SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.09661 · 2020

Implicit Neural Representations with Periodic Activation Functions

Vincent Sitzmann, Julien Martel, David Lindell, Gordon Wetzstein, Alexander Bergman

arXiv · PDF · Open in the Atlas

Code that ran

54 of 71 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (30 of them constructed an object rather than computing a result; 4 honoured the contract we drafted, 0 violated it and 37 have no contract verdict; 13 failures of our instrument; 17 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/2006.09661.json · badge: /api/ran/2006.09661.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
vsitzmann/siren canonical (classifier label)
classifier basis: census
12 ran · 2 no recorded run
TalFurman/Implict_neural_representation_of_images community
Papers with Code archive
8 ran · 8 no recorded run
linusnie/diffcd community
Papers with Code archive
4 ran · 3 no recorded run
marccoru/locationencoder community
Papers with Code archive
4 ran · 0 no recorded run
dalmia/siren community
Papers with Code archive
3 ran · 0 no recorded run
titu1994/tf_SIREN community
Papers with Code archive
3 ran · 0 no recorded run
zion-king/Tensorflow-Sinusodial-Representation-Networks-SIREN community
Papers with Code archive
3 ran · 0 no recorded run
palombom/sirenmri community
Papers with Code archive
3 ran · 0 no recorded run
ecovision-uzh/climplicit community
Papers with Code archive
3 ran · 0 no recorded run
mjhwright/error-correction community
Papers with Code archive
2 ran · 1 no recorded run
arihaz/ImplicitNeuralRepresentations community
Papers with Code archive
2 ran · 0 no recorded run
MaximeVandegar/Papers-in-100-Lines-of-Code community
Papers with Code archive
2 ran · 0 no recorded run
jankrepl/mildlyoverfitted community
Papers with Code archive
1 ran · 1 no recorded run
Daniel-H-99/SIREN community
Papers with Code archive
1 ran · 0 no recorded run
locuslab/impsq community
Papers with Code archive
1 ran · 0 no recorded run
lucidrains/siren-pytorch community
Papers with Code archive
1 ran · 0 no recorded run
scart97/Siren-fastai2 community
Papers with Code archive
1 ran · 0 no recorded run
TheStageAI/TorchIntegral community
Papers with Code archive
0 ran · 2 no recorded run
FunctionStatusWhere it lives
BatchLinear Ran
ran on a synthesized input
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("976f5acdf961701c")
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
from collections import OrderedDict

# INIT_ARGS for BatchLinear.__init__
# BatchLinear inherits from nn.Linear, so we need in_features and out_features
INIT_ARGS = {
    'in_features': 5,
    'out_features': 7,
    'bias': True
}

# ARGS for BatchLinear.forward
# Create input tensor and params dict
# The params dict should contain 'weight' and optionally 'bias'
# For batched linear, weight can have extra batch dimensions

batch_size = 2
seq_len = 3
in_features = 5
out_features = 7

# Input tensor: (batch_size, seq_len, in_features)
input_tensor = torch.randn(batch_size, seq_len, in_features)

# Batched weight: (batch_size, out_features, in_features)
batched_weight = torch.randn(batch_size, out_features, in_features)

# Batched bias: (batch_size, out_features)
batched_bias = torch.randn(batch_size, out_features)

params_dict = OrderedDict([
    ('weight', batched_weight),
    ('bias', batched_bias)
])

ARGS = {
    'input': input_tensor,
    'params': params_dict
}
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)
FCBlock Ran
ran on a synthesized input · constructed an object; no result was computed
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("8ede319230f3b9c5")
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, FCBlock: 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
import numpy as np
from collections import OrderedDict
import math
import re

torch.manual_seed(0)

# The function signature is:
# def __init__(self, in_features, out_features, num_hidden_layers, hidden_features, 
#              outermost_linear=False, nonlinearity='relu', weight_init=None):

ARGS = {
    'in_features': 4,
    'out_features': 3,
    'num_hidden_layers': 2,
    'hidden_features': 8,
    'outermost_linear': False,
    'nonlinearity': 'relu',
    'weight_init': None,
}
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 FCBlock — 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 FCBlock 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
ImageDownsampling Ran
ran on a synthesized input · constructed an object; no result was computed
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("1ad216217b5c6292")
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, ImageDownsampling: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

class ImageDownsampling(nn.Module):
    """Generate samples in u,v plane according to downsampling blur kernel"""

    def __init__(self, sidelength, downsample=False):
        super().__init__()
        if isinstance(sidelength, int):
            self.sidelength = (sidelength, sidelength)
        else:
            self.sidelength = sidelength
        if self.sidelength is not None:
            self.sidelength = torch.Tensor(self.sidelength).float()
        else:
            assert downsample is False
        self.downsample = downsample

    def forward(self, coords):
        if self.downsample:
            return coords + self.forward_bilinear(coords)
        else:
            return coords

    def forward_box(self, coords):
        return 2 * (torch.rand_like(coords) - 0.5) / self.sidelength

    def forward_bilinear(self, coords):
        Y = torch.sqrt(torch.rand_like(coords)) - 1
        Z = 1 - torch.sqrt(torch.rand_like(coords))
        b = torch.rand_like(coords) < 0.5
        Q = (b * Y + ~b * Z) / self.sidelength
        return Q

ARGS = {
    "sidelength": 32,
    "downsample": True
}
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 ImageDownsampling — 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 ImageDownsampling 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
PosEncodingNeRF Ran
ran on a synthesized input · constructed an object; no result was computed
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("96a71871fc866141")
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, PosEncodingNeRF: 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 numpy as np
import math
from torch import nn

torch.manual_seed(0)

# Test case 1: in_features=3 (3D coordinates like in NeRF)
ARGS = {
    'in_features': 3,
    'sidelength': None,
    'fn_samples': None,
    'use_nyquist': True,
}

# Alternative test cases could be:
# For 2D coordinates:
# ARGS = {
#     'in_features': 2,
#     'sidelength': 64,
#     'fn_samples': None,
#     'use_nyquist': True,
# }

# For 1D coordinates:
# ARGS = {
#     'in_features': 1,
#     'sidelength': None,
#     'fn_samples': 128,
#     'use_nyquist': True,
# }
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 PosEncodingNeRF — 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 PosEncodingNeRF 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
RBFLayer Ran
ran on a synthesized input · constructed an object; no result was computed
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("74c875b042d85c09")
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, RBFLayer: 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
import numpy as np

torch.manual_seed(0)

