SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1509.00519 · 2015

Importance Weighted Autoencoders

Ruslan Salakhutdinov, Roger Grosse, Yuri Burda

arXiv · PDF · Open in the Atlas

Code that ran

20 of 29 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (6 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 18 have no contract verdict; 2 failures of our instrument; 9 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/1509.00519.json · badge: /api/ran/1509.00519.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
jurgisp/pydreamer community (classifier label)
classifier basis not recorded on the link
7 ran · 0 no recorded run
alexey-pronkin/annealed community
Papers with Code archive
3 ran · 2 no recorded run
madhubabuv/TightIWAE community
Papers with Code archive
3 ran · 0 no recorded run
harvardnlp/sa-vae community
Papers with Code archive
1 ran · 5 no recorded run
adrianjav/heterogeneous_vaes community
Papers with Code archive
1 ran · 2 no recorded run
andrecavalcante/iwae community
Papers with Code archive
1 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
la3ma/IWAE-torch community
Papers with Code archive
1 ran · 0 no recorded run
JohanYe/IWAE-pytorch community
Papers with Code archive
1 ran · 0 no recorded run
angzhifan/Importance_Weighted_Autoencoders community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
Decoder Ran
ran on a synthesized input · constructed an object; no result was computed
madhubabuv/TightIWAE/model.py
pointer only (licence: GPL-3.0) · get_code("345529606c7ce3f1")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, madhubabuv/TightIWAE, is not in the permissive set Syntology serves inline. 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.

Encoder Ran
ran on a synthesized input · constructed an object; no result was computed
madhubabuv/TightIWAE/model.py
pointer only (licence: GPL-3.0) · get_code("95c347831411d92d")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, madhubabuv/TightIWAE, is not in the permissive set Syntology serves inline. 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.

HVITransition Ran
ran on a synthesized input
alexey-pronkin/annealed/avo/models/vae_hvi.py
pointer only (licence: NONE) · get_code("09532121d9834645")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, alexey-pronkin/annealed 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.

IWAE Ran
ran on a synthesized input
andrecavalcante/iwae/iwae.py
pointer only (licence: NONE) · get_code("35d63939c447a9f8")
How it ran

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

IWAE_KL Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from denproc/Taming-VAEs
pointer only; the stored copy get_code serves, from denproc/Taming-VAEs, is pointer only (licence: NONE) · get_code("5fb1d231ba9d1ff8")
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. A run is recorded.

ImportanceWeightedAutoencoder Ran
ran on a synthesized input
la3ma/IWAE-torch/src/models.py
code served (permissive licence) · get_code("9b0570a716391574")
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
import numpy as np

# INIT_ARGS for ImportanceWeightedAutoencoder.__init__
INIT_ARGS = {
    'latent_dim': 3,
    'data_dim': 5,
    'hidden_dim': 4
}

# ARGS for forward method (which is called when instance is invoked)
# The forward method signature is: forward(self, x, mc_samples=1)
# We need to provide x and optionally mc_samples
batch_size = 2
data_dim = 5  # must match INIT_ARGS['data_dim']

ARGS = {
    'x': torch.randn(batch_size, data_dim),
    'mc_samples': 2
}
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 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)
MaskedConv2d Ran
ran on a synthesized input
harvardnlp/sa-vae/models_img.py
pointer only (licence: NONE) · get_code("0df0f85652bb400e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, harvardnlp/sa-vae 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.

PytorchIWAE Ran
ran on a synthesized input · constructed an object; no result was computed
JohanYe/IWAE-pytorch/model/PytorchIWAE.py
pointer only (licence: NONE) · get_code("4bb10cc9c1249911")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, JohanYe/IWAE-pytorch 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.

RefinementOperatorModule Ran
ran on a synthesized input
alexey-pronkin/annealed/avo/models/vae_hvi.py
pointer only (licence: NONE) · get_code("c6fa73ec552e34c9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, alexey-pronkin/annealed 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.

VAE Ran
ran on a synthesized input · constructed an object; no result was computed
madhubabuv/TightIWAE/model.py
pointer only (licence: GPL-3.0) · get_code("a4a3b579a1c261fb")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, madhubabuv/TightIWAE, is not in the permissive set Syntology serves inline. 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.

VAEHead Ran
ran on a synthesized input · constructed an object; no result was computed
adrianjav/heterogeneous_vaes/src/models/iwae.py
pointer only (licence: GPL-3.0) · get_code("22dff8581dcfbf0a")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, adrianjav/heterogeneous_vaes, is not in the permissive set Syntology serves inline. 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.

