SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2010.05646 · NeurIPS · 2020

HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis

Jungil Kong, Jaehyeon Kim, Jaekyoung Bae

arXiv · PDF · Open in the Atlas

Code that ran

17 of 25 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (11 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 14 have no contract verdict; 3 failures of our instrument; 8 no recorded run); 1 of them was also run on the host, not in the jail, by the peer-agreement lane that set its level, 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/2010.05646.json · badge: /api/ran/2010.05646.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
jik876/hifi-gan canonical (classifier label)
classifier basis: classifier_2025
5 ran · 0 no recorded run
TensorSpeech/TensorflowTTS community
Papers with Code archive
4 ran · 4 no recorded run
takaaki-saeki/ssl_speech_restoration community
Papers with Code archive
3 ran · 0 no recorded run
rishikksh20/HiFi-GAN reimplementation (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
keonlee9420/Comprehensive-E2E-TTS community
Papers with Code archive
2 ran · 0 no recorded run
maum-ai/univnet community
Papers with Code archive
0 ran · 3 no recorded run
NVIDIA/DeepLearningExamples community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
load_checkpoint Ran
ran; our contract draft was wrong, not the code
jik876/hifi-gan/inference.py
code served (permissive licence) · get_code("1644a2f2942ffe4c")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned dict. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T22:41:19Z). 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 os
import torch
import tempfile

torch.manual_seed(0)

# Create a temporary checkpoint file
with tempfile.NamedTemporaryFile(mode='wb', delete=False, suffix='.pt') as f:
    temp_filepath = f.name
    # Create a simple checkpoint dict and save it
    checkpoint_data = {
        'model_state': torch.randn(2, 4),
        'optimizer_state': torch.randn(3, 5),
        'epoch': 10
    }
    torch.save(checkpoint_data, temp_filepath)

ARGS = {
    'filepath': temp_filepath,
    'device': torch.device('cpu')
}
Setup block drafted by
a language model, Claude Haiku 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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
DiscriminatorP Ran
ran on a synthesized input · constructed an object; no result was computed
keonlee9420/Comprehensive-E2E-TTS/model/E2ETTS.py
pointer only (licence: NONE) · get_code("1cbd5983017dbccf")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, keonlee9420/Comprehensive-E2E-TTS 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.

Generator Ran
ran on a synthesized input · constructed an object; no result was computed
takaaki-saeki/ssl_speech_restoration/hifigan/models.py
code served (permissive licence) · get_code("a8588b06fe6ee564")
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, Generator: 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)

class HConfig:
    def __init__(self):
        self.feat_order = 80
        self.upsample_initial_channel = 512
        self.resblock = '1'
        self.upsample_rates = [8, 8, 2, 2]
        self.upsample_kernel_sizes = [16, 16, 4, 4]
        self.resblock_kernel_sizes = [3, 7, 11]
        self.resblock_dilation_sizes = [(1, 3, 5), (1, 3, 5), (1, 3, 5)]

h = HConfig()

ARGS = {
    'h': h
}
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 Generator — 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 Generator 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
Generator Ran
ran on a synthesized input · constructed an object; no result was computed
rishikksh20/HiFi-GAN/model/generator.py
code served (permissive licence) · get_code("315da86169871441")
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, Generator: 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
from torch import nn

torch.manual_seed(0)

# The Generator class requires:
# - input_channel: int (default 80)
# - hu: int (default 512)
# - ku: list of ints (default [16, 16, 4, 4])
# - kr: list of ints (default [3, 7, 11])
# - Dr: list of ints (default [1, 3, 5])

ARGS = {
    'input_channel': 80,
    'hu': 512,
    'ku': [16, 16, 4, 4],
    'kr': [3, 7, 11],
    'Dr': [1, 3, 5]
}
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 Generator — 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 Generator 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
Generator Ran
ran on a synthesized input
jik876/hifi-gan/models.py
code served (permissive licence) · get_code("a02c998d88e55794")
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; 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 torch
import numpy as np

# Config dict for Generator.__init__
INIT_ARGS = {
    'h': type('Config', (), {
        'upsample_initial_channel': 16,
        'resblock': '1',
        'resblock_kernel_sizes': [3, 5],
        'resblock_dilation_sizes': [(1, 3, 5), (1, 3, 5)],
        'upsample_rates': [4, 4],
        'upsample_kernel_sizes': [8, 8],
    })
}

# Args for forward pass
x = torch.randn(2, 80, 4)  # (batch, features, seq_len)
ARGS = {'x': x}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 3, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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)
MRF Ran
ran on a synthesized input · constructed an object; no result was computed
rishikksh20/HiFi-GAN/model/generator.py
code served (permissive licence) · get_code("296fba6a2e3878dd")
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, MRF: 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 = {
    'kernels': [3, 7, 11],
    'channel': 64,
    'dilations': [[1, 1], [3, 1], [5, 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 MRF — 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 MRF 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
MultiPeriodDiscriminator Ran
ran on a synthesized input · constructed an object; no result was computed
keonlee9420/Comprehensive-E2E-TTS/model/E2ETTS.py
pointer only (licence: NONE) · get_code("e51b96d3f2bca97c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, keonlee9420/Comprehensive-E2E-TTS 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.

