SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1803.09017 · 2018

Style Tokens: Unsupervised Style Modeling, Control and Transfer in End-to-End Speech Synthesis

arXiv · PDF · Open in the Atlas

Code that ran

19 of 21 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 12 have no contract verdict; 7 failures of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; 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/1803.09017.json · badge: /api/ran/1803.09017.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
cnlinxi/style-token_tacotron2 community (classifier label)
classifier basis not recorded on the link
5 ran · 1 no recorded run
jinhan/tacotron2-gst community (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
foamliu/GST-Tacotron-v2 community (classifier label)
classifier basis not recorded on the link
4 ran · 0 no recorded run
CODEJIN/GST_Tacotron community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
Kyubyong/expressive_tacotron reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
KinglittleQ/GST-Tacotron reimplementation (classifier label)
classifier basis: classifier_2025
0 ran · 1 no recorded run
FunctionStatusWhere it lives
Get_Path Ran
ran on a synthesized input
CODEJIN/GST_Tacotron/Get_Path.py
code served (permissive licence) · get_code("6611d63ef3343fb1")
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 tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 os
from random import sample
from unittest.mock import patch, MagicMock
import tempfile

# Create a temporary directory structure to simulate the audio dataset paths
temp_dir = tempfile.mkdtemp()

# Create mock subdirectories and files for each speaker
path_list = [
    ('LJ(F)', os.path.join(temp_dir, 'LJ', 'wavs')),
    ('CLB(F)', os.path.join(temp_dir, 'CLB', 'wav')),
    ('SLT(F)', os.path.join(temp_dir, 'SLT', 'wav')),
    ('AWB(M)', os.path.join(temp_dir, 'AWB', 'wav')),
    ('BDL(M)', os.path.join(temp_dir, 'BDL', 'wav')),
    ('JMK(M)', os.path.join(temp_dir, 'JMK', 'wav')),
    ('KSP(M)', os.path.join(temp_dir, 'KSP', 'wav')),
    ('RMS(M)', os.path.join(temp_dir, 'RMS', 'wav'))
]

# Create directories and dummy wav files
for tag, path in path_list:
    os.makedirs(path, exist_ok=True)
    for i in range(100):
        open(os.path.join(path, f'audio_{i:04d}.wav'), 'w').close()

ARGS = {
    'sample_count': 50
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
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
apply_gradient_allreduce Ran
ran on a synthesized input
jinhan/tacotron2-gst/distributed.py
pointer only (licence: BSD-3-Clause) · get_code("8b84f60b5ff9e153")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, jinhan/tacotron2-gst, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

conversion_helper Ran
ran on a synthesized input
jinhan/tacotron2-gst/fp16_optimizer.py
pointer only (licence: BSD-3-Clause) · get_code("f887d5ffcbc6ad3b")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, jinhan/tacotron2-gst, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

dynamic_range_compression Ran
ran; our contract draft was wrong, not the code
foamliu/GST-Tacotron-v2/audio_processing.py
code served (permissive licence) · get_code("83e2c862f7374ac7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) 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)

# Create a small tensor with positive values for dynamic range compression
# This simulates magnitude spectrogram or similar audio processing data
x = torch.rand(2, 4) * 10  # Random values between 0 and 10

# C is compression factor (default 1)
C = 1

# clip_val is minimum clipping value (default 1e-05)
clip_val = 1e-05

ARGS = {
    'x': x,
    'C': C,
    'clip_val': clip_val
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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
fp16_to_fp32 Ran
ran on a synthesized input
jinhan/tacotron2-gst/fp16_optimizer.py
pointer only (licence: BSD-3-Clause) · get_code("38034aa1d784464d")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, jinhan/tacotron2-gst, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

fp32_to_fp16 Ran
ran on a synthesized input
jinhan/tacotron2-gst/fp16_optimizer.py
pointer only (licence: BSD-3-Clause) · get_code("ea22baceb9a219a1")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, jinhan/tacotron2-gst, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

fused_add_tanh_sigmoid_multiply Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
foamliu/GST-Tacotron-v2/glow.py
pointer only (licence: MIT); the stored copy get_code serves, from isletennos/mmvc_trainer, is pointer only (licence: NONE) · get_code("b38f9c28c5117398")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from isletennos/mmvc_trainer, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

griffin_lim Ran
ran; our contract draft was wrong, not the code
jinhan/tacotron2-gst/audio_processing.py
pointer only (licence: BSD-3-Clause) · get_code("9f9ec26d4cdfdf7d")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, jinhan/tacotron2-gst, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

griffin_lim Ran
ran; our contract draft was wrong, not the code
foamliu/GST-Tacotron-v2/audio_processing.py
code served (permissive licence) · get_code("e274cfa39f3d6bf9")
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, 64) 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 numpy as np

torch.manual_seed(0)

class MockSTFT:
    """Mock STFT class with transform and inverse methods."""
    