# RBFLayer.__init__ takes in_features and out_features
in_features = 4
out_features = 3

ARGS = {
    "in_features": in_features,
    "out_features": out_features,
}
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 RBFLayer — 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 RBFLayer 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
SIREN Ran
ran on a synthesized input · constructed an object; no result was computed
dalmia/siren/siren/siren.py
code served (permissive licence) · get_code("d974796661d1e3ea")
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, SIREN: 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 = {
    'layers': [16, 32, 16],
    'in_features': 4,
    'out_features': 3,
    'w0': 1.0,
    'w0_initial': 30.0,
    'bias': True,
    'initializer': 'siren',
    'c': 6
}
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 SIREN — 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 SIREN 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
SIREN Ran
ran on a synthesized input · constructed an object; no result was computed
Daniel-H-99/SIREN/models.py
pointer only (licence: NONE) · get_code("25b93817248bf804")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Daniel-H-99/SIREN 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.

SIREN Ran
ran on a synthesized input · constructed an object; no result was computed
locuslab/impsq/modules/models.py
pointer only (licence: NONE) · get_code("9dbc91c61f8d9f89")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, locuslab/impsq 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.

SIRENInitializer Ran
ran on a synthesized input · constructed an object; no result was computed
titu1994/tf_SIREN/tf_siren/siren.py
code served (permissive licence) · get_code("b3b681c70d532217")
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, SIRENInitializer: 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

ARGS = {
    'w0': 1.0,
    'c': 6.0,
    '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 SIRENInitializer — 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 SIRENInitializer 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
SIRENInitializer Ran
ran on a synthesized input · constructed an object; no result was computed
zion-king/Tensorflow-Sinusodial-Representation-Networks-SIREN/tf_siren/siren.py
code served (permissive licence) · get_code("7850486bfddc6c83")
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, SIRENInitializer: 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

ARGS = {
    'w0': 1.0,
    'c': 6.0,
    '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 SIRENInitializer — 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 SIRENInitializer 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
Sine Ran
ran on a synthesized input · constructed an object; no result was computed
marccoru/locationencoder/locationencoder/nn/siren.py
pointer only (licence: NONE) · get_code("9ea2ac2f593eb96a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, marccoru/locationencoder 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.

Sine Ran
ran on a synthesized input · constructed an object; no result was computed
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("7bbf8646a3306d9e")
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, Sine: 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)

class Sine(nn.Module):
    def __init__(self):
        super().__init__()

    def forward(self, input):
        return torch.sin(30 * input)

# The Sine class only takes self as parameter in __init__
# The forward method takes input tensor
# We need to construct valid arguments for creating an instance of Sine

ARGS = {}
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 Sine — 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 Sine 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
Sine Ran
ran on a synthesized input · constructed an object; no result was computed
titu1994/tf_SIREN/tf_siren/siren.py
code served (permissive licence) · get_code("1d752eb8bb968496")
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, Sine: 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

class Sine(tf.keras.layers.Layer):
    def __init__(self, w0: float = 1.0, **kwargs):
        """
        Sine activation function with w0 scaling support.

        Args:
            w0: w0 in the activation step `act(x; w0) = sin(w0 * x)`
        """
        super(Sine, self).__init__(**kwargs)
        self.w0 = w0

    def call(self, inputs):
        return tf.sin(self.w0 * inputs)

    def get_config(self):
        config = {'w0': self.w0}
        base_config = super(Sine, self).get_config()
        return dict(list(base_config.items()) + list(config.items()))

ARGS = {
    'w0': 2.5,
}
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 Sine — 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 Sine 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
Sine Ran
ran on a synthesized input · constructed an object; no result was computed
mjhwright/error-correction/network.py
code served (permissive licence) · get_code("dc2056da4db84fc9")
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, Sine: 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

torch.manual_seed(0)

class Sine(torch.nn.Module):
    def __init__(self):
        super().__init__()

    def forward(self, input):
        return torch.sin(input)

# The Sine class __init__ takes no parameters beyond self
# So ARGS should be an empty dict since there are no parameters to pass
ARGS = {}
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 Sine — 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 Sine 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
Sine Ran
ran on a synthesized input · constructed an object; no result was computed
palombom/sirenmri/siren.py
code served (permissive licence) · get_code("377aa085c43d01c6")
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, Sine: 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)

ARGS = {
    "w0": 2.0,
}
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 Sine — 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 Sine 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
Sine Ran
ran on a synthesized input · constructed an object; no result was computed
dalmia/siren/siren/siren.py
code served (permissive licence) · get_code("5dfb17433aff0114")
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, Sine: 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)

# Construct arguments for Sine.__init__
ARGS = {
    'w0': 2.5
}
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 Sine — 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 Sine 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
Sine Ran
ran on a synthesized input · constructed an object; no result was computed
ecovision-uzh/climplicit/quickstart/siren.py
code served (permissive licence) · get_code("2389bbb56014ff60")
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, Sine: 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)

ARGS = {
    'w0': 2.0
}
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 Sine — 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 Sine 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
Sine Ran
ran on a synthesized input · constructed an object; no result was computed
zion-king/Tensorflow-Sinusodial-Representation-Networks-SIREN/tf_siren/siren.py
code served (permissive licence) · get_code("7c31228c86957fbf")
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, Sine: 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 test input tensor
test_input = tf.constant(np.random.randn(2, 4).astype(np.float32))

# Define ARGS dict with all parameters for Sine
ARGS = {
    'w0': 2.0,
}
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 Sine — 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 Sine 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
SineLayer Ran
ran on a synthesized input · constructed an object; no result was computed
jankrepl/mildlyoverfitted/github_adventures/siren/core.py
code served (permissive licence) · get_code("2043cfd7f293fb9e")
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, SineLayer: 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
import numpy as np

torch.manual_seed(0)

ARGS = {
    "in_features": 4,
    "out_features": 3,
    "bias": True,
    "is_first": False,
    "omega": 30,
    "custom_init_function_": None,
}
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 SineLayer — 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 SineLayer 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
SineLayer Ran
ran on a synthesized input · constructed an object; no result was computed
arihaz/ImplicitNeuralRepresentations/network.py
pointer only (licence: NONE) · get_code("7877368eae84c292")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, arihaz/ImplicitNeuralRepresentations 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.