ResBlock1 Ran
ran on a synthesized input · constructed an object; no result was computed
takaaki-saeki/ssl_speech_restoration/hifigan/models.py
code served (permissive licence) · get_code("6dd04160990ce7aa")
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, ResBlock1: 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)

class MockConfig:
    def __init__(self):
        self.channels = 64
        self.kernel_size = 3
        self.dilation = (1, 3, 5)

h = MockConfig()
channels = 64
kernel_size = 3
dilation = (1, 3, 5)

ARGS = {
    "h": h,
    "channels": channels,
    "kernel_size": kernel_size,
    "dilation": dilation,
}
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 ResBlock1 — 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 ResBlock1 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
ResBlock1 Ran
ran on a synthesized input
jik876/hifi-gan/models.py
code served (permissive licence) · get_code("e9c63f1724d52ed8")
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, 4, 16) 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
import torch.nn.functional as F
from torch.nn import Conv1d
from torch.nn.utils import weight_norm, remove_weight_norm

LRELU_SLOPE = 0.1

def get_padding(kernel_size, dilation=1):
    return int((kernel_size * dilation - dilation) / 2)

def init_weights(m, mean=0.0, std=0.01):
    classname = m.__class__.__name__
    if classname.find('Conv') != -1:
        m.weight.data.normal_(mean, std)

class ResBlock1(torch.nn.Module):

    def __init__(self, h, channels, kernel_size=3, dilation=(1, 3, 5)):
        super(ResBlock1, self).__init__()
        self.h = h
        self.convs1 = nn.ModuleList([weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[0], padding=get_padding(kernel_size, dilation[0]))), weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[1], padding=get_padding(kernel_size, dilation[1]))), weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[2], padding=get_padding(kernel_size, dilation[2])))])
        self.convs1.apply(init_weights)
        self.convs2 = nn.ModuleList([weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1, padding=get_padding(kernel_size, 1))), weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1, padding=get_padding(kernel_size, 1))), weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=1, padding=get_padding(kernel_size, 1)))])
        self.convs2.apply(init_weights)

    def forward(self, x):
        for c1, c2 in zip(self.convs1, self.convs2):
            xt = F.leaky_relu(x, LRELU_SLOPE)
            xt = c1(xt)
            xt = F.leaky_relu(xt, LRELU_SLOPE)
            xt = c2(xt)
            x = xt + x
        return x

    def remove_weight_norm(self):
        for l in self.convs1:
            remove_weight_norm(l)
        for l in self.convs2:
            remove_weight_norm(l)


INIT_ARGS = {
    'h': None,
    'channels': 4,
    'kernel_size': 3,
    'dilation': (1, 3, 5)
}

ARGS = {
    'x': torch.randn(2, 4, 16)
}
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, 4, 16) — shape (2, 4, 16) 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)
ResBlock2 Ran
ran on a synthesized input · constructed an object; no result was computed
takaaki-saeki/ssl_speech_restoration/hifigan/models.py
code served (permissive licence) · get_code("29a9e53b37e5a897")
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, ResBlock2: 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)

# Create a minimal config object for h
class ConfigStandIn:
    pass

h = ConfigStandIn()

# Define ARGS dict with all parameters for ResBlock2.__init__
ARGS = {
    'h': h,
    'channels': 4,
    'kernel_size': 3,
    'dilation': (1, 3)
}
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 ResBlock2 — 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 ResBlock2 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
ResBlock2 Ran
ran on a synthesized input
jik876/hifi-gan/models.py
code served (permissive licence) · get_code("d7373d3c5d20dd29")
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, 4, 16) 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
import torch.nn.functional as F
from torch.nn import Conv1d
from torch.nn.utils import weight_norm, remove_weight_norm

LRELU_SLOPE = 0.1

def get_padding(kernel_size, dilation=1):
    return int((kernel_size * dilation - dilation) / 2)

def init_weights(m, mean=0.0, std=0.01):
    classname = m.__class__.__name__
    if classname.find('Conv') != -1:
        m.weight.data.normal_(mean, std)

class ResBlock2(torch.nn.Module):

    def __init__(self, h, channels, kernel_size=3, dilation=(1, 3)):
        super(ResBlock2, self).__init__()
        self.h = h
        self.convs = nn.ModuleList([weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[0], padding=get_padding(kernel_size, dilation[0]))), weight_norm(Conv1d(channels, channels, kernel_size, 1, dilation=dilation[1], padding=get_padding(kernel_size, dilation[1])))])
        self.convs.apply(init_weights)

    def forward(self, x):
        for c in self.convs:
            xt = F.leaky_relu(x, LRELU_SLOPE)
            xt = c(xt)
            x = xt + x
        return x

    def remove_weight_norm(self):
        for l in self.convs:
            remove_weight_norm(l)