VAE_2 Ran
ran on a synthesized input · constructed an object; no result was computed
angzhifan/Importance_Weighted_Autoencoders/model.py
pointer only (licence: NONE) · get_code("331f0ac646f63a0a")
How it ran

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

flatten Ran
ran on a synthesized input
jurgisp/pydreamer/pydreamer/models/functions.py
code served (permissive licence) · get_code("a3ec62a9c649b4eb")
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 (8, 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

torch.manual_seed(0)

# The function takes a Tensor x and reshapes it
# It flattens dimensions 0 and 1 while keeping dimensions 2 onward intact
# We need a tensor with at least 2 dimensions to test this properly

x = torch.randn(2, 4, 3, 5)  # batch=2, dim1=4, spatial dims=3x5

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 3, dtype family float, shape (8, 3, 5) — shape (8, 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
flatten_batch Ran
ran on a synthesized input
jurgisp/pydreamer/pydreamer/models/functions.py
code served (permissive licence) · get_code("8e9e702c0c99753a")
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
from torch import Tensor, Size

torch.manual_seed(0)

# Test case 1: 3D tensor with nonbatch_dims=1
x1 = torch.randn(2, 3, 4)
nonbatch_dims1 = 1

# Test case 2: 4D tensor with nonbatch_dims=2
x2 = torch.randn(2, 3, 5, 6)
nonbatch_dims2 = 2

# Test case 3: 2D tensor with nonbatch_dims=0
x3 = torch.randn(2, 7)
nonbatch_dims3 = 0

# We'll use the first test case as the primary argument set
ARGS = {
    "x": x1,
    "nonbatch_dims": nonbatch_dims1
}
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
get_activation Ran
ran; our contract draft was wrong, not the code
alexey-pronkin/annealed/avo/models/vae_hvi.py
pointer only (licence: NONE) · get_code("14576598bdf9154f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, alexey-pronkin/annealed 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.

img_to_onehot Ran
ran on a synthesized input
jurgisp/pydreamer/pydreamer/preprocessing.py
code served (permissive licence) · get_code("76b1176ac4bedca4")
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 5, dtype family float, shape (2, 4, 3, 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

# Analyze the function:
# img_to_onehot takes:
# - x: np.ndarray (image tensor with shape that will be indexed and transposed)
# - n_categories: int (number of categories for one-hot encoding)
#
# The function:
# 1. Calls to_onehot(x, n_categories) which does e[x] where e is (n_categories, n_categories)
#    This means x should contain integer indices in range [0, n_categories)
#    After e[x], output shape is x.shape + (n_categories,)
# 2. Transposes result with (0, 1, 4, 2, 3)
#    This means the output from to_onehot should have 5 dimensions
#    So x.shape must have 4 dimensions (since to_onehot adds one dimension at the end)
#
# So x should be a 4D array of shape (batch, channels, height, width) or similar
# with integer values in [0, n_categories)

n_categories = 3

# Create a 4D image-like array with shape (2, 4, 3, 5) - batch=2, channels=4, height=3, width=5
# After to_onehot with n_categories=3, shape becomes (2, 4, 3, 5, 3)
# After transpose(0, 1, 4, 2, 3), shape becomes (2, 4, 3, 3, 5)
x = np.random.randint(0, n_categories, size=(2, 4, 3, 5), dtype=np.int32)

ARGS = {
    "x": x,
    "n_categories": n_categories
}
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 5, dtype family float, shape (2, 4, 3, 3, 5) — shape (2, 4, 3, 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
lenb Ran
ran on a synthesized input
jurgisp/pydreamer/pydreamer/data.py
code served (permissive licence) · get_code("72f12f98266796c4")
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 int, ndim 0, dtype family int, 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)

batch = {
    'reward': torch.randn(2, 4)
}

ARGS = {
    'batch': batch
}
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 int, ndim 0, dtype family int, 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
to_image Ran
ran on a synthesized input
jurgisp/pydreamer/pydreamer/preprocessing.py
code served (permissive licence) · get_code("291954b28ca216e7")
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 5, dtype family float, shape (2, 3, 6, 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 numpy as np

# Create a small valid input array
# The function expects either uint8 or float32 in range [0, 1]
# Shape analysis: x.transpose(0, 1, 4, 2, 3) means input needs dims (0, 1, 2, 3, 4)
# So input shape should be (batch, ?, ?, ?, channels)
# Let's use shape (2, 3, 4, 5, 6) for distinct dimensions

# Test with float32 in valid range [0, 1]
x_float = np.random.rand(2, 3, 4, 5, 6).astype(np.float32)

ARGS = {
    "x": x_float
}
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 5, dtype family float, shape (2, 3, 6, 4, 5) — shape (2, 3, 6, 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
to_onehot Ran
ran on a synthesized input
jurgisp/pydreamer/pydreamer/preprocessing.py
code served (permissive licence) · get_code("5f3793a255e72f19")
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 (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 numpy as np

# Create a small integer array to be converted to one-hot
x = np.array([0, 1, 2, 1], dtype=np.int32)

# Number of categories
n_categories = 3

ARGS = {
    "x": x,
    "n_categories": n_categories,
}
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 (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
unflatten Ran
ran on a synthesized input
jurgisp/pydreamer/pydreamer/models/functions.py
code served (permissive licence) · get_code("b26ca1623f8bfc99")
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, 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 Tensor

torch.manual_seed(0)

# Create a small tensor to unflatten
# Shape: (12,) - a 1D tensor with 12 elements
x = torch.randn(12)

# n parameter: number of rows to reshape into
n = 3

ARGS = {
    "x": x,
    "n": n,
}
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, 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
CNNVAE Not yet run
unverified (harvested, no recorded run)
harvardnlp/sa-vae/models_img.py
pointer only (licence: NONE) · get_code("7c34da1793d4495f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, harvardnlp/sa-vae 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. No invocation recorded.

