SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.00626 · NeurIPS · 2024

Maximum Entropy Inverse Reinforcement Learning of Diffusion Models with Energy-Based Models

Sangwoong Yoon, Yung-Kyun Noh, Frank Park, Dohyun Kwon, Himchan Hwang

arXiv · PDF · Open in the Atlas

Code that ran

11 of 18 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; 4 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 6 failures of our instrument; 7 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/2407.00626.json · badge: /api/ran/2407.00626.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
swyoon/Diffusion-by-MaxEntIRL canonical (classifier label)
classifier basis: census
9 ran · 5 no recorded run
swyoon/diffusion-by-maxentirl canonical (classifier label)
classifier basis: classifier_2025
2 ran · 2 no recorded run
FunctionStatusWhere it lives
calculate_frechet_distance Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
swyoon/Diffusion-by-MaxEntIRL/pytorch_fid/fid_score.py
pointer only (licence: MIT); the stored copy get_code serves, from NVlabs/denoising-diffusion-gan, is pointer only (licence: NOASSERTION) · get_code("fcdc92d3d2e8edc1")
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 NVlabs/denoising-diffusion-gan, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

get_activation Ran
ran; our contract draft was wrong, not the code
swyoon/Diffusion-by-MaxEntIRL/models/modules.py
code served (permissive licence) · get_code("6c1e362234350951")
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 ReLU. 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 nn

torch.manual_seed(0)

ARGS = {
    's_act': 'relu'
}
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 ReLU — a ReLU output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ReLU 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_weightings Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
swyoon/Diffusion-by-MaxEntIRL/models/cm/karras_diffusion.py
code served (permissive licence) · get_code("b50e02eda098b1ea")
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 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 torch as th

th.manual_seed(0)

# Create small tensors for snrs and sigma_data
# snrs: signal-to-noise ratios tensor
snrs = th.tensor([1.0, 2.0, 3.0, 4.0])

# sigma_data: noise level parameter
sigma_data = th.tensor([0.5, 0.6, 0.7, 0.8])

# weight_schedule: string parameter that determines which weighting scheme to use
weight_schedule = 'snr'

ARGS = {
    'weight_schedule': weight_schedule,
    'snrs': snrs,
    'sigma_data': sigma_data,
}
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 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
make_beta_schedule Ran
ran, honoured the contract we drafted
swyoon/diffusion-by-maxentirl/models/DxMI/trainer.py
code served (permissive licence) · get_code("01baa8e1e4e273e2")
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) and returned Tensor, ndim 1, dtype family float, shape (1000,) 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 = {
    'schedule': 'linear',
    'n_timesteps': 1000,
    'start': 1e-05,
    'end': 0.01
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 1, dtype family float, shape (1000,) — shape (1000,) 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
parse_arg_type Ran
ran, honoured the contract we drafted
swyoon/Diffusion-by-MaxEntIRL/cmd_utils.py
code served (permissive licence) · get_code("3a94534d8de246ee")
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
import numpy as np

torch.manual_seed(0)

ARGS = {
    'val': '42'
}
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 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
parse_nested_args Ran
ran; our contract draft was wrong, not the code
swyoon/Diffusion-by-MaxEntIRL/cmd_utils.py
code served (permissive licence) · get_code("2312bb33e056cfa7")
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 torch

def parse_nested_args(d_cmd_cfg):
    """produce a nested dictionary by parsing dot-separated keys
    e.g. {key1.key2 : 1}  --> {key1: {key2: 1}}"""
    d_new_cfg = {}
    for key, val in d_cmd_cfg.items():
        l_key = key.split('.')
        d = d_new_cfg
        for i_key, each_key in enumerate(l_key):
            if i_key == len(l_key) - 1:
                d[each_key] = val
            else:
                if each_key not in d:
                    d[each_key] = {}
                d = d[each_key]
    return d_new_cfg

ARGS = {
    'd_cmd_cfg': {
        'model.hidden_size': 256,
        'model.num_layers': 3,
        'training.learning_rate': 0.001,
        'training.batch_size': 32,
        'data.path': '/tmp/data',
        'optimizer.weight_decay': 0.0001
    }
}
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 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
parse_unknown_args Ran
ran; our contract draft was wrong, not the code
swyoon/Diffusion-by-MaxEntIRL/cmd_utils.py
code served (permissive licence) · get_code("24050cdd9df51454")
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 torch
import numpy as np

torch.manual_seed(0)

def parse_arg_type(val):
    if val.isnumeric():
        return int(val)
    try:
        return float(val)
    except ValueError:
        if val.lower() == 'true':
            return True
        elif val.lower() == 'false':
            return False
        elif val.lower() == 'null' or val.lower() == 'none':
            return None
        elif val.startswith('[') and val.endswith(']'):
            return eval(val)
        return val

def parse_unknown_args(l_args):
    """convert the list of unknown args into dict
    this does similar stuff to OmegaConf.from_cli()
    I may have invented the wheel again..."""
    n_args = len(l_args) // 2
    kwargs = {}
    for i_args in range(n_args):
        key = l_args[i_args * 2]
        val = l_args[i_args * 2 + 1]
        assert '=' not in key, 'optional arguments should be separated by space'
        kwargs[key.strip('-')] = parse_arg_type(val)
    return kwargs

ARGS = {
    'l_args': [
        '--learning_rate', '0.001',
        '--epochs', '100',
        '--enable_training', 'true',
        '--use_cache', 'false',
        '--config_path', 'null',
        '--hidden_dims', '[64, 128, 256]',
        '--model_name', 'bert-base',
        '-momentum', '0.9'
    