SineLayer Ran
ran on a synthesized input
MaximeVandegar/Papers-in-100-Lines-of-Code/Implicit_Neural_Representations_with_Periodic_Activation_Functions/siren.py
code served (permissive licence) · get_code("4b770f57b782b848")
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 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
import torch.nn as nn

class SineLayer(nn.Module):

    def __init__(self, w0):
        super(SineLayer, self).__init__()
        self.w0 = w0

    def forward(self, x):
        return torch.sin(self.w0 * x)


INIT_ARGS = {
    'w0': 1.0
}

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

if __name__ == '__main__':
    layer = SineLayer(**INIT_ARGS)
    output = layer(**ARGS)
    print(output)
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 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
SingleBVPNet Ran
ran on a synthesized input · constructed an object; no result was computed
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("92f860754ce944d1")
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, SingleBVPNet: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
import numpy as np
from collections import OrderedDict
import math

torch.manual_seed(0)

class ImageDownsampling(nn.Module):
    """Generate samples in u,v plane according to downsampling blur kernel"""

    def __init__(self, sidelength, downsample=False):
        super().__init__()
        if isinstance(sidelength, int):
            self.sidelength = (sidelength, sidelength)
        else:
            self.sidelength = sidelength
        if self.sidelength is not None:
            self.sidelength = torch.Tensor(self.sidelength).float()
        else:
            assert downsample is False
        self.downsample = downsample

    def forward(self, coords):
        if self.downsample:
            return coords + self.forward_bilinear(coords)
        else:
            return coords

    def forward_box(self, coords):
        return 2 * (torch.rand_like(coords) - 0.5) / self.sidelength

    def forward_bilinear(self, coords):
        Y = torch.sqrt(torch.rand_like(coords)) - 1
        Z = 1 - torch.sqrt(torch.rand_like(coords))
        b = torch.rand_like(coords) < 0.5
        Q = (b * Y + ~b * Z) / self.sidelength
        return Q


class PosEncodingNeRF(nn.Module):
    """Module to add positional encoding as in NeRF [Mildenhall et al. 2020]."""

    def __init__(self, in_features, sidelength=None, fn_samples=None, use_nyquist=True):
        super().__init__()
        self.in_features = in_features
        if self.in_features == 3:
            self.num_frequencies = 10
        elif self.in_features == 2:
            assert sidelength is not None
            if isinstance(sidelength, int):
                sidelength = (sidelength, sidelength)
            self.num_frequencies = 4
            if use_nyquist:
                self.num_frequencies = self.get_num_frequencies_nyquist(min(sidelength[0], sidelength[1]))
        elif self.in_features == 1:
            assert fn_samples is not None
            self.num_frequencies = 4
            if use_nyquist:
                self.num_frequencies = self.get_num_frequencies_nyquist(fn_samples)
        self.out_dim = in_features + 2 * in_features * self.num_frequencies

    def get_num_frequencies_nyquist(self, samples):
        nyquist_rate = 1 / (2 * (2 * 1 / samples))
        return int(math.floor(math.log(nyquist_rate, 2)))

    def forward(self, coords):
        coords = coords.view(coords.shape[0], -1, self.in_features)
        coords_pos_enc = coords
        for i in range(self.num_frequencies):
            for j in range(self.in_features):
                c = coords[..., j]
                sin = torch.unsqueeze(torch.sin(2 ** i * np.pi * c), -1)
                cos = torch.unsqueeze(torch.cos(2 ** i * np.pi * c), -1)
                coords_pos_enc = torch.cat((coords_pos_enc, sin, cos), axis=-1)
        return coords_pos_enc.reshape(coords.shape[0], -1, self.out_dim)


class RBFLayer(nn.Module):
    """Transforms incoming data using a given radial basis function."""

    def __init__(self, in_features, out_features):
        super().__init__()
        self.in_features = in_features
        self.out_features = out_features
        self.centres = nn.Parameter(torch.Tensor(out_features, in_features))
        self.sigmas = nn.Parameter(torch.Tensor(out_features))
        self.reset_parameters()
        self.freq = nn.Parameter(np.pi * torch.ones((1, self.out_features)))

    def reset_parameters(self):
        nn.init.uniform_(self.centres, -1, 1)
        nn.init.uniform_(self.sigmas, 0.5, 1.5)


ARGS = {}
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 SingleBVPNet — 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 SingleBVPNet 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
SinusodialRepresentationDense Ran
ran on a synthesized input · constructed an object; no result was computed
titu1994/tf_SIREN/tf_siren/siren.py
code served (permissive licence) · get_code("e7f217d13dd1e5d9")
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, SinusodialRepresentationDense: 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

ARGS = {
    'units': 16,
    'w0': 1.0,
    'c': 6.0,
    'activation': 'sine',
    'use_bias': True,
    'kernel_initializer': 'siren_uniform',
    'bias_initializer': 'he_uniform',
    'kernel_regularizer': None,
    'bias_regularizer': None,
    'activity_regularizer': None,
    'kernel_constraint': None,
    'bias_constraint': None,
}
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 SinusodialRepresentationDense — 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 SinusodialRepresentationDense 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
SinusodialRepresentationDense Ran
ran on a synthesized input · constructed an object; no result was computed
zion-king/Tensorflow-Sinusodial-Representation-Networks-SIREN/tf_siren/siren.py
code served (permissive licence) · get_code("184b236f39ccaca1")
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, SinusodialRepresentationDense: 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 sample input tensor
batch_size = 2
input_dim = 4
sample_input = tf.constant(np.random.randn(batch_size, input_dim), dtype=tf.float32)

# Define all arguments for SinusodialRepresentationDense
ARGS = {
    'units': 3,
    'w0': 1.0,
    'c': 6.0,
    'activation': 'sine',
    'use_bias': True,
    'kernel_initializer': 'siren_initializer',
    'bias_initializer': 'he_uniform',
    'kernel_regularizer': None,
    'bias_regularizer': None,
    'activity_regularizer': None,
    'kernel_constraint': None,
    'bias_constraint': None,
}
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 SinusodialRepresentationDense — 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 SinusodialRepresentationDense 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
Siren Ran
ran on a synthesized input · constructed an object; no result was computed
marccoru/locationencoder/locationencoder/nn/siren.py
pointer only (licence: NONE) · get_code("2aa2597d8bd3bb93")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, marccoru/locationencoder 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.