INIT_ARGS = {
    'h': None,
    'channels': 4,
    'kernel_size': 3,
    'dilation': (1, 3)
}

ARGS = {
    'x': torch.randn(2, 4, 16)
}

if __name__ == '__main__':
    model = ResBlock2(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Expected shape: {ARGS['x'].shape}")
    assert output.shape == ARGS['x'].shape, "Output shape mismatch"
    print("Test passed!")
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, 4, 16) — shape (2, 4, 16) 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)
ResStack Ran
ran on a synthesized input · constructed an object; no result was computed
rishikksh20/HiFi-GAN/model/generator.py
code served (permissive licence) · get_code("cec55aeff5fee127")
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, ResStack: 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 = {
    "kernel": 3,
    "channel": 4,
    "padding": 1,
    "dilations": [1, 3, 5]
}
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 ResStack — 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 ResStack 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
TFMultiHifiResBlock Ran
ran on a synthesized input · constructed an object; no result was computed
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("47841da7a9730802")
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, TFMultiHifiResBlock: 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 tensorflow as tf
import numpy as np

# Create a mock resblock layer that can be called with (x, training=...)
class MockResBlock(tf.keras.layers.Layer):
    def __init__(self):
        super().__init__()
        self.dense = tf.keras.layers.Dense(4)
    
    def call(self, x, training=False):
        # Process input and return output of same shape
        return self.dense(x)

# Create list of resblocks
list_resblock = [
    MockResBlock(),
    MockResBlock(),
    MockResBlock()
]

# Create input tensor (batch_size=2, feature_dim=4)
x = tf.constant(np.random.randn(2, 4).astype(np.float32))

ARGS = {
    "list_resblock": list_resblock
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type TFMultiHifiResBlock — 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 TFMultiHifiResBlock 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
TFReflectionPad1d Ran
ran on a synthesized input · constructed an object; no result was computed
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("4ab0d0201baa1a43")
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, TFReflectionPad1d: 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 tensorflow as tf
import numpy as np

# Create a small input tensor (B, T, C) = (2, 4, 3)
# B=2 (batch), T=4 (time steps), C=3 (channels)
input_tensor = tf.constant(np.random.randn(2, 4, 3), dtype=tf.float32)

ARGS = {
    'padding_size': 2,
    'padding_type': 'REFLECT',
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type TFReflectionPad1d — 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 TFReflectionPad1d 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
WeightNormalization Ran
ran on a synthesized input · constructed an object; no result was computed
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("b73dd2a7fa701451")
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, WeightNormalization: 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 tensorflow as tf
import numpy as np

# Create a simple Dense layer to wrap
base_layer = tf.keras.layers.Dense(
    units=8,
    activation='relu',
    input_shape=(4,)
)

# Create small dummy input data for data-dependent initialization
dummy_input = tf.constant(np.random.randn(2, 4).astype(np.float32))

ARGS = {
    'layer': base_layer,
    'data_init': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type WeightNormalization — 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 WeightNormalization 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
get_initializer Ran
ran; our contract draft was wrong, not the code
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("e3ecb51f94828d3b")
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 GlorotNormal. 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 tensorflow as tf

ARGS = {
    "initializer_seed": 42,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type GlorotNormal — a GlorotNormal output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a GlorotNormal 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
scan_checkpoint Ran
ran; our contract draft was wrong, not the code
jik876/hifi-gan/inference.py
code served (permissive licence) · get_code("404632cb4f4669da")
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) and returned str. 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 tempfile
import os

# Create a temporary directory for testing
temp_dir = tempfile.mkdtemp()

# Create some dummy checkpoint files to test the glob pattern
checkpoint_files = [
    os.path.join(temp_dir, "model_step_100.pt"),
    os.path.join(temp_dir, "model_step_200.pt"),
    os.path.join(temp_dir, "model_step_150.pt"),
]

# Create the dummy files
for cp_file in checkpoint_files:
    with open(cp_file, 'w') as f:
        f.write("dummy checkpoint")

ARGS = {
    "cp_dir": temp_dir,
    "prefix": "model_step_"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
BaseModel Not yet run
unverified (harvested, no recorded run)
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("3193c0ceab116156")
How it ran

no invocation recorded.

Generator Not yet run
unverified (harvested, no recorded run)
maum-ai/univnet/model/generator.py
code served (permissive licence) · get_code("793b82863c1fdc65")
How it ran

no invocation recorded.

