// https://github.com/flashlight/sequence/blob/main/flashlight/lib/sequence/criterion/cpu/ConnectionistTemporalClassificationCriterion.cpp
template <typename scalar_t, at::ScalarType target_scalar_type>
template <typename scalar_t, ScalarType target_scalar_type>
void forced_align_impl(
const torch::Tensor& logProbs,
const torch::Tensor& targets,
const Tensor& logProbs,
const Tensor& targets,
const int64_t blank,
torch::Tensor& paths) {
Tensor& paths) {
const scalar_t kNegInfinity = -std::numeric_limits<scalar_t>::infinity();
using target_t = typename std::
conditional<target_scalar_type == torch::kInt, int, int64_t>::type;
conditional<target_scalar_type == ScalarType::Int, int, int64_t>::type;
const auto batchIndex =
0; // TODO: support batch version and use the real batch index
const auto T = logProbs.size(1);
Expand All @@ -36,17 +37,16 @@ void forced_align_impl(
for (int i = 0; i < T * S; i++) {
backPtr_a[i] = -1;
}

auto logProbs_a = logProbs.accessor<scalar_t, 3>();
auto targets_a = targets.accessor<target_t, 2>();
auto paths_a = paths.accessor<target_t, 2>();
auto logProbs_a = torchaudio::stable::accessor<scalar_t, 3>(logProbs);
auto targets_a = torchaudio::stable::accessor<target_t, 2>(targets);
auto paths_a = torchaudio::stable::accessor<target_t, 2>(paths);
auto R = 0;
for (auto i = 1; i < L; i++) {
if (targets_a[batchIndex][i] == targets_a[batchIndex][i - 1]) {
++R;
}
}
TORCH_CHECK(
STD_TORCH_CHECK(
T >= L + R,
"targets length is too long for CTC. Found log_probs length: ",
T,
Expand Down Expand Up @@ -138,73 +138,93 @@ void forced_align_impl(
delete[] backPtr_a;
}

std::tuple<torch::Tensor, torch::Tensor> compute(
const torch::Tensor& logProbs,
const torch::Tensor& targets,
const torch::Tensor& inputLengths,
const torch::Tensor& targetLengths,
std::tuple<Tensor, Tensor> compute(
const Tensor& logProbs,
const Tensor& targets,
const Tensor& inputLengths,
const Tensor& targetLengths,
const int64_t blank) {
TORCH_CHECK(logProbs.is_cpu(), "log_probs must be a CPU tensor");
TORCH_CHECK(targets.is_cpu(), "targets must be a CPU tensor");
TORCH_CHECK(
logProbs.device() == targets.device(),
"log_probs and targets need to be on the same device");
TORCH_CHECK(
logProbs.dtype() == torch::kFloat64 ||
logProbs.dtype() == torch::kFloat32 ||
logProbs.dtype() == torch::kFloat16,
STD_TORCH_CHECK(logProbs.is_cpu(), "log_probs must be a CPU tensor");
STD_TORCH_CHECK(targets.is_cpu(), "targets must be a CPU tensor");
STD_TORCH_CHECK(inputLengths.is_cpu(), "input_lengths must be a CPU tensor");
STD_TORCH_CHECK(
targetLengths.is_cpu(), "target_lengths must be a CPU tensor");
STD_TORCH_CHECK(
logProbs.scalar_type() == ScalarType::Double ||
logProbs.scalar_type() == ScalarType::Float ||
logProbs.scalar_type() == ScalarType::Half,
"log_probs must be float64, float32 or float16 (half) type");
TORCH_CHECK(
targets.dtype() == torch::kInt32 || targets.dtype() == torch::kInt64,
STD_TORCH_CHECK(
targets.scalar_type() == ScalarType::Int ||
targets.scalar_type() == ScalarType::Long,
"targets must be int32 or int64 type");
TORCH_CHECK(logProbs.is_contiguous(), "log_probs must be contiguous");
TORCH_CHECK(targets.is_contiguous(), "targets must be contiguous");
TORCH_CHECK(
STD_TORCH_CHECK(logProbs.is_contiguous(), "log_probs must be contiguous");
STD_TORCH_CHECK(targets.is_contiguous(), "targets must be contiguous");
STD_TORCH_CHECK(
logProbs.dim() == 3,
"log_probs must be 3-D (batch_size, input length, num classes)");
TORCH_CHECK(
STD_TORCH_CHECK(
targets.dim() == 2, "targets must be 2-D (batch_size, target length,)");
TORCH_CHECK(
STD_TORCH_CHECK(
inputLengths.dim() == 1, "input_lengths must be 1-D (batch_size,)");
TORCH_CHECK(
STD_TORCH_CHECK(
targetLengths.dim() == 1, "target_lengths must be 1-D (batch_size,)");
TORCH_CHECK(
STD_TORCH_CHECK(
logProbs.size(0) == 1,
"The batch dimension for log_probs must be 1 at the current version.")
TORCH_CHECK(
STD_TORCH_CHECK(
targets.size(0) == 1,
"The batch dimension for targets must be 1 at the current version.")
TORCH_CHECK(
STD_TORCH_CHECK(
blank >= 0 && blank < logProbs.size(-1),
"blank must be within [0, num classes)");