Siren Ran
ran on a synthesized input · constructed an object; no result was computed
lucidrains/siren-pytorch/siren_pytorch/siren_pytorch.py
code served (permissive licence) · get_code("5b35db3519256685")
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, Siren: 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 math
import torch
from torch import nn
import torch.nn.functional as F

torch.manual_seed(0)

class Sine(nn.Module):
    def __init__(self, w0=1.0):
        super().__init__()
        self.w0 = w0

    def forward(self, x):
        return torch.sin(self.w0 * x)

def exists(val):
    return val is not None

class Siren(nn.Module):
    def __init__(self, dim_in, dim_out, w0=1.0, c=6.0, is_first=False, use_bias=True, activation=None, dropout=0.0):
        super().__init__()
        self.dim_in = dim_in
        self.is_first = is_first
        weight = torch.zeros(dim_out, dim_in)
        bias = torch.zeros(dim_out) if use_bias else None
        self.init_(weight, bias, c=c, w0=w0)
        self.weight = nn.Parameter(weight)
        self.bias = nn.Parameter(bias) if use_bias else None
        self.activation = Sine(w0) if activation is None else activation
        self.dropout = nn.Dropout(dropout)

    def init_(self, weight, bias, c, w0):
        dim = self.dim_in
        w_std = 1 / dim if self.is_first else math.sqrt(c / dim) / w0
        weight.uniform_(-w_std, w_std)
        if exists(bias):
            bias.uniform_(-w_std, w_std)

    def forward(self, x):
        out = F.linear(x, self.weight, self.bias)
        out = self.activation(out)
        out = self.dropout(out)
        return out

ARGS = {
    'dim_in': 4,
    'dim_out': 3,
    'w0': 1.0,
    'c': 6.0,
    'is_first': False,
    'use_bias': True,
    'activation': None,
    'dropout': 0.0
}
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 Siren — 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 Siren 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
Siren Ran
ran; our contract draft was wrong, not the code
mjhwright/error-correction/network.py
code served (permissive licence) · get_code("08022569ff28ebd6")
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 Sequential. 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 numpy as np

torch.manual_seed(0)

ARGS = {
    'hparams': {
        'depth': 3,
        'width': 8,
        'num_features': 2,
        'dim': 4
    }
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
Siren Ran
ran on a synthesized input · constructed an object; no result was computed
palombom/sirenmri/siren.py
code served (permissive licence) · get_code("4030b7a0e9748b71")
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, Siren: 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)

ARGS = {
    "dim_in": 4,
    "dim_hidden": 8,
    "dim_out": 2,
    "num_layers": 3,
    "w0": 30.0,
    "w0_initial": 30.0,
    "use_bias": True,
    "final_activation": 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 Siren — 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 Siren 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
Siren Ran
ran on a synthesized input · constructed an object; no result was computed
ecovision-uzh/climplicit/quickstart/siren.py
code served (permissive licence) · get_code("7b1fb4633b5400fd")
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, Siren: 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
import math

torch.manual_seed(0)

ARGS = {
    'dim_in': 4,
    'dim_out': 3,
    'w0': 1.0,
    'c': 6.0,
    'is_first': False,
    'use_bias': True,
    'activation': None,
    'dropout': False,
    'residual_connections': False,
    'h_siren': 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 Siren — 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 Siren 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
Siren Ran
ran on a synthesized input · constructed an object; no result was computed
arihaz/ImplicitNeuralRepresentations/network.py
pointer only (licence: NONE) · get_code("5a2ba67aad35b0e7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, arihaz/ImplicitNeuralRepresentations 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.

Siren Ran
ran on a synthesized input
scart97/Siren-fastai2/siren.py
code served (permissive licence) · get_code("5961b5f54fb8d3f8")
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 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
from torch import nn

class Siren(nn.Module):
    """
        Siren activation
        https://arxiv.org/abs/2006.09661
    """

    def __init__(self, w0=1):
        """
            w0 comes from the end of section 3
            it should be 30 for the first layer
            and 1 for the rest
        """
        super().__init__()
        self.w0 = torch.tensor(w0)

    def forward(self, x):
        return torch.sin(self.w0 * x)

    def extra_repr(self):
        return 'w0={}'.format(self.w0)


INIT_ARGS = {
    'w0': 30
}

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

if __name__ == '__main__':
    model = Siren(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
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 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Siren Ran
ran on a synthesized input
MaximeVandegar/Papers-in-100-Lines-of-Code/Implicit_Neural_Representations_with_Periodic_Activation_Functions/siren.py
code served (permissive licence) · get_code("5569fd5a29189c82")
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 2, dtype family float, shape (2, 3) 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 numpy as np

class SineLayer(nn.Module):

    def __init__(self, w0):
        super(SineLayer, self).__init__()
        self.w0 = w0

    def forward(self, x):
        return torch.sin(self.w0 * x)

class Siren(nn.Module):

    def __init__(self, w0=30, in_dim=2, hidden_dim=256, out_dim=1):
        super(Siren, self).__init__()
        self.net = nn.Sequential(nn.Linear(in_dim, hidden_dim), SineLayer(w0), nn.Linear(hidden_dim, hidden_dim), SineLayer(w0), nn.Linear(hidden_dim, hidden_dim), SineLayer(w0), nn.Linear(hidden_dim, hidden_dim), SineLayer(w0), nn.Linear(hidden_dim, out_dim))
        with torch.no_grad():
            self.net[0].weight.uniform_(-1.0 / in_dim, 1.0 / in_dim)
            self.net[2].weight.uniform_(-np.sqrt(6.0 / hidden_dim) / w0, np.sqrt(6.0 / hidden_dim) / w0)
            self.net[4].weight.uniform_(-np.sqrt(6.0 / hidden_dim) / w0, np.sqrt(6.0 / hidden_dim) / w0)
            self.net[6].weight.uniform_(-np.sqrt(6.0 / hidden_dim) / w0, np.sqrt(6.0 / hidden_dim) / w0)
            self.net[8].weight.uniform_(-np.sqrt(6.0 / hidden_dim) / w0, np.sqrt(6.0 / hidden_dim) / w0)

    def forward(self, x):
        return self.net(x)


INIT_ARGS = {
    'w0': 30,
    'in_dim': 4,
    'hidden_dim': 64,
    'out_dim': 3
}

ARGS = {
    'x': torch.randn(2, 4)
}
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 2, dtype family float, shape (2, 3) — shape (2, 3) 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)
SirenLayer Ran
ran on a synthesized input · constructed an object; no result was computed
palombom/sirenmri/siren.py
code served (permissive licence) · get_code("8ac1c1befb2f7857")
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, SirenLayer: 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
from math import sqrt

class Sine(nn.Module):
    """Sine activation with scaling.

