SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1609.04468 · 2016

Sampling Generative Networks

arXiv · PDF · Open in the Atlas

Code that ran

15 of 15 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 14 have no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-21; 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/1609.04468.json · badge: /api/ran/1609.04468.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
dribnet/plat canonical (classifier label)
classifier basis: census
14 ran · 0 no recorded run
gitaar9/MLDAGAN reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
add_shoulders Ran
ran on a synthesized input
dribnet/plat/plat/sampling.py
code served (permissive licence) · get_code("3fa640a95cb4a3ae")
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
import numpy as np

torch.manual_seed(0)

# Create small dummy images (just numpy arrays or tensors)
# The function only uses len() and indexing on these, so simple arrays work
image1 = np.random.rand(3, 32, 32)
image2 = np.random.rand(3, 32, 32)
image3 = np.random.rand(3, 32, 32)
image4 = np.random.rand(3, 32, 32)
image5 = np.random.rand(3, 32, 32)
image6 = np.random.rand(3, 32, 32)

images = [image1, image2, image3, image4, image5, image6]

# Create anchor images (typically corner/edge images)
anchor1 = np.random.rand(3, 32, 32)
anchor2 = np.random.rand(3, 32, 32)
anchor3 = np.random.rand(3, 32, 32)

anchor_images = [anchor1, anchor2, anchor3]

# Grid dimensions
rows = 3
cols = 2

ARGS = {
    "images": images,
    "anchor_images": anchor_images,
    "rows": rows,
    "cols": cols,
}
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 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
additive_composite Ran
ran on a synthesized input
dribnet/plat/plat/bin/canvas.py
code served (permissive licence) · get_code("4695bd5cc5fe570c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family float, shape (4, 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 numpy as np

# Examine the function to understand what it needs:
# - src: accessed with [rgb] and [alpha] indexing, needs shape (4, height, width) with channels
# - src_mask: either None or accessed with [alpha] indexing, can be None or array
# - dst: accessed with [alpha] and [rgb] indexing, needs shape (4, height, width)

# Create small numpy arrays with distinct dimensions
height = 3
width = 5

# src: 4 channels (RGBA), height=3, width=5
src = np.random.rand(4, height, width).astype('float32')

# src_mask: either None or a mask array for alpha channel
src_mask = np.random.rand(height, width).astype('float32')

# dst: 4 channels (RGBA), height=3, width=5
dst = np.random.rand(4, height, width).astype('float32')

ARGS = {
    'src': src,
    'src_mask': src_mask,
    'dst': dst,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (4, 3, 5) — shape (4, 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
alpha_composite Ran
ran on a synthesized input
dribnet/plat/plat/bin/canvas.py
code served (permissive licence) · get_code("c001793a18ea82fe")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family float, shape (4, 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 numpy as np

# Create small RGBA images with shape (4, height, width)
# src: source image with alpha channel
src = np.random.rand(4, 2, 3).astype('float32')

# src_mask: alpha mask for source (can be None or a mask array)
src_mask = np.random.rand(2, 3).astype('float32')

# dst: destination image with alpha channel
dst = np.random.rand(4, 2, 3).astype('float32')

ARGS = {
    'src': src,
    'src_mask': src_mask,
    'dst': dst
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (4, 2, 3) — shape (4, 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: none by the harness; any seed is in the setup block
filter_attributes Ran
ran on a synthesized input
dribnet/plat/plat/bin/atvec.py
code served (permissive licence) · get_code("e87e50f4ec870631")
How it ran

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

import torch
import numpy as np

torch.manual_seed(0)

# Create a small 3D tensor: (batch=2, sequence=4, attributes=5)
attribs = torch.randn(2, 4, 5)

# Test case 1: which_attribs is None (returns original)
ARGS = {
    'attribs': attribs,
    'which_attribs': None
}

# Alternative test case 2: which_attribs filters to specific indices
# ARGS = {
#     'attribs': attribs,
#     'which_attribs': '0,2,4'
# }
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 Tensor, ndim 3, dtype family float, shape (2, 4, 5) — shape (2, 4, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_averages Ran
ran on a synthesized input
dribnet/plat/plat/bin/atvec.py
code served (permissive licence) · get_code("69c23b3721421350")
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 numpy as np
import torch

# Set seed for reproducibility
torch.manual_seed(0)
np.random.seed(0)

# Create sample data
num_items = 2
num_attribs = 3
encoded_dim = 4

# attribs should be shape (num_items, 1, num_attribs)
# with values of 0 or 1 indicating presence/absence of attributes
attribs = np.random.randint(0, 2, size=(num_items, 1, num_attribs)).astype(np.float32)

