SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2105.04551 · CVPR · 2021

Stochastic Image-to-Video Synthesis using cINNs

Timo Milbich, Björn Ommer, Robin Rombach, Konstantinos Derpanis, Andreas Blattmann, Michael Dorkenwald

arXiv · PDF · Open in the Atlas

Code that ran

6 of 8 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 0 failures of our instrument; 2 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/2105.04551.json · badge: /api/ran/2105.04551.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
CompVis/image2video-synthesis-using-cINNs marked official in Papers with Code
Papers with Code archive
6 ran · 2 no recorded run
FunctionStatusWhere it lives
ActNorm Ran
ran on a synthesized input · constructed an object; no result was computed
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("d2651b54a261ca07")
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, ActNorm: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a minimal ActNorm module instance
actnorm = nn.Module()

# Define ARGS dictionary with all parameters for ActNorm.__init__
ARGS = {
    'num_features': 4,
    'logdet': True,
    'affine': 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 ActNorm — 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 ActNorm 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
BasicFullyConnectedNet Ran
ran on a synthesized input · constructed an object; no result was computed
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("f21483d52215e239")
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, BasicFullyConnectedNet: 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": 4,
    "depth": 2,
    "hidden_dim": 256,
    "use_tanh": False,
    "use_bn": False,
    "out_dim": 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 BasicFullyConnectedNet — 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 BasicFullyConnectedNet 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
ConditionalDoubleVectorCouplingBlock Ran
ran on a synthesized input · constructed an object; no result was computed
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("8bcd8f96cfc92e2f")
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, ConditionalDoubleVectorCouplingBlock: 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)

# Define the ARGS dictionary with all required parameters
ARGS = {
    'in_channels': 4,
    'cond_channels': 3,
    'hidden_dim': 64,
    'depth': 2,
    'mode': 'normal'
}

# Create sample tensors for testing the forward pass
x = torch.randn(2, 4, 1, 1)  # batch_size=2, in_channels=4, spatial dims 1x1
xc = torch.randn(2, 3, 1, 1)  # batch_size=2, cond_channels=3, spatial dims 1x1
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 ConditionalDoubleVectorCouplingBlock — 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 ConditionalDoubleVectorCouplingBlock 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
IgnoreLeakyRelu Ran
ran on a synthesized input
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("04e7c12c96f4fe9b")
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 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 torch.nn as nn

class IgnoreLeakyRelu(nn.Module):
    """performs identity op."""

    def __init__(self):
        super().__init__()

    def forward(self, input, reverse=False):
        if reverse:
            return self.reverse(input)
        h = input
        return (h, 0.0)

    def reverse(self, input):
        h = input
        return h


INIT_ARGS = {}

ARGS = {
    'input': torch.randn(2, 4),
    'reverse': False
}

if __name__ == '__main__':
    model = IgnoreLeakyRelu(**INIT_ARGS)
    output = model(**ARGS)
    print("Output:", output)
    print("Output type:", type(output))
    print("Output[0] shape:", output[0].shape)
    print("Output[1]:", output[1])
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 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
InvLeakyRelu Ran
ran on a synthesized input
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("aa63e950dd47091c")
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 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 torch.nn as nn

class InvLeakyRelu(nn.Module):

    def __init__(self, alpha=0.9):
        super().__init__()
        self.alpha = alpha

    def forward(self, input, reverse=False):
        if reverse:
            return self.reverse(input)
        scaling = (input >= 0).to(input) + (input < 0).to(input) * self.alpha
        h = input * scaling
        return (h, 0.0)

    def reverse(self, input):
        scaling = (input >= 0).to(input) + (input < 0).to(input) * self.alpha
        h = input / scaling
        return h


INIT_ARGS = {
    'alpha': 0.9
}

ARGS = {
    'input': torch.randn(2, 4),
    'reverse': False
}

if __name__ == '__main__':
    model = InvLeakyRelu(**INIT_ARGS)
    output = model(**ARGS)
    print("Forward output:", output)
    
    ARGS_reverse = {
        'input': torch.randn(2, 4),
        'reverse': True
    }
    output_reverse = model(**ARGS_reverse)
    print("Reverse output:", output_reverse)
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 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Shuffle Ran
ran on a synthesized input · constructed an object; no result was computed
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("8527bd367d67eed6")
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, Shuffle: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

class Shuffle(nn.Module):

    def __init__(self, in_channels):
        super(Shuffle, self).__init__()
        self.in_channels = in_channels
        idx = torch.randperm(in_channels)
        self.register_buffer('forward_shuffle_idx', idx)
        self.register_buffer('backward_shuffle_idx', torch.argsort(idx))

    def forward(self, x, reverse=False, conditioning=None):
        if not reverse:
            return (x[:, self.forward_shuffle_idx, ...], 0)
        else:
            return x[:, self.backward_shuffle_idx, ...]

ARGS = {
    'in_channels': 4
}
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 Shuffle — 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 Shuffle 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
ConditionalFlatDoubleCouplingFlowBlock Not yet run
unverified (harvested, no recorded run)
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("53efe854b4513e3d")
How it ran

no invocation recorded.

ConditionalFlow Not yet run
unverified (harvested, no recorded run)
CompVis/image2video-synthesis-using-cINNs/stage2_cINN/modules/flow_blocks.py
code served (permissive licence) · get_code("a025da90487eae60")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 8 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.

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

Video understanding calls for a model to learn the characteristic interplay between static scene content and its dynamics: Given an image, the model must be able to predict a future progression of the portrayed scene and, conversely, a video should be explained in terms of its static image content and all the remaining characteristics not present in the initial frame. This naturally suggests a bijective mapping between the video domain and the static content as well as residual information. In contrast to common stochastic image-to-video synthesis, such a model does not merely generate arbitrary videos progressing the initial image. Given this image, it rather provides a one-to-one mapping between the residual vectors and the video with stochastic outcomes when sampling. The approach is naturally implemented using a conditional invertible neural network (cINN) that can explain videos by independently modelling static and other video characteristics, thus laying the basis for controlled video synthesis. Experiments on diverse video datasets demonstrate the effectiveness of our approach in terms of both the quality and diversity of the synthesized results. Our project page is available at https://bit.ly/3dg90fV . * Indicates equal supervision.

For agents

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

get_harvested_code_for_paper("2105.04551")
get_code_for_paper("2105.04551")
have("2105.04551")

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

curl https://syntology.ai/api/ran/2105.04551.json

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

[![Syntology run record](https://syntology.ai/api/ran/2105.04551.svg)](https://syntology.ai/paper/2105.04551)

Connect an agent — have() is free.