    Args:
        w0 (float): Omega_0 parameter from SIREN paper.
    """

    def __init__(self, w0=1.0):
        super().__init__()
        self.w0 = w0

    def forward(self, x):
        return torch.sin(self.w0 * x)

class SirenLayer(nn.Module):
    """Implements a single SIREN layer.

    Args:
        dim_in (int): Dimension of input.
        dim_out (int): Dimension of output.
        w0 (float):
        c (float): c value from SIREN paper used for weight initialization.
        is_first (bool): Whether this is first layer of model.
        use_bias (bool):
        activation (torch.nn.Module): Activation function. If None, defaults to
            Sine activation.
    """

    def __init__(self, dim_in, dim_out, w0=30.0, c=6.0, is_first=False, use_bias=True, activation=None):
        super().__init__()
        self.dim_in = dim_in
        self.is_first = is_first
        self.linear = nn.Linear(dim_in, dim_out, bias=use_bias)
        w_std = 1 / dim_in if self.is_first else sqrt(c / dim_in) / w0
        nn.init.uniform_(self.linear.weight, -w_std, w_std)
        if use_bias:
            nn.init.uniform_(self.linear.bias, -w_std, w_std)
        self.activation = Sine(w0) if activation is None else activation

    def forward(self, x):
        out = self.linear(x)
        out = self.activation(out)
        return out


torch.manual_seed(0)

ARGS = {
    'dim_in': 4,
    'dim_out': 3,
    'w0': 30.0,
    'c': 6.0,
    'is_first': False,
    'use_bias': True,
    'activation': 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 SirenLayer — 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 SirenLayer 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
SirenNet Ran
ran on a synthesized input · constructed an object; no result was computed
marccoru/locationencoder/locationencoder/nn/siren.py
pointer only (licence: NONE) · get_code("da5b5f2242dabaf8")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, marccoru/locationencoder 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.

SirenNet Ran
ran on a synthesized input · constructed an object; no result was computed
ecovision-uzh/climplicit/quickstart/siren.py
code served (permissive licence) · get_code("116c403f9b64d6f9")
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, SirenNet: 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 = {
    "dim_in": 4,
    "dim_hidden": 8,
    "dim_out": 3,
    "num_layers": 3,
    "w0": 1.0,
    "w0_initial": 30.0,
    "use_bias": True,
    "final_activation": None,
    "dropout": False,
    "residual_connections": False,
    "h_siren": False,
    "return_hidden_embs": 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 SirenNet — 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 SirenNet 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
cast_tuple Ran
ran, honoured the contract we drafted
marccoru/locationencoder/locationencoder/nn/siren.py
pointer only (licence: NONE) · get_code("7bd1641f670e602d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, marccoru/locationencoder 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.

channel_last Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vsitzmann/siren/modules.py
code served (permissive licence) · get_code("ca25202da0ba3ba4")
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 4, dtype family float, shape (2, 4, 5, 3) 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)

# The function takes a tensor x and performs two transpose operations
# x.transpose(1, 2).transpose(2, 3)
# This means x needs to be a tensor with at least 4 dimensions (0, 1, 2, 3)
# We'll create a 4D tensor with distinct sizes for each dimension

x = torch.randn(2, 3, 4, 5)  # batch=2, dim1=3, dim2=4, dim3=5

ARGS = {
    "x": x
}
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 4, dtype family float, shape (2, 4, 5, 3) — shape (2, 4, 5, 3) 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
compl_conj Ran
ran; our contract draft was wrong, not the code
vsitzmann/siren/modules.py
code served (permissive licence) · get_code("50122beab0f810e1")
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)

# The function takes a tensor x, clones it, negates every other element
# along the last dimension (starting from index 1), and returns it.
# x should be a torch.Tensor with at least 1 dimension.

x = torch.randn(2, 4)

ARGS = {
    "x": x
}
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
conv3x3 Ran
ran; our contract draft was wrong, not the code
vsitzmann/siren/modules.py
code served (permissive licence) · get_code("46ad0f97079d1fd8")
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 PartialConv2d. 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)

# Define ARGS dict for conv3x3 function
ARGS = {
    'in_planes': 3,
    'out_planes': 4,
    'stride': 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 PartialConv2d — a PartialConv2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PartialConv2d 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
densely_sample_activations Ran
ran on a synthesized input
vsitzmann/siren/utils.py
code served (permissive licence) · get_code("d0cf6c9ce260f1a4")
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 (1, 100, 4) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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

torch.manual_seed(0)

class MockModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(1, 4)
    
    def forward_with_activations(self, input_dict):
        coords = input_dict['coords']
        batch_size = coords.shape[0]
        num_points = coords.shape[1]
        activations = torch.randn(batch_size, num_points, 4)
        return {'activations': activations}

model = MockModel()

ARGS = {
    'model': model,
    'num_dim': 1,
    'num_steps': 100
}
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 (1, 100, 4) — shape (1, 100, 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
divergence Ran
ran; our contract draft was wrong, not the code
vsitzmann/siren/diff_operators.py
code served (permissive licence) · get_code("87e822f4bcfbdac7")
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, 1) 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)

# The divergence function computes the divergence of a vector field y with respect to x
# It requires:
# - y: a tensor with shape [..., n] where n is the output dimension
# - x: a tensor that y depends on, requires grad

# Create a small input tensor x with requires_grad=True
x = torch.randn(2, 4, requires_grad=True)

# Create y as a function of x
# y should have shape [..., output_dim]
# We'll create y with output dimension 3 (distinct from input dim 4)
# y = some function of x, e.g., a simple linear-like transformation
y = torch.sum(x, dim=1, keepdim=True) * torch.ones(2, 3)
y.requires_grad_(True)