GatedMaskedConv2d Not yet run
unverified (harvested, no recorded run)
harvardnlp/sa-vae/models_img.py
pointer only (licence: NONE) · get_code("f16612739812b34c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, harvardnlp/sa-vae 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. No invocation recorded.

IWAE Not yet run
unverified (harvested, no recorded run)
adrianjav/heterogeneous_vaes/src/models/iwae.py
pointer only (licence: GPL-3.0) · get_code("215e8a7b9e225449")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, adrianjav/heterogeneous_vaes, is not in the permissive set Syntology serves inline. 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.

ResidualBlock Not yet run
unverified (harvested, no recorded run)
harvardnlp/sa-vae/models_img.py
pointer only (licence: NONE) · get_code("8caea4e84f7e6c2d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, harvardnlp/sa-vae 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. No invocation recorded.

StackedGatedMaskedConv2d Not yet run
unverified (harvested, no recorded run)
harvardnlp/sa-vae/models_img.py
pointer only (licence: NONE) · get_code("df38b0e37366392c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, harvardnlp/sa-vae 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. No invocation recorded.

VAE Not yet run
unverified (harvested, no recorded run)
adrianjav/heterogeneous_vaes/src/models/iwae.py
pointer only (licence: GPL-3.0) · get_code("a4f76a4d4d881d17")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, adrianjav/heterogeneous_vaes, is not in the permissive set Syntology serves inline. 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.

VAE Not yet run
unverified (harvested, no recorded run)
alexey-pronkin/annealed/avo/models/vae_hvi.py
pointer only (licence: NONE) · get_code("f79c21f1eeefc8f3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, alexey-pronkin/annealed 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. No invocation recorded.

VAEHVI Not yet run
unverified (harvested, no recorded run)
alexey-pronkin/annealed/avo/models/vae_hvi.py
pointer only (licence: NONE) · get_code("a9665ebff625b64f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, alexey-pronkin/annealed 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. No invocation recorded.

he_init Not yet run
unverified (harvested, no recorded run)
harvardnlp/sa-vae/models_img.py
pointer only (licence: NONE) · get_code("a61c21baef53a292")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, harvardnlp/sa-vae 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. No invocation recorded.

Also present in other papers' code

Of 29 samples lifted from this paper's linked repositories, 29 carry a function identity and 0 do not. Of the 29 distinct functions with an identity, 5 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 2 later) and 24 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
IWAE_KL
brick of this paper's assembly: entry
1 other paper
1 later
1810.00597 (2018-10)
Taming VAEs
VAE
brick of this paper's assembly: model
1 other paper
1 dated earlier
1312.6114 (2013-12)
Auto-Encoding Variational Bayes
VAE
brick of this paper's assembly: helper
1 other paper
1 dated earlier
1312.6114 (2013-12)
Auto-Encoding Variational Bayes
VAEHead
brick of this paper's assembly: helper
1 other paper
1 dated earlier
1312.6114 (2013-12)
Auto-Encoding Variational Bayes
he_init
brick of this paper's assembly: helper
1 other paper
1 later · 1 only as an assembly brick
2004.04795 (2020-04)
Exemplar VAE: Linking Generative Models, Nearest Neighbor Retrieval, and Data Augmentation

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

More are linked than shown here.

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

Abstract

The variational autoencoder (VAE; Kingma & Welling (2014)) is a recently proposed generative model pairing a top-down generative network with a bottom-up recognition network which approximates posterior inference. It typically makes strong assumptions about posterior inference, for instance that the posterior distribution is approximately factorial, and that its parameters can be approximated with nonlinear regression from the observations. As we show empirically, the VAE objective can lead to overly simplified representations which fail to use the network's entire modeling capacity. We present the importance weighted autoencoder (IWAE), a generative model with the same architecture as the VAE, but which uses a strictly tighter log-likelihood lower bound derived from importance weighting. In the IWAE, the recognition network uses multiple samples to approximate the posterior, giving it increased flexibility to model complex posteriors which do not fit the VAE modeling assumptions. We show empirically that IWAEs learn richer latent space representations than VAEs, leading to improved test log-likelihood on density estimation benchmarks.

For agents

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

get_harvested_code_for_paper("1509.00519")
get_code_for_paper("1509.00519")
have("1509.00519")

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

curl https://syntology.ai/api/ran/1509.00519.json

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

[![Syntology run record](https://syntology.ai/api/ran/1509.00519.svg)](https://syntology.ai/paper/1509.00519)

Connect an agent — have() is free.