SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2606.20137 · 2026

PASQA: Pitch-Accent-Focused Speech Quality Assessment Model Trained on Synthetic Speech with Accent Errors

Kentaro Mitsui, Masaya Kawamura, Reo Shimizu, Yuma Shirahata

arXiv · PDF · Open in the Atlas

Code that ran

5 of 7 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 2 failures of our instrument; 2 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2606.20137.json · badge: /api/ran/2606.20137.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
lycorp-jp/PASQA community
Papers with Code archive
5 ran · 2 no recorded run
FunctionStatusWhere it lives
MoraCrossAttention Ran
ran on a synthesized input · constructed an object; no result was computed
lycorp-jp/PASQA/src/pasqa/_sheet/model.py
code served (permissive licence) · get_code("016f849e4fa57702")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, MoraCrossAttention: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Small encoder outputs: batch_size=2, seq_len=5, encoder_dim=16
encoder_outputs = torch.randn(2, 5, 16)

# Mora indices: batch_size=2, mora_seq_len=8, vocab indices in range [0, mora_vocab_size)
mora_idxs = torch.tensor([
    [1, 2, 3, 4, 0, 0, 0, 0],
    [2, 3, 4, 5, 6, 0, 0, 0]
])

# Mora lengths: batch_size=2
mora_lengths = torch.tensor([4, 5])

ARGS = {
    'encoder_dim': 16,
    'mora_vocab_size': 10,
    'mora_emb_dim': 8,
    'mora_transformer_layers': 1,
    'mora_transformer_heads': 2,
    'mora_ffn_dim': 32,
    'mora_dropout': 0.1,
    'mora_max_len': 128,
    'attn_dim': 8,
    'attn_heads': 2,
    'attn_dropout': 0.1,
    'attn_alpha_init': 0.1,
    'mora_pos_encoding': 'rope'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MoraCrossAttention — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MoraCrossAttention output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
Projection Ran
ran on a synthesized input · constructed an object; no result was computed
lycorp-jp/PASQA/src/pasqa/_sheet/model.py
code served (permissive licence) · get_code("e13b245c236fcc69")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, Projection: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'in_dim': 4,
    'hidden_dim': 8,
    'activation': nn.ReLU,
    'output_type': 'categorical',
    '_output_dim': 3,
    'output_step': 1.0,
    'range_clipping': False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Projection — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Projection output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
RotaryPositionalEmbedding Ran
ran on a synthesized input
lycorp-jp/PASQA/src/pasqa/_sheet/model.py
code served (permissive licence) · get_code("6b937564d9a16f82")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float, shape (2, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

class RotaryPositionalEmbedding(nn.Module):
    """Rotary Position Embedding (RoPE) implementation using PyTorch only."""

    def __init__(self, dim: int, max_seq_len: int=2048, base: float=10000.0):
        super().__init__()
        if dim % 2 != 0:
            raise ValueError(f'RotaryPositionalEmbedding requires even dim, got {dim}')
        self.dim = dim
        self.max_seq_len = max_seq_len
        self.base = base
        inv_freq = 1.0 / base ** (torch.arange(0, dim, 2).float() / dim)
        self.register_buffer('inv_freq', inv_freq, persistent=False)
        self._seq_len_cached = 0
        self._cos_cached = None
        self._sin_cached = None

    def _update_cache(self, seq_len: int, device: torch.device, dtype: torch.dtype):
        if seq_len > self._seq_len_cached:
            self._seq_len_cached = seq_len
            t = torch.arange(seq_len, device=device, dtype=dtype)
            freqs = torch.outer(t, self.inv_freq.to(dtype))
            emb = torch.cat([freqs, freqs], dim=-1)
            self._cos_cached = emb.cos()
            self._sin_cached = emb.sin()

    def _rotate_half(self, x: torch.Tensor) -> torch.Tensor:
        x1 = x[..., :x.shape[-1] // 2]
        x2 = x[..., x.shape[-1] // 2:]
        return torch.cat([-x2, x1], dim=-1)