# Alternatively, if y needs to be explicitly computed from x:
# Let's make y depend on x through autograd graph
y = x[:, :3] * 2.0 + x[:, 1:4].sum(dim=1, keepdim=True)

ARGS = {
    "y": y,
    "x": x,
}
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, 1) — shape (2, 1) 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
get_mgrid Ran
ran, honoured the contract we drafted
vsitzmann/siren/dataio.py
code served (permissive licence) · get_code("c3c376e19f06a842")
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 (256, 2) 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)

ARGS = {
    "sidelen": 16,
    "dim": 2,
}
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 (256, 2) — shape (256, 2) 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
get_subdict Ran
ran; our contract draft was wrong, not the code
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("cc9c74f5df741aef")
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 OrderedDict. 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.

from collections import OrderedDict
import re

# Test case 1: None dictionary
test_dict_1 = None

# Test case 2: Empty key
test_dict_2 = OrderedDict([
    ('layer.weight', 0.5),
    ('layer.bias', 0.1),
    ('model.fc.weight', 0.3),
])

# Test case 3: Dictionary with nested structure
test_dict_3 = OrderedDict([
    ('encoder.layer1.weight', [1, 2, 3]),
    ('encoder.layer1.bias', [4, 5, 6]),
    ('encoder.layer2.weight', [7, 8, 9]),
    ('decoder.output', [10, 11]),
])

# Test case 4: Key that matches some entries
test_dict_4 = OrderedDict([
    ('model.encoder.weight', 'w1'),
    ('model.encoder.bias', 'b1'),
    ('model.decoder.weight', 'w2'),
    ('other.param', 'x'),
])

# Test case 5: Simple dictionary with various keys
test_dict_5 = OrderedDict([
    ('prefix.a', 1),
    ('prefix.b', 2),
    ('prefix.c.d', 3),
    ('other', 4),
])

ARGS = {
    'dictionary': test_dict_5,
    'key': 'prefix',
}
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 OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict 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
grads2img Ran
ran; our contract draft was wrong, not the code
vsitzmann/siren/dataio.py
code served (permissive licence) · get_code("2d7e4df5345f8188")
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 (3, 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
import numpy as np
import math

torch.manual_seed(0)

# The function expects a gradient tensor
# Looking at the operations:
# - squeeze(0): removes batch dimension
# - permute(-2, -1, -3): reorders dimensions
# - The result should have shape (nRows, nCols, 2) after permute
# So we need to construct a tensor that after squeeze(0).permute(-2, -1, -3)
# gives us (nRows, nCols, 2)
# 
# If we start with shape (1, 2, nRows, nCols):
# - squeeze(0) -> (2, nRows, nCols)
# - permute(-2, -1, -3) -> permute(1, 2, 0) -> (nRows, nCols, 2)

gradients = torch.randn(1, 2, 4, 5)

ARGS = {
    "gradients": gradients
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, 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, shape (3, 4, 5) — shape (3, 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
hessian Ran
ran; our contract draft was wrong, not the code
vsitzmann/siren/diff_operators.py
code served (permissive licence) · get_code("fd089b9f4b8749ab")
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. 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
torch.manual_seed(0)

# The function expects:
# y: shape (meta_batch_size, num_observations, channels)
# x: shape (meta_batch_size, num_observations, 2)
# Both need to require grad for autograd to work

meta_batch_size = 2
num_observations = 3
channels = 4
x_dim = 2

# x needs to be differentiable
x = torch.randn(meta_batch_size, num_observations, x_dim, requires_grad=True, dtype=torch.float32)

# y should be a function of x to compute meaningful gradients
# Create a simple function: y = x^2 summed in a way that gives us channels outputs
y = torch.sum(x.unsqueeze(-1) ** 2, dim=2)  # shape (meta_batch_size, num_observations, 1)
y = y.expand(meta_batch_size, num_observations, channels)  # expand to (meta_batch_size, num_observations, channels)
y.requires_grad_(True)

ARGS = {
    "y": y,
    "x": x,
}
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 tuple — a tuple output has no array shape of its own; none is 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
image_l1 Ran
ran on a synthesized input
vsitzmann/siren/loss_functions.py
code served (permissive licence) · get_code("81fe748ce57e9fae")
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 dict. 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

torch.manual_seed(0)

# Construct tensors
batch_size = 2
height = 4
width = 3
channels = 4

# model_output dict with 'model_out' key
model_output = {
    'model_out': torch.randn(batch_size, channels, height, width)
}

# gt dict with 'img' key
gt = {
    'img': torch.randn(batch_size, channels, height, width)
}

# mask tensor - same shape as model_out and gt['img']
mask = torch.randn(batch_size, channels, height, width)

ARGS = {
    'mask': mask,
    'model_output': model_output,
    'gt': gt
}
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 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
image_mse Ran
ran on a synthesized input
vsitzmann/siren/loss_functions.py
code served (permissive licence) · get_code("196981f676d3a9d8")
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 dict. 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)

# The function expects:
# - mask: either None or a tensor that broadcasts with model_output and gt
# - model_output: a dict with 'model_out' key containing a tensor
# - gt: a dict with 'img' key containing a tensor

# Create small tensors with distinct dimensions for clarity
batch_size = 2
img_height = 4
img_width = 3
channels = 5

# Create model_output dict with 'model_out' tensor
model_output = {
    'model_out': torch.randn(batch_size, channels, img_height, img_width)
}

# Create gt dict with 'img' tensor (same shape as model_out['model_out'])
gt = {
    'img': torch.randn(batch_size, channels, img_height, img_width)
}

# Create mask tensor (same shape as the image tensors)
mask = torch.ones(batch_size, channels, img_height, img_width)

ARGS = {
    'mask': mask,
    'model_output': model_output,
    'gt': gt
}
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 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
laplace Ran
ran; our contract draft was wrong, not the code
vsitzmann/siren/diff_operators.py
code served (permissive licence) · get_code("a2be4be6558869af")
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, 1) 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 y that depends on x via autograd graph
# y needs to have a last dimension so we can iterate over it in divergence()
# Shape: (batch=2, feature_dim=4)
x = torch.randn(2, 4, dtype=torch.float32, requires_grad=True)

