SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1909.13082 · 2019

Wasserstein-2 Generative Networks

arXiv · PDF · Open in the Atlas

Code that ran

10 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 9 have no contract verdict; 1 failure of our instrument; 5 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/1909.13082.json · badge: /api/ran/1909.13082.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
lit-leo/inner-convex-nn community (classifier label)
classifier basis not recorded on the link
10 ran · 4 no recorded run
copy not recorded community
not in the paper's declared repositories
0 ran · 1 no recorded run
FunctionStatusWhere it lives
analyzeTask Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/RL/analyze.random-search.py
code served (permissive licence) · get_code("092f4c3b216fbb95")
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 os
import tempfile
import numpy as np
import pandas as pd

# Create a temporary directory structure with the expected files
tmpdir = tempfile.mkdtemp()

# Create directory structure for each algorithm
for alg in ['DDPG', 'NAF', 'ICNN']:
    alg_dir = os.path.join(tmpdir, alg)
    os.makedirs(alg_dir, exist_ok=True)
    
    # Create a couple of experiment directories
    for exp_idx in range(2):
        exp_dir = os.path.join(alg_dir, f'exp_{exp_idx}')
        os.makedirs(exp_dir, exist_ok=True)
        
        # Create a test.log file with dummy data
        # Format: 2 columns, first column is iteration/step, second is reward
        test_data = np.column_stack([
            np.arange(50),  # 50 steps
            np.random.randn(50) + np.linspace(0, 10, 50)  # Rewards with upward trend
        ])
        np.savetxt(os.path.join(exp_dir, 'test.log'), test_data)

ARGS = {
    'taskDir': tmpdir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — 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
get_data Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/RL/plot-all.py
code served (permissive licence) · get_code("e8f061b46629b5d1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
import numpy as np

# Create temporary directory structure with log files
tmpdir = tempfile.mkdtemp()

# Create subdirectories with log files
for i in range(3):
    subdir = os.path.join(tmpdir, f'exp_{i}')
    os.makedirs(subdir, exist_ok=True)
    
    # Create synthetic data: x values from 0 to 10, y values with some variation
    x_data = np.linspace(0, 10, 50)
    y_data = np.sin(x_data) + np.random.randn(50) * 0.1
    
    # Stack x and y as columns and save
    data = np.column_stack([x_data, y_data])
    log_path = os.path.join(subdir, 'log.txt')
    np.savetxt(log_path, data)

ARGS = {
    'dirpath': tmpdir
}
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 2, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_step Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/lib/bundle_entropy.py
code served (permissive licence) · get_code("355190b9807011bc")
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 numpy as np

# Construct arguments for get_step function
# The function takes v and dv as numpy arrays
# It checks if any element in dv is negative
# If so, it computes a = -v / dv and returns the minimum of a where dv < 0
# Otherwise returns 1.0

# Create test arrays
v = np.array([2.0, 4.0, 6.0, 8.0])
dv = np.array([-1.0, 2.0, -0.5, 3.0])

ARGS = {
    "v": v,
    "dv": dv,
}
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 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
logexp1p Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/lib/bundle_entropy_dual.py
code served (permissive licence) · get_code("0a767f9ffa609a3e")
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 (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 numpy as np

# The function logexp1p takes a single parameter x
# which should be a numpy array (based on the use of np.zeros_like, indexing with boolean masks, etc.)

# Create a small test array with mixed values (some > 1, some <= 1)
x = np.array([[-2.0, 0.5, 1.5, 3.0],
              [-0.5, 2.0, 0.1, 5.0]])

ARGS = {
    "x": x
}
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 (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
logistic Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/lib/bundle_entropy_dual.py
code served (permissive licence) · get_code("99b488c1fe141b25")
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

def logistic(x):
    return 1.0 / (1.0 + np.exp(-x))

# Create a small numpy array to pass to logistic
x_input = np.array([0.0, 1.0, -1.0, 2.5])

ARGS = {
    "x": x_input
}
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
loss_function Ran
ran; our contract draft was wrong, not the code
lit-leo/inner-convex-nn/vae/vanila_vae.py
pointer only (licence: MIT); the stored copy get_code serves, from bhpfelix/Variational-Autoencoder-PyTorch, is pointer only (licence: NONE) · get_code("151aee3ffaced99a")
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 bhpfelix/Variational-Autoencoder-PyTorch, 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.

pdipm_boyd Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/lib/bundle_entropy.py
code served (permissive licence) · get_code("0bac3bc140b305fd")
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 scipy as sp

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

# Create a small valid optimization problem
# G should be nDual x nPrimal
nDual = 4
nPrimal = 3

# Create a small G matrix
G = np.random.randn(nDual, nPrimal) * 0.1

# Create h vector (nDual,)
h = -np.abs(np.random.randn(nDual)) * 0.5

# Create z0 (initial dual variable, nDual,)
z0 = np.ones(nDual) / nDual

ARGS = {
    "G": G,
    "h": h,
    "z0": z0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
pdipm_pc Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/lib/bundle_entropy.py
code served (permissive licence) · get_code("453775314c379efe")
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 scipy as sp

