Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
5 changes: 4 additions & 1 deletion .github/workflows/compass-ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -207,10 +207,13 @@ jobs:

dependency-audit:
runs-on: ubuntu-24.04
permissions:
contents: read
checks: write
steps:
- uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6
- name: Audit Rust dependencies
uses: rustsec/audit-check@69366f33c96575abad1ee0dba8212993eecbe998 # v2
uses: rustsec/audit-check@858dc40f52ca2b8570b7a997c1c4e35c6fc9a432 # Node 24 update
with:
working-directory: .
token: ${{ secrets.GITHUB_TOKEN }}
Expand Down
59 changes: 57 additions & 2 deletions crates/compass-whisper/src/timing.rs
Original file line number Diff line number Diff line change
Expand Up @@ -72,6 +72,17 @@ pub fn dtw(x: &[f64], n: usize, m: usize) -> (Vec<usize>, Vec<usize>) {
let mut text_indices = Vec::new();
let mut time_indices = Vec::new();
while i > 0 || j > 0 {
// Boundary cells do not represent a text/time pair. They are
// unreachable for finite costs, but can be selected after a caller
// supplies a non-finite cost matrix.
if i == 0 {
j -= 1;
continue;
}
if j == 0 {
i -= 1;
continue;
}
text_indices.push(i - 1);
time_indices.push(j - 1);
match trace[i * w + j] {
Expand Down Expand Up @@ -179,8 +190,17 @@ pub fn find_alignment(
var += d * d;
}
let std = (var / seq_len as f64).sqrt();
for t in 0..seq_len {
w[t * half + f] = ((w[t * half + f] as f64 - mean) / std) as f32;
if std.is_finite() && std > 0.0 {
for t in 0..seq_len {
w[t * half + f] = ((w[t * half + f] as f64 - mean) / std) as f32;
}
} else {
// A constant or non-finite attention column has no usable
// alignment signal. Keep it neutral instead of creating NaNs
// that can send DTW through an invalid boundary path.
for t in 0..seq_len {
w[t * half + f] = 0.0;
}
}
}
median_filter_rows(w, half, medfilt_width);
Expand Down Expand Up @@ -478,6 +498,15 @@ mod tests {
assert!(alignment[2].word.is_empty());
}

#[test]
fn dtw_keeps_boundary_indices_in_range_for_non_finite_costs() {
let (text, time) = dtw(&[f64::NAN, 0.0], 2, 1);
assert_eq!(text.len(), time.len());
assert!(!text.is_empty());
assert!(text.iter().all(|&index| index < 2));
assert!(time.iter().all(|&index| index < 1));
}

#[test]
fn empty_alignment_and_segment_sets_are_noops() -> Result<()> {
let device = Device::Cpu;
Expand All @@ -489,6 +518,32 @@ mod tests {
Ok(())
}

#[test]
fn zero_cross_attention_has_finite_word_alignment() -> Result<()> {
let device = Device::Cpu;
let mut model = WhisperModel::for_test(tiny_config())?;
let tokenizer = get_tokenizer(false, 99, Some("en"), Some(Task::Transcribe))?;
let mel = Tensor::zeros((1, 2, 8), DType::F32, &device)?;

let alignment = find_alignment(
&mut model,
&tokenizer,
&tokenizer.encode(" hello"),
&mel,
8,
3,
1.0,
)?;
assert!(!alignment.is_empty());
assert!(alignment.iter().all(|timing| {
timing.start.is_finite()
&& timing.end.is_finite()
&& timing.start <= timing.end
&& timing.probability.is_finite()
}));
Ok(())
}

#[test]
fn tiny_random_model_exercises_alignment_and_segment_attachment() -> Result<()> {
let device = Device::Cpu;
Expand Down
Loading