# Create y as a function of x so the autograd graph is built
# y should also have shape compatible with the divergence function
# We'll make y = x^2 to create a simple dependency
y = (x ** 2).sum(dim=1, keepdim=True)  # Shape: (2, 1)

ARGS = {
    "y": y,
    "x": x,
}
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, 1) — shape (2, 1) 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
lin2img Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
vsitzmann/siren/dataio.py
pointer only (licence: MIT); the stored copy get_code serves, from computational-imaging/bacon, is pointer only (licence: NONE) · get_code("9d902c538061fea5")
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 computational-imaging/bacon, 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.

non_zero_mean Ran
ran; our contract draft was wrong, not the code
linusnie/diffcd/diffcd/networks.py
code served (permissive licence) · get_code("a543cc7ec13c4381")
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 ArrayImpl, 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 jax
import jax.numpy as jnp
import jax.random as jrnd

# Create a random key for JAX
key = jrnd.PRNGKey(0)

# Define shape - using distinct dimensions (2, 4) for batch and feature dims
shape = (2, 4)

# Define dtype
dtype = jnp.float32

# Construct the ARGS dictionary
ARGS = {
    "key": key,
    "shape": shape,
    "dtype": dtype
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, 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
safe_softplus Ran
ran, honoured the contract we drafted
linusnie/diffcd/diffcd/networks.py
code served (permissive licence) · get_code("74a2fc2b7bb9535e")
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 ArrayImpl, 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 jax.numpy as jnp
import numpy as np

# Create test inputs for safe_softplus
x_np = np.array([[0.5, 1.0, -0.5, 2.0],
                  [0.1, -1.0, 0.3, 1.5]], dtype=np.float32)
x_jax = jnp.asarray(x_np)

beta_value = 100

ARGS = {
    "x": x_jax,
    "beta": beta_value
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, 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
siren_uniform_ Ran
ran; our contract draft was wrong, not the code
dalmia/siren/siren/siren.py
code served (permissive licence) · get_code("a87fd526d4a3314f")
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 (3, 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
import math
from torch.nn.init import _calculate_correct_fan

torch.manual_seed(0)

# Create a small tensor for initialization
tensor = torch.empty(3, 5)

ARGS = {
    'tensor': tensor,
    'mode': 'fan_in',
    'c': 6
}
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 (3, 5) — shape (3, 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
softplus Ran
ran, honoured the contract we drafted
linusnie/diffcd/diffcd/networks.py
code served (permissive licence) · get_code("412ae5f90dd4d2df")
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 ArrayImpl, 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 jax.numpy as jnp
import numpy as np

# The function softplus takes:
# - x: a numeric array/tensor that will be used in jnp operations
# - beta: a numeric scalar (default 100)

# Create a small JAX array for x
x_value = jnp.array([[-2.0, -1.0, 0.0, 1.0],
                      [2.0, 3.0, -0.5, 0.5]])

# Define ARGS with both parameters
ARGS = {
    "x": x_value,
    "beta": 100
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, 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
zero_mean Ran
ran; our contract draft was wrong, not the code
linusnie/diffcd/diffcd/networks.py
code served (permissive licence) · get_code("bd03e58827e3f72d")
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 ArrayImpl, ndim 2, dtype family float, shape (4, 3) 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 jax
import jax.numpy as jnp
import jax.random as jrnd

# Create a random key for JAX
key = jrnd.PRNGKey(0)

# Define shape - using distinct dimensions (4, 3) for probing
shape = (4, 3)

# Define dtype
dtype = jnp.float32

# Construct ARGS dict with all parameters
ARGS = {
    "key": key,
    "shape": shape,
    "dtype": dtype,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 2, dtype family float, shape (4, 3) — shape (4, 3) 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
ActivationFunction Not yet run
unverified (harvested, no recorded run)
linusnie/diffcd/diffcd/networks.py
code served (permissive licence) · get_code("c156c41aef66579a")
How it ran

no invocation recorded.

GridSampleWeightsBase Not yet run
unverified (harvested, no recorded run)
TheStageAI/TorchIntegral/torch_integral/parametrizations/interpolation_weights.py
code served (permissive licence) · get_code("b273c9f1b2897ba2")
How it ran

no invocation recorded.

IWeights Not yet run
unverified (harvested, no recorded run)
TheStageAI/TorchIntegral/torch_integral/parametrizations/interpolation_weights.py
code served (permissive licence) · get_code("7d35faf2895d98fb")
How it ran

no invocation recorded.

MLP Not yet run
unverified (harvested, no recorded run)
linusnie/diffcd/diffcd/networks.py
code served (permissive licence) · get_code("d09b082f03330456")
How it ran

no invocation recorded.

MetaModule Not yet run
unverified (harvested, no recorded run)
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("78371e32f4b5df03")
How it ran

no invocation recorded.

MetaSequential Not yet run
unverified (harvested, no recorded run)
TalFurman/Implict_neural_representation_of_images/modules.py
code served (permissive licence) · get_code("fa9f394eb26ff78f")
How it ran

no invocation recorded.

Showing the first 60. An agent gets every one from get_harvested_code_for_paper("2006.09661").