# Analyze pdipm_pc function signature and usage:
# - G: used as G.shape, G.T.dot(), G.dot() -> needs to be 2D numpy array
# - h: used in concatenation and arithmetic -> needs to be 1D numpy array
# - z0: optional, used with .copy() -> needs to be 1D numpy array or None

# From the code:
# nDual, nPrimal = G.shape
# G is (nDual, nPrimal)
# h has length nDual (used in: G.dot(y) + h - t * np.ones(nDual) + s)
# y has length nPrimal (initialized as np.full(nPrimal, 0.5))
# z has length nDual (initialized as np.ones(nDual) / nDual)
# s has length nDual (initialized as np.ones(nDual))

# Let's use distinct dimensions:
# nDual = 3 (number of constraints/dual variables)
# nPrimal = 4 (number of primal variables)

nDual = 3
nPrimal = 4

# G should be (nDual, nPrimal)
G = np.random.randn(nDual, nPrimal)

# h should be (nDual,)
h = np.random.randn(nDual)

# z0 should be (nDual,) - optional initial dual variable
z0 = np.ones(nDual) / nDual

ARGS = {
    "G": G,
    "h": h,
    "z0": z0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
proj_newton_logistic Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/lib/bundle_entropy_dual.py
code served (permissive licence) · get_code("ec0bdd07185128b8")
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

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

# Create small test matrices
# A should be (n, m) where n is the number of constraints and m is the dimension
# b should be (n,)
n = 4  # number of rows
m = 3  # number of columns

A = np.random.randn(n, m)
b = np.random.randn(n)

# lam0 is optional, should be (n,) and sum to 1 if provided
lam0 = np.ones(n) / n

# line_search is a boolean flag
line_search = False

ARGS = {
    "A": A,
    "b": b,
    "lam0": lam0,
    "line_search": line_search,
}
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
rolling Ran
ran on a synthesized input
lit-leo/inner-convex-nn/icnn_fixed/RL/analyze.random-search.py
code served (permissive licence) · get_code("fe1e956663360614")
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 (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

# The rolling function takes:
# N: an integer representing the window size for the rolling average
# rew: a numpy array of rewards/values to convolve

# Create a simple reward array with distinct size
rew = np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0])

# Create window size N
N = 3

ARGS = {
    'N': N,
    'rew': rew
}
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 (5,) — shape (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
load_batch Not yet run
unverified (harvested, no recorded run)
lit-leo/inner-convex-nn/vae/vanila_vae.py
code served (permissive licence) · get_code("79775596097e420c")
How it ran

no invocation recorded.

load_pickle Not yet run
unverified (harvested, no recorded run)
lit-leo/inner-convex-nn/vae/util.py
code served (permissive licence) · get_code("773fab84a811e1fd")
How it ran

no invocation recorded.

makeTable Not yet run
unverified (harvested, no recorded run)
lit-leo/inner-convex-nn/icnn_fixed/RL/analyze.random-search.py
code served (permissive licence) · get_code("c110d70bcd7b5b6a")
How it ran

no invocation recorded.

plot_conj Not yet run
unverified (harvested, no recorded run)
this paper's copy was not recorded; identical code first harvested from facebookresearch/w2ot
pointer only · get_code("7fb9bfa9b31ec695")
How it ran

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

train Not yet run
unverified (harvested, no recorded run)
lit-leo/inner-convex-nn/vae/vanila_vae.py
code served (permissive licence) · get_code("c5f18230130906c0")
How it ran

no invocation recorded.

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, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 1 later) and 13 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
plot_conj 4 other papers
1 dated earlier, 3 later
1902.07197 (2019-02)
2-Wasserstein Approximation via Restricted Convex Potentials with Application to Improved Training for GANs
2106.01954 (2021-06)
Do Neural Optimal Transport Solvers Work? A Continuous Wasserstein-2 Benchmark
2110.02999 (2021-10)
Generative Modeling with Optimal Transport Maps
and 1 more not named here
loss_function 1 other paper
1 dated earlier
1610.00291 (2016-10)
Deep Feature Consistent Variational Autoencoder

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

We propose a novel end-to-end non-minimax algorithm for training optimal transport mappings for the quadratic cost (Wasserstein-2 distance). The algorithm uses input convex neural networks and a cycle-consistency regularization to approximate Wasserstein-2 distance. In contrast to popular entropic and quadratic regularizers, cycle-consistency does not introduce bias and scales well to high dimensions. From the theoretical side, we estimate the properties of the generative mapping fitted by our algorithm. From the practical side, we evaluate our algorithm on a wide range of tasks: image-to-image color transfer, latent space optimal transport, image-to-image style transfer, and domain adaptation.

For agents

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

get_harvested_code_for_paper("1909.13082")
get_code_for_paper("1909.13082")
have("1909.13082")

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

curl https://syntology.ai/api/ran/1909.13082.json

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

[![Syntology run record](https://syntology.ai/api/ran/1909.13082.svg)](https://syntology.ai/paper/1909.13082)

Connect an agent — have() is free.