TORCH_CHECK(
logProbs.size(1) == at::max(inputLengths).item().toInt(),
"input length mismatch");
TORCH_CHECK(
targets.size(1) == at::max(targetLengths).item().toInt(),
"target length mismatch");

STABLE_DISPATCH_INDEX_TYPES(
inputLengths.scalar_type(), "forced_align_impl", [&] {
STD_TORCH_CHECK(
logProbs.size(1) == torchaudio::util::max<index_t>(inputLengths),
"input length mismatch");
});
STABLE_DISPATCH_INDEX_TYPES(
targetLengths.scalar_type(), "forced_align_impl", [&] {
STD_TORCH_CHECK(
targets.size(1) == torchaudio::util::max<index_t>(targetLengths),
"target length mismatch");
});
Copy link
Member

@NicolasHug NicolasHug Nov 6, 2025

Choose a reason for hiding this comment

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

On the above, can you help me understand why we need to use STABLE_DISPATCH_INDEX_TYPES now? Basically, which part of at::max(inputLengths).item().toInt() do we need to work-around? Is this a permanent workaround, or something we'll be able to simplify in the future when more APIs are ported to the stable part?

Copy link
Collaborator Author

@pearu pearu Nov 6, 2025

Choose a reason for hiding this comment

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

Currently, targetLengths can be a tensor with int32 or int64 items. If one uses, say, torchaudio::util::max<int32_t>(targetLengths) when targetLengths dtype is int64, the result of max may be wrong. The usage of STABLE_DISPATCH_INDEX_TYPES makes calling max safe.

We can simplify this in future after we are able to rewrite at::max(inputLengths).item().toInt() using stable ABI methods.

const auto B = logProbs.size(0);
const auto T = logProbs.size(1);
auto paths = torch::zeros(
{B, T},
torch::TensorOptions().device(targets.device()).dtype(targets.dtype()));
AT_DISPATCH_FLOATING_TYPES_AND_HALF(
Tensor paths = torchaudio::stable::new_zeros(targets, {B, T});

STABLE_DISPATCH_FLOATING_TYPES_AND_HALF(
logProbs.scalar_type(), "forced_align_impl", [&] {
if (targets.scalar_type() == torch::kInt64) {
forced_align_impl<scalar_t, torch::kInt64>(
if (targets.scalar_type() == ScalarType::Long) {
forced_align_impl<scalar_t, ScalarType::Long>(
logProbs, targets, blank, paths);
} else {
forced_align_impl<scalar_t, torch::kInt32>(
forced_align_impl<scalar_t, ScalarType::Int>(
logProbs, targets, blank, paths);
}
});
return std::make_tuple(paths, logProbs);
}

TORCH_LIBRARY_IMPL(torchaudio, CPU, m) {
m.impl("forced_align", &compute);
void boxed_forced_align_cpu(
StableIValue* stack,
uint64_t num_args,
uint64_t num_outputs) {
STD_TORCH_CHECK(num_args == 5, "num_args must be 5");
STD_TORCH_CHECK(num_outputs == 2, "num_outputs must be 2");
std::tuple<Tensor, Tensor> res = compute(
/*logProbs*/ torch::stable::detail::to<Tensor>(stack[0]),
/*targets*/ torch::stable::detail::to<Tensor>(stack[1]),
/*logit_lengths*/ torch::stable::detail::to<Tensor>(stack[2]),
/*target_lengths*/ torch::stable::detail::to<Tensor>(stack[3]),
/*blank*/ float(torch::stable::detail::to<int64_t>(stack[4])));
stack[0] = torch::stable::detail::from(std::get<0>(res));
stack[1] = torch::stable::detail::from(std::get<1>(res));
}

STABLE_TORCH_LIBRARY_IMPL(torchaudio, CPU, m) {
m.impl("forced_align", &boxed_forced_align_cpu);
}

} // namespace cpu
Expand Down
Loading
Loading