Also present in other papers' code

Of 71 samples lifted from this paper's linked repositories, 71 carry a function identity and 0 do not. Of the 69 distinct functions with an identity, 23 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 23 later) and 46 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
Sine
brick of this paper's assembly: helper
8 other papers
8 later · 4 only as an assembly brick
2104.03866 (2021-04)
SMD-Nets: Stereo Mixture Density Networks
2305.19185 (2023-05)
Compression with Bayesian Implicit Neural Representations
2310.19629 (2023-10)
RayDF: Neural Ray-surface Distance Fields with Multi-view Consistency
and 5 more not named here
cast_tuple
brick of this paper's assembly: helper
8 other papers
8 later · 1 only as an assembly brick
2111.00195 (2021-11)
Learning Continuous Representation of Audio for Arbitrary Scale Super Resolution
2210.06823 (2022-10)
Scalable Neural Video Representations with Learnable Positional Features
2211.03329 (2022-11)
Implicit Graphon Neural Representation
and 5 more not named here
laplace 6 other papers
6 later
2007.09601 (2020-07)
Hypersolvers: Toward Fast Continuous-Depth Models
2108.08478 (2021-08)
Learning Anchored Unsigned Distance Functions with Gradient Direction Alignment for Single-view Garment Reconstruction
2303.14092 (2023-03)
NeuFace: Realistic 3D Neural Face Rendering from Multi-view Images
and 3 more not named here
divergence 4 other papers
4 later
2108.08478 (2021-08)
Learning Anchored Unsigned Distance Functions with Gradient Direction Alignment for Single-view Garment Reconstruction
2303.14092 (2023-03)
NeuFace: Realistic 3D Neural Face Rendering from Multi-view Images
2312.04028 (2023-12)
ImFace++: A Sophisticated Nonlinear 3D Morphable Face Model with Implicit Neural Representations
and 1 more not named here
get_mgrid 3 other papers
3 later
2105.02788 (2021-05)
ACORN: Adaptive Coordinate Networks for Neural Scene Representation
2202.10847 (2022-02)
UncertaINR: Uncertainty Quantification of End-to-End Implicit Neural Representations for Computed Tomography
2409.19472 (2024-09)
Towards Croppable Implicit Neural Representations
get_subdict
brick of this paper's assembly: helper
3 other papers
3 later
2303.14092 (2023-03)
NeuFace: Realistic 3D Neural Face Rendering from Multi-view Images
2310.05174 (2023-10)
GSLB: The Graph Structure Learning Benchmark
2312.04028 (2023-12)
ImFace++: A Sophisticated Nonlinear 3D Morphable Face Model with Implicit Neural Representations
hessian 3 other papers
3 later
2108.08478 (2021-08)
Learning Anchored Unsigned Distance Functions with Gradient Direction Alignment for Single-view Garment Reconstruction
2303.14092 (2023-03)
NeuFace: Realistic 3D Neural Face Rendering from Multi-view Images
2312.04028 (2023-12)
ImFace++: A Sophisticated Nonlinear 3D Morphable Face Model with Implicit Neural Representations
lin2img 3 other papers
3 later
2105.02788 (2021-05)
ACORN: Adaptive Coordinate Networks for Neural Scene Representation
2112.04645 (2021-12)
BACON: Band-limited Coordinate Networks for Multiscale Scene Representation
2202.10847 (2022-02)
UncertaINR: Uncertainty Quantification of End-to-End Implicit Neural Representations for Computed Tomography
Sine
brick of this paper's assembly: entry, model
2 other papers
2 later · 1 only as an assembly brick
2207.03691 (2022-07)
Neural Implicit Dictionary Learning via Mixture-of-Expert Training
2409.19472 (2024-09)
Towards Croppable Implicit Neural Representations
SineLayer
brick of this paper's assembly: helper
2 other papers
2 later
2107.04589 (2021-07)
ViTGAN: Training GANs with Vision Transformers
2206.04647 (2022-06)
VideoINR: Learning Video Implicit Neural Representation for Continuous Space-Time Super-Resolution
channel_last 2 other papers
2 later
2310.09192 (2023-10)
Does Graph Distillation See Like Vision Dataset Counterpart?
2503.01287 (2025-03)
Robust Simulation-Based Inference under Missing Data via Neural Processes
first_layer_sine_init
brick of this paper's assembly: helper
2 other papers
2 later · 2 only as an assembly brick
2203.15065 (2022-03)
DeepShadow: Neural Shape from Shadow
2207.03691 (2022-07)
Neural Implicit Dictionary Learning via Mixture-of-Expert Training
image_mse 2 other papers
2 later
2202.10847 (2022-02)
UncertaINR: Uncertainty Quantification of End-to-End Implicit Neural Representations for Computed Tomography
2210.06823 (2022-10)
Scalable Neural Video Representations with Learnable Positional Features
sine_init
brick of this paper's assembly: helper
2 other papers
2 later · 2 only as an assembly brick
2203.15065 (2022-03)
DeepShadow: Neural Shape from Shadow
2207.03691 (2022-07)
Neural Implicit Dictionary Learning via Mixture-of-Expert Training
Siren
brick of this paper's assembly: helper
1 other paper
1 later
2606.08046 (2026-06)
OSMGraphCLIP: Learning Global Location Representations from OpenStreetMap Graphs
Siren
brick of this paper's assembly: entry
1 other paper
1 later
2503.04835 (2025-03)
Distilling Dataset into Neural Field
Siren
brick of this paper's assembly: model
1 other paper
1 later
2504.05089 (2025-04)
Climplicit: Climatic Implicit Embeddings for Global Ecological Tasks
SirenLayer
brick of this paper's assembly: model
1 other paper
1 later
2503.04835 (2025-03)
Distilling Dataset into Neural Field
SirenNet
brick of this paper's assembly: entry
1 other paper
1 later
2504.05089 (2025-04)
Climplicit: Climatic Implicit Embeddings for Global Ecological Tasks
compl_conj 1 other paper
1 later
2310.09192 (2023-10)
Does Graph Distillation See Like Vision Dataset Counterpart?

Shown: the 20 functions present in the most other papers. 3 more shared functions are not listed here.

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

Implicitly defined, continuous, differentiable signal representations parameterized by neural networks have emerged as a powerful paradigm, offering many possible benefits over conventional representations. However, current network architectures for such implicit neural representations are incapable of modeling signals with fine detail, and fail to represent a signal's spatial and temporal derivatives, despite the fact that these are essential to many physical signals defined implicitly as the solution to partial differential equations. We propose to leverage periodic activation functions for implicit neural representations and demonstrate that these networks, dubbed sinusoidal representation networks or SIRENs, are ideally suited for representing complex natural signals and their derivatives. We analyze SIREN activation statistics to propose a principled initialization scheme and demonstrate the representation of images, wavefields, video, sound, and their derivatives. Further, we show how SIRENs can be leveraged to solve challenging boundary value problems, such as particular Eikonal equations (yielding signed distance functions), the Poisson equation, and the Helmholtz and wave equations. Lastly, we combine SIRENs with hypernetworks to learn priors over the space of SIREN functions. Please see the project website for a video overview of the proposed method and all applications.

For agents

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

get_harvested_code_for_paper("2006.09661")
get_code_for_paper("2006.09661")
have("2006.09661")

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

curl https://syntology.ai/api/ran/2006.09661.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.09661.svg)](https://syntology.ai/paper/2006.09661)

Connect an agent — have() is free.