KernelPredictor Not yet run
unverified (harvested, no recorded run)
maum-ai/univnet/model/generator.py
code served (permissive licence) · get_code("788fd01770c36aed")
How it ran

no invocation recorded.

LVCBlock Not yet run
unverified (harvested, no recorded run)
maum-ai/univnet/model/generator.py
code served (permissive licence) · get_code("ddf586275727eb6a")
How it ran

no invocation recorded.

TFConvTranspose1d Not yet run
unverified (harvested, no recorded run)
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("770d53e113359a0f")
How it ran

no invocation recorded.

TFHifiGANGenerator Not yet run
unverified (harvested, no recorded run)
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("ed16adff529066e2")
How it ran

no invocation recorded.

TFHifiResBlock Not yet run
unverified (harvested, no recorded run)
TensorSpeech/TensorflowTTS/tensorflow_tts/models/hifigan.py
code served (permissive licence) · get_code("90a00ab9b14ddecf")
How it ran

no invocation recorded.

_convert_ts_to_trt_hifigan Not yet run
unverified (harvested, no recorded run)
NVIDIA/DeepLearningExamples/PyTorch/SpeechSynthesis/HiFiGAN/models.py
pointer only (licence: NONE) · get_code("b170cd48d4da4c05")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, NVIDIA/DeepLearningExamples 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. No invocation recorded.

Also present in other papers' code

Of 25 samples lifted from this paper's linked repositories, 25 carry a function identity and 0 do not. Of the 23 distinct functions with an identity, 4 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 4 later) and 19 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
load_checkpoint 13 other papers
13 later
2106.07889 (2021-06)
UnivNet: A Neural Vocoder with Multi-Resolution Spectrogram Discriminators for High-Fidelity Waveform Generation
2203.02395 (2022-03)
iSTFTNet: Fast and Lightweight Mel-Spectrogram Vocoder Incorporating Inverse Short-Time Fourier Transform
2203.13086 (2022-03)
HiFi++: a Unified Framework for Bandwidth Extension and Speech Enhancement
and 10 more not named here
scan_checkpoint 10 other papers
10 later
2203.02395 (2022-03)
iSTFTNet: Fast and Lightweight Mel-Spectrogram Vocoder Incorporating Inverse Short-Time Fourier Transform
2203.13086 (2022-03)
HiFi++: a Unified Framework for Bandwidth Extension and Speech Enhancement
2305.13686 (2023-05)
MP-SENet: A Speech Enhancement Model with Parallel Denoising of Magnitude and Phase Spectra
and 7 more not named here
ResBlock1
brick of this paper's assembly: helper, model
1 other paper
1 later
2110.14513 (2021-10)
Neural Analysis and Synthesis: Reconstructing Speech from Self-Supervised Representations
ResBlock2
brick of this paper's assembly: helper, model
1 other paper
1 later
2110.14513 (2021-10)
Neural Analysis and Synthesis: Reconstructing Speech from Self-Supervised Representations

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

Several recent work on speech synthesis have employed generative adversarial networks (GANs) to produce raw waveforms. Although such methods improve the sampling efficiency and memory usage, their sample quality has not yet reached that of autoregressive and flow-based generative models. In this work, we propose HiFi-GAN, which achieves both efficient and high-fidelity speech synthesis. As speech audio consists of sinusoidal signals with various periods, we demonstrate that modeling periodic patterns of an audio is crucial for enhancing sample quality. A subjective human evaluation (mean opinion score, MOS) of a single speaker dataset indicates that our proposed method demonstrates similarity to human quality while generating 22.05 kHz high-fidelity audio 167.9 times faster than real-time on a single V100 GPU. We further show the generality of HiFi-GAN to the melspectrogram inversion of unseen speakers and end-to-end speech synthesis. Finally, a small footprint version of HiFi-GAN generates samples 13.4 times faster than real-time on CPU with comparable quality to an autoregressive counterpart. Voice is one of the most frequent and naturally used communication interfaces for humans. With recent developments in technology, voice is being used as a main interface in artificial intelligence (AI) voice assistant services such as Amazon Alexa, and it is also widely used in automobiles, smart homes and so forth. Accordingly, with the increase in demand for people to converse with machines, technology that synthesizes natural speech like human speech is being actively studied. Recently, with the development of neural networks, speech synthesis technology has made a rapid progress. Most neural speech synthesis models use a two-stage pipeline: 1) predicting a low resolution intermediate representation such as mel-spectrograms (

For agents

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

get_harvested_code_for_paper("2010.05646")
get_code_for_paper("2010.05646")
have("2010.05646")

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

curl https://syntology.ai/api/ran/2010.05646.json

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

[![Syntology run record](https://syntology.ai/api/ran/2010.05646.svg)](https://syntology.ai/paper/2010.05646)

Connect an agent — have() is free.