# encoded should be shape (num_items, encoded_dim)
# contains the encoded representations for each item
encoded = np.random.randn(num_items, encoded_dim).astype(np.float32)

ARGS = {
    'attribs': attribs,
    'encoded': encoded
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, tqdm — 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
get_class_averages Ran
ran on a synthesized input
dribnet/plat/plat/bin/atvec.py
code served (permissive licence) · get_code("88d5514f4345aafe")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, 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 numpy as np

# Create small stand-in data
num_classes = 3
num_samples = 25

# attribs: list of arrays where attribs[i][0] is the class index
attribs = [np.array([i % num_classes]) for i in range(num_samples)]

# encoded: list of feature vectors (or arrays)
# Using shape (num_samples, 4) where 4 is the feature dimension
encoded = [np.random.randn(4) for i in range(num_samples)]

ARGS = {
    "attribs": attribs,
    "encoded": encoded,
    "num_classes": num_classes,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_json_vectors Ran
ran on a synthesized input
dribnet/plat/plat/utils.py
code served (permissive licence) · get_code("c04dd2fcc1e23ad9")
How it ran

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

import numpy as np
import json
import tempfile
import os

# Create a temporary JSON file with vector data
temp_dir = tempfile.mkdtemp()
filename = os.path.join(temp_dir, "vectors.json")

# Write sample vector data to JSON file
vectors_data = [
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0],
    [9.0, 10.0, 11.0, 12.0]
]

with open(filename, 'w') as f:
    json.dump(vectors_data, f)

ARGS = {
    "filename": filename
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 4) — shape (3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
grid2img Ran
ran on a synthesized input
dribnet/plat/plat/grid_layout.py
code served (permissive licence) · get_code("874d0c68ae78d3bb")
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 Image. 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 numpy as np
from PIL import Image

# Create a small grid of images
# Each image has shape (channels, height, width)
arr = [
    np.random.rand(3, 4, 5).astype(np.float32),  # 3 channels, 4 height, 5 width
    np.random.rand(3, 4, 5).astype(np.float32),
    np.random.rand(3, 4, 5).astype(np.float32),
    np.random.rand(3, 4, 5).astype(np.float32),
]

ARGS = {
    "arr": arr,
    "rows": 2,
    "cols": 2,
    "with_space": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
lerp Ran
ran on a synthesized input
dribnet/plat/plat/interpolate.py
code served (permissive licence) · get_code("ee023c21ffe38ef3")
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 float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

ARGS = {
    "val": 0.5,
    "low": 1.0,
    "high": 10.0,
}
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 float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
lerp_gaussian Ran
ran on a synthesized input
dribnet/plat/plat/interpolate.py
code served (permissive licence) · get_code("2491fa1841f23907")
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 float64, ndim 0, 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 scipy.stats import norm

torch.manual_seed(0)

# Analyze the function signature:
# lerp_gaussian(val, low, high)
# The function takes three numeric parameters and performs:
# - norm.cdf(low)
# - norm.cdf(high)
# - lerp(val, low_gau, high_gau) which does: low_gau + (high_gau - low_gau) * val
# - norm.ppf(lerped_gau)

# All parameters are used in arithmetic operations with scalar/array values
# They should be numeric types (float, int, or numpy/torch arrays)

ARGS = {
    "val": 0.5,           # scalar between 0 and 1 for interpolation
    "low": -2.0,          # scalar for lower bound
    "high": 2.0,          # scalar for upper bound
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
load_model_with_interface Ran
ran on a synthesized input
dribnet/plat/plat/zoo.py
code served (permissive licence) · get_code("b6f6cf45195382a5")
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 MockModelClass. 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 os
import sys
import importlib
import tempfile
from unittest.mock import Mock, patch, MagicMock

# Create a mock model class that will be importable
class MockModelClass:
    def __init__(self, filename):
        self.filename = filename
        self.loaded = True

# Create a temporary module to simulate the model interface
mock_module = MagicMock()
mock_module.MockModel = MockModelClass

# Patch importlib.import_module to return our mock module
original_import_module = importlib.import_module

def patched_import_module(name):
    if name == 'my_models.interface':
        return mock_module
    return original_import_module(name)

# Create a temporary model file path (won't actually be created, just a path string)
model_file_path = '/tmp/mock_model.pt'