    def transform(self, signal):
        """
        Returns magnitude and phase/angles.
        signal shape: (batch, time)
        Returns: (magnitudes, angles)
        """
        # Create synthetic magnitudes and angles
        batch_size = signal.shape[0]
        freq_bins = 5
        time_frames = 3
        
        magnitudes = torch.randn(batch_size, freq_bins, time_frames, dtype=torch.float32).abs()
        angles = torch.randn(batch_size, freq_bins, time_frames, dtype=torch.float32)
        
        return magnitudes, angles
    
    def inverse(self, magnitudes, angles):
        """
        Reconstructs signal from magnitudes and angles.
        magnitudes shape: (batch, freq_bins, time_frames)
        angles shape: (batch, freq_bins, time_frames)
        Returns: signal of shape (batch, 1, time) - note the 1 for squeeze(1)
        """
        # Simple reconstruction: create a signal that can be squeezed
        batch_size = magnitudes.shape[0]
        signal_length = 64
        
        # Convert angles to tensor if it's a Variable
        if isinstance(angles, torch.autograd.Variable):
            angles_val = angles.data
        else:
            angles_val = angles
        
        # Create synthetic signal
        signal = torch.randn(batch_size, 1, signal_length, dtype=torch.float32)
        
        return signal

# Create magnitudes spectrogram (batch=2, freq_bins=5, time_frames=3)
magnitudes = torch.randn(2, 5, 3, dtype=torch.float32).abs()

# Create STFT function object
stft_fn = MockSTFT()

ARGS = {
    "magnitudes": magnitudes,
    "stft_fn": stft_fn,
    "n_iters": 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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, 64) — shape (2, 64) 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
inv_preemphasis Ran
ran on a synthesized input
cnlinxi/style-token_tacotron2/datasets/audio.py
code served (permissive licence) · get_code("bd89c559f716b290")
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 ndarray, ndim 1, dtype family float, shape (6,) 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 numpy as np
from scipy import signal

# Set seed for reproducibility
torch.manual_seed(0)

# Create a small audio waveform (wav parameter)
# Using numpy array as scipy.signal.lfilter works with array-like inputs
wav = np.array([0.1, 0.2, -0.15, 0.05, 0.3, -0.1], dtype=np.float32)

# Preemphasis coefficient (k parameter)
# Typical values range from 0.95 to 0.99
k = 0.97

# Boolean flag for whether to apply inverse preemphasis
inv_preemphasize = True

ARGS = {
    "wav": wav,
    "k": k,
    "inv_preemphasize": inv_preemphasize,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (6,) — shape (6,) 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
inv_preemphasis Ran
ran on a synthesized input
CODEJIN/GST_Tacotron/Audio.py
code served (permissive licence) · get_code("f1b4c5cfbaef7ff5")
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 ndarray, ndim 1, dtype family float, shape (6,) 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 numpy as np
from scipy import signal

torch.manual_seed(0)

# The function inv_preemphasis takes:
# - x: input signal (array-like, typically numpy array or list)
# - preemphasis: float parameter (default 0.97)

# Create a small input signal for testing
x = np.array([0.1, 0.2, 0.3, 0.4, 0.5, 0.6], dtype=np.float32)

ARGS = {
    'x': x,
    'preemphasis': 0.97
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (6,) — shape (6,) 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
is_mulaw Ran
ran on a synthesized input
cnlinxi/style-token_tacotron2/wavenet_vocoder/util.py
code served (permissive licence) · get_code("203a81d997da01e5")
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 bool, ndim 0, dtype family float, shape () 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

def _assert_valid_input_type(s):
    assert s == 'mulaw-quantize' or s == 'mulaw' or s == 'raw'

def is_mulaw(s):
    _assert_valid_input_type(s)
    return s == 'mulaw'