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        batch, seq_len, dim = x.shape
        assert dim == self.dim, f'Expected dim={self.dim}, got {dim}'
        self._update_cache(seq_len, x.device, x.dtype)
        assert self._cos_cached is not None and self._sin_cached is not None
        cos = self._cos_cached[:seq_len]
        sin = self._sin_cached[:seq_len]
        return x * cos + self._rotate_half(x) * sin


INIT_ARGS = {
    'dim': 4,
    'max_seq_len': 2048,
    'base': 10000.0,
}

ARGS = {
    'x': torch.randn(2, 3, 4),
}

if __name__ == '__main__':
    model = RotaryPositionalEmbedding(**INIT_ARGS)
    output = model(**ARGS)
    print(f'Output shape: {output.shape}')
    print(f'Output dtype: {output.dtype}')
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
make_non_pad_mask Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lycorp-jp/PASQA/src/pasqa/_sheet/model.py
code served (permissive licence) · get_code("e56e75bf59733e04")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Test case 1: Basic usage with lengths as tensor
lengths_tensor = torch.tensor([2, 3])
xs_2d = torch.randn(2, 4)

# Test case 2: lengths as list
lengths_list = [2, 3]

ARGS = {
    "lengths": lengths_list,
    "xs": xs_2d,
    "length_dim": -1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_pad_mask Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lycorp-jp/PASQA/src/pasqa/_sheet/model.py
code served (permissive licence) · get_code("4f4f6b69a51a9e11")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Test case 1: Basic usage with lengths as tensor and xs provided
lengths_tensor = torch.tensor([2, 3])
xs = torch.randn(2, 4, 5)  # batch_size=2, length_dim=-1, seq_len=5

ARGS = {
    "lengths": lengths_tensor,
    "xs": xs,
    "length_dim": -1,
    "maxlen": None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 4, 5) — shape (2, 4, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
SSLMOS Not yet run
unverified (harvested, no recorded run)
lycorp-jp/PASQA/src/pasqa/_sheet/model.py
code served (permissive licence) · get_code("cdc703842b85e610")
How it ran

no invocation recorded.

_patch_torchaudio_for_s3prl Not yet run
unverified (harvested, no recorded run)
lycorp-jp/PASQA/src/pasqa/_sheet/model.py
code served (permissive licence) · get_code("fa7a9118df7f4ce2")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 0 later) and 5 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
make_pad_mask
brick of this paper's assembly: helper
3 other papers
3 dated earlier
2110.06280 (2021-10)
S3PRL-VC: Open-source Voice Conversion Framework with Self-supervised Speech Representations
2212.04559 (2022-12)
SpeechLMScore: Evaluating speech generation using speech language model
2505.15061 (2025-05)
SHEET: A Multi-purpose Open-source Speech Human Evaluation Estimation Toolkit
make_non_pad_mask
brick of this paper's assembly: helper
2 other papers
2 dated earlier
2110.06280 (2021-10)
S3PRL-VC: Open-source Voice Conversion Framework with Self-supervised Speech Representations
2505.15061 (2025-05)
SHEET: A Multi-purpose Open-source Speech Human Evaluation Estimation Toolkit

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Existing mean opinion score (MOS) prediction models typically predict utterance-level naturalness MOS and can be insensitive to localized pitch-accent errors. We propose Pitch-Accentfocused Speech Quality Assessment (PASQA), which explicitly targets pitch-accent correctness. To train our model, we construct a controlled Japanese accent-error dataset by changing accent patterns using an accent-controllable text-to-speech system, and compute a pseudo accent-quality score from the accent-error rate. PASQA builds on self-supervised representations and employs mora-conditioned fusion, ranking loss, an auxiliary accent-error localization task, and speaker-invariant training. Experiments show that conventional models fail to preserve the ordering by accent-error severity, whereas PASQA achieves high ordering accuracy on both seen and unseen speakers. Further, PASQA shows stronger agreement with human accent-correctness judgments. The code is available at https: //github.com/lycorp-jp/PASQA.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2606.20137")
get_code_for_paper("2606.20137")
have("2606.20137")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2606.20137.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2606.20137.svg)](https://syntology.ai/paper/2606.20137)

Connect an agent — have() is free.