SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2002.09741 · 2020

VFlow: More Expressive Generative Flows with Variational Data Augmentation

arXiv · PDF · Open in the Atlas

Code that ran

4 of 9 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 4 have no contract verdict; 0 failures of our instrument; 5 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/2002.09741.json · badge: /api/ran/2002.09741.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
thu-ml/vflow canonical (classifier label)
classifier basis: census
4 ran · 5 no recorded run
FunctionStatusWhere it lives
at_least_float32 Ran
ran on a synthesized input
thu-ml/vflow/flows_imagenet/imagenet32.py
code served (permissive licence) · get_code("5f0f74ae9d431c1c")
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 EagerTensor, 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 tensorflow as tf

# Create test tensors with different dtypes
x_float16 = tf.constant([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0]], dtype=tf.float16)
x_float32 = tf.constant([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0]], dtype=tf.float32)
x_float64 = tf.constant([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0]], dtype=tf.float64)

# The function takes one parameter 'x' which should be a tensor with dtype in [tf.float16, tf.float32, tf.float64]
ARGS = {
    "x": x_float32
}
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 EagerTensor, 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
tile_imgs Ran
ran on a synthesized input
thu-ml/vflow/flows_imagenet/utils.py
code served (permissive licence) · get_code("f7bc5864104e41ee")
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 int, shape (10, 12, 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 numpy as np

# Create a small batch of images for testing
# Create 3 small grayscale images (1 channel each)
img1 = np.random.randint(0, 256, size=(4, 5, 1), dtype=np.uint8)
img2 = np.random.randint(0, 256, size=(4, 5, 1), dtype=np.uint8)
img3 = np.random.randint(0, 256, size=(4, 5, 1), dtype=np.uint8)

imgs = np.stack([img1, img2, img3], axis=0)

ARGS = {
    "imgs": imgs,
    "pad_pixels": 1,
    "pad_val": 255,
}
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 int, shape (10, 12, 1) — shape (10, 12, 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
tile_imgs Ran
ran on a synthesized input
thu-ml/vflow/flows/utils.py
code served (permissive licence) · get_code("77ffd1982c6a4f9a")
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 int, shape (10, 8, 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 test images
# Using 2 images of size 4x3 with 3 channels (RGB)
img1 = np.random.randint(0, 256, size=(4, 3, 3), dtype=np.uint8)
img2 = np.random.randint(0, 256, size=(4, 3, 3), dtype=np.uint8)
imgs = np.stack([img1, img2], axis=0)

ARGS = {
    'imgs': imgs,
    'pad_pixels': 1,
    'pad_val': 255,
}
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 int, shape (10, 8, 3) — shape (10, 8, 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
to_default_floatx Ran
ran on a synthesized input
thu-ml/vflow/flows_imagenet/imagenet32.py
code served (permissive licence) · get_code("60187f793be7f93d")
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 EagerTensor, 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 tensorflow as tf

DEFAULT_FLOATX = tf.float32

def to_default_floatx(x):
    return tf.cast(x, DEFAULT_FLOATX)

# Create a small tensor with a different dtype to cast
x_int32 = tf.constant([1, 2, 3, 4], dtype=tf.int32)

ARGS = {
    "x": x_int32
}
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 EagerTensor, 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
assert_in_range Not yet run
unverified (harvested, no recorded run)
thu-ml/vflow/flows_imagenet/logistic.py
code served (permissive licence) · get_code("65754dd997703c12")
How it ran

no invocation recorded.

conv2d Not yet run
unverified (harvested, no recorded run)
thu-ml/vflow/flows/flows.py
code served (permissive licence) · get_code("6e9d0a8c6f8d3fae")
How it ran

no invocation recorded.

dense Not yet run
unverified (harvested, no recorded run)
thu-ml/vflow/flows/flows.py
code served (permissive licence) · get_code("9bab91e30fb7e13f")
How it ran

no invocation recorded.

get_var Not yet run
unverified (harvested, no recorded run)
thu-ml/vflow/flows_imagenet/imagenet32.py
code served (permissive licence) · get_code("952f606576c632c7")
How it ran

no invocation recorded.

get_var Not yet run
unverified (harvested, no recorded run)
thu-ml/vflow/flows/flows.py
code served (permissive licence) · get_code("9d6d831a4e81b501")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 7 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 0 later) and 2 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
assert_in_range 1 other paper
1 dated earlier
1902.00275 (2019-02)
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
at_least_float32 1 other paper
1 dated earlier
1902.00275 (2019-02)
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
get_var 1 other paper
1 dated earlier
1902.00275 (2019-02)
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
get_var 1 other paper
1 dated earlier
1902.00275 (2019-02)
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
tile_imgs 1 other paper
1 dated earlier
1902.00275 (2019-02)
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
tile_imgs 1 other paper
1 dated earlier
1902.00275 (2019-02)
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design
to_default_floatx 1 other paper
1 dated earlier
1902.00275 (2019-02)
Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design

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

Generative flows are promising tractable models for density modeling that define probabilistic distributions with invertible transformations. However, tractability imposes architectural constraints on generative flows, making them less expressive than other types of generative models. In this work, we study a previously overlooked constraint that all the intermediate representations must have the same dimensionality with the original data due to invertibility, limiting the width of the network. We tackle this constraint by augmenting the data with some extra dimensions and jointly learning a generative flow for augmented data as well as the distribution of augmented dimensions under a variational inference framework. Our approach, VFlow, is a generalization of generative flows and therefore always performs better. Combining with existing generative flows, VFlow achieves a new state-of-the-art 2.98 bits per dimension on the CIFAR-10 dataset and is more compact than previous models to reach similar modeling quality.

For agents

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

get_harvested_code_for_paper("2002.09741")
get_code_for_paper("2002.09741")
have("2002.09741")

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

curl https://syntology.ai/api/ran/2002.09741.json

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

[![Syntology run record](https://syntology.ai/api/ran/2002.09741.svg)](https://syntology.ai/paper/2002.09741)

Connect an agent — have() is free.