ARGS = {
    's': 'mulaw'
}
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 bool, ndim 0, dtype family float, shape () — shape () 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
is_mulaw_quantize Ran
ran on a synthesized input
cnlinxi/style-token_tacotron2/wavenet_vocoder/util.py
code served (permissive licence) · get_code("9f7f281fd2dc3db3")
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 bool, ndim 0, dtype family float, shape () 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)

ARGS = {
    's': 'mulaw-quantize'
}
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 bool, ndim 0, dtype family float, shape () — shape () 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
is_raw Ran
ran on a synthesized input
cnlinxi/style-token_tacotron2/wavenet_vocoder/util.py
code served (permissive licence) · get_code("0ce2fef6ac56c0bd")
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 bool, ndim 0, dtype family float, shape () 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

def _assert_valid_input_type(s):
    assert s == 'mulaw-quantize' or s == 'mulaw' or s == 'raw'

def is_raw(s):
    _assert_valid_input_type(s)
    return s == 'raw'

ARGS = {
    's': 'raw'
}
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 bool, ndim 0, dtype family float, shape () — shape () 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
looper Ran
ran; our contract draft was wrong, not the code
Kyubyong/expressive_tacotron/synthesize.py
pointer only (licence: NONE) · get_code("3494c70cf20413e6")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Kyubyong/expressive_tacotron when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

preemphasis Ran
ran on a synthesized input
cnlinxi/style-token_tacotron2/datasets/audio.py
code served (permissive licence) · get_code("56a333de6e89087a")
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 ndarray, ndim 1, dtype family float, shape (6,) 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 numpy as np
from scipy import signal

# Create a small audio waveform (wav parameter)
wav = np.array([0.1, 0.2, 0.3, 0.4, 0.5, 0.6], dtype=np.float32)

# Create preemphasis coefficient (k parameter)
k = 0.97

# Create preemphasize flag (boolean parameter)
preemphasize = True

ARGS = {
    "wav": wav,
    "k": k,
    "preemphasize": preemphasize,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (6,) — shape (6,) 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
preemphasis Ran
ran on a synthesized input
CODEJIN/GST_Tacotron/Audio.py
code served (permissive licence) · get_code("31d5837f91baf6c1")
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 ndarray, ndim 1, dtype family float, shape (6,) 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 numpy as np
from scipy import signal

torch.manual_seed(0)

# The function preemphasis takes:
# - x: an array-like input (audio signal)
# - preemphasis: a float coefficient (default 0.97)
# 
# The function uses signal.lfilter which expects array-like inputs
# x can be a numpy array, list, or torch tensor

# Create a small audio signal
x = np.array([0.1, 0.2, 0.3, 0.4, 0.5, 0.6], dtype=np.float32)

ARGS = {
    "x": x,
    "preemphasis": 0.97
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (6,) — shape (6,) 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
remove Ran
ran; our contract draft was wrong, not the code
foamliu/GST-Tacotron-v2/glow.py
pointer only (licence: MIT); the stored copy get_code serves, from ANLGBOY/WaveNODE, is pointer only (licence: NONE) · get_code("b13a72233912c0b1")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ANLGBOY/WaveNODE, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

split_func Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from Rayhane-mamah/Tacotron-2
pointer only · get_code("b1491d3c2e709fdc")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

load_model Not yet run
unverified (harvested, no recorded run)
KinglittleQ/GST-Tacotron/generate.py
code served (permissive licence) · get_code("d3e8204b867d34ac")
How it ran

no invocation recorded.

norm_data Not yet run
unverified (harvested, no recorded run)
cnlinxi/style-token_tacotron2/preprocess.py
code served (permissive licence) · get_code("ae38e046c059c78f")
How it ran

no invocation recorded.