# Define the arguments
ARGS = {
    'model_file_name': model_file_path,
    'model_interface': 'my_models.interface.MockModel'
}

# Monkey-patch importlib to use our mock
importlib.import_module = patched_import_module
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 MockModelClass — a MockModelClass output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockModelClass 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
remove_duplicates Ran
ran; our contract draft was wrong, not the code
gitaar9/MLDAGAN/dagan_architectures.py
code served (permissive licence) · get_code("e73857dab63b5298")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned list, ndim 1, 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 torch.nn as nn

torch.manual_seed(0)

class FeatureLayer:
    def __init__(self, name, tensor):
        self.name = name
        self.tensor = tensor

# Create a list of feature layers with some duplicates
input_features = [
    FeatureLayer("layer1", torch.randn(2, 4)),
    FeatureLayer("layer2", torch.randn(2, 3)),
    FeatureLayer("layer1", torch.randn(2, 4)),  # duplicate name
    FeatureLayer("layer3", torch.randn(2, 5)),
    FeatureLayer("layer2", torch.randn(2, 3)),  # duplicate name
    FeatureLayer("layer4", torch.randn(2, 6)),
]

ARGS = {
    "input_features": input_features
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
resolve_model_to_filename Ran
ran on a synthesized input
dribnet/plat/plat/zoo.py
code served (permissive licence) · get_code("aee9dcd8433bb6d9")
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 str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
import shutil

# Create a temporary directory to act as the platzoo directory
temp_dir = tempfile.mkdtemp()
plat_dir = os.path.join(temp_dir, 'plat')
os.makedirs(plat_dir, exist_ok=True)
zoo_dir = os.path.join(plat_dir, 'zoo')
os.makedirs(zoo_dir, exist_ok=True)

# Set environment variable for the test
os.environ['PLAT_DIR'] = plat_dir

# Define arguments for resolve_model_to_filename
ARGS = {
    'filename': 'model.pkl'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
resolve_model_type_from_filename Ran
ran on a synthesized input
dribnet/plat/plat/zoo.py
code served (permissive licence) · get_code("4ee8364367586c42")
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 str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function takes a filename string and splits it by '.' to get the extension
# It checks if the last part is 'pth' and returns a model type string
# No complex objects needed - just a string parameter

ARGS = {
    'filename': 'model.pth'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
slerp Ran
ran on a synthesized input
dribnet/plat/plat/interpolate.py
code served (permissive licence) · get_code("9c1b90f679a7f761")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (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 numpy as np

# Construct small vector arguments for slerp
# The function expects val (scalar), low (vector), and high (vector)

val = 0.5

# Create two random unit vectors for low and high
# These should be normalized so arccos can work properly
np.random.seed(0)
low = np.random.randn(4)
low = low / np.linalg.norm(low)

high = np.random.randn(4)
high = high / np.linalg.norm(high)

ARGS = {
    "val": val,
    "low": low,
    "high": high
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (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

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 4 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 4 later) and 11 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
lerp 1 other paper
1 later
1711.04340 (2017-11)
Data Augmentation Generative Adversarial Networks
lerp_gaussian 1 other paper
1 later
1711.04340 (2017-11)
Data Augmentation Generative Adversarial Networks
remove_duplicates 1 other paper
1 later
1711.04340 (2017-11)
Data Augmentation Generative Adversarial Networks
slerp 1 other paper
1 later
1711.04340 (2017-11)
Data Augmentation Generative Adversarial Networks

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 2026-11-21.

Repositories linked to this paper

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

Abstract

We introduce several techniques for sampling and visualizing the latent spaces of generative models. Replacing linear interpolation with spherical linear interpolation prevents diverging from a model's prior distribution and produces sharper samples. J-Diagrams and MINE grids are introduced as visualizations of manifolds created by analogies and nearest neighbors. We demonstrate two new techniques for deriving attribute vectors: bias-corrected vectors with data replication and synthetic vectors with data augmentation. Binary classification using attribute vectors is presented as a technique supporting quantitative analysis of the latent space. Most techniques are intended to be independent of model type and examples are shown on both Variational Autoencoders and Generative Adversarial Networks.

For agents

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

get_harvested_code_for_paper("1609.04468")
get_code_for_paper("1609.04468")
have("1609.04468")

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

curl https://syntology.ai/api/ran/1609.04468.json

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

[![Syntology run record](https://syntology.ai/api/ran/1609.04468.svg)](https://syntology.ai/paper/1609.04468)

Connect an agent — have() is free.