Also present in other papers' code

Of 21 samples lifted from this paper's linked repositories, 21 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 11 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 9 later) and 9 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
griffin_lim 14 other papers
14 later
1811.02122 (2018-11)
Robust and fine-grained prosody control of end-to-end speech synthesis
1812.04342 (2018-12)
Learning latent representations for style control and transfer in end-to-end speech synthesis
1907.04448 (2019-07)
Learning to Speak Fluently in a Foreign Language: Multilingual Speech Synthesis and Cross-Language Voice Cloning
and 11 more not named here
dynamic_range_compression 13 other papers
13 later
1811.02122 (2018-11)
Robust and fine-grained prosody control of end-to-end speech synthesis
1812.04342 (2018-12)
Learning latent representations for style control and transfer in end-to-end speech synthesis
1905.11286 (2019-05)
Stochastic Gradient Methods with Layer-wise Adaptive Moments for Training of Deep Networks
and 10 more not named here
fused_add_tanh_sigmoid_multiply 6 other papers
6 later · 1 only as an assembly brick
1811.00002 (2018-11)
WaveGlow: A Flow-based Generative Network for Speech Synthesis
2006.04598 (2020-06)
WaveNODE: A Continuous Normalizing Flow for Speech Synthesis
2106.06103 (2021-06)
Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech
and 3 more not named here
conversion_helper 4 other papers
4 later
1812.04342 (2018-12)
Learning latent representations for style control and transfer in end-to-end speech synthesis
1909.03329 (2019-09)
LAMOL: LAnguage MOdeling for Lifelong Language Learning
2203.10652 (2022-03)
Continual Sequence Generation with Adaptive Compositional Modules
and 1 more not named here
fp16_to_fp32 4 other papers
4 later
1812.04342 (2018-12)
Learning latent representations for style control and transfer in end-to-end speech synthesis
1909.03329 (2019-09)
LAMOL: LAnguage MOdeling for Lifelong Language Learning
2203.10652 (2022-03)
Continual Sequence Generation with Adaptive Compositional Modules
and 1 more not named here
fp32_to_fp16 4 other papers
4 later
1812.04342 (2018-12)
Learning latent representations for style control and transfer in end-to-end speech synthesis
1909.03329 (2019-09)
LAMOL: LAnguage MOdeling for Lifelong Language Learning
2203.10652 (2022-03)
Continual Sequence Generation with Adaptive Compositional Modules
and 1 more not named here
remove 3 other papers
3 later
1811.00002 (2018-11)
WaveGlow: A Flow-based Generative Network for Speech Synthesis
2006.04598 (2020-06)
WaveNODE: A Continuous Normalizing Flow for Speech Synthesis
2106.12699 (2021-06)
Distilling the Knowledge from Conditional Normalizing Flows
inv_preemphasis 2 other papers
2 later
2005.08209 (2020-05)
Learning Individual Speaking Styles for Accurate Lip to Speech Synthesis
2211.14758 (2022-11)
VideoReTalking: Audio-based Lip Synchronization for Talking Head Video Editing In the Wild
preemphasis 2 other papers
2 later
2005.08209 (2020-05)
Learning Individual Speaking Styles for Accurate Lip to Speech Synthesis
2211.14758 (2022-11)
VideoReTalking: Audio-based Lip Synchronization for Talking Head Video Editing In the Wild
looper 1 other paper
1 the same month
1803.09047 (2018-03)
Towards End-to-End Prosody Transfer for Expressive Speech Synthesis with Tacotron
split_func 1 other paper
1 dated earlier
1712.05884 (2017-12)
Natural TTS Synthesis by Conditioning WaveNet on Mel Spectrogram Predictions

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

In this work, we propose "global style tokens" (GSTs), a bank of embeddings that are jointly trained within Tacotron, a state-of-the-art end-to-end speech synthesis system. The embeddings are trained with no explicit labels, yet learn to model a large range of acoustic expressiveness. GSTs lead to a rich set of significant results. The soft interpretable "labels" they generate can be used to control synthesis in novel ways, such as varying speed and speaking style - independently of the text content. They can also be used for style transfer, replicating the speaking style of a single audio clip across an entire long-form text corpus. When trained on noisy, unlabeled found data, GSTs learn to factorize noise and speaker identity, providing a path towards highly scalable but robust speech synthesis.

For agents

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

get_harvested_code_for_paper("1803.09017")
get_code_for_paper("1803.09017")
have("1803.09017")

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

curl https://syntology.ai/api/ran/1803.09017.json

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

[![Syntology run record](https://syntology.ai/api/ran/1803.09017.svg)](https://syntology.ai/paper/1803.09017)

Connect an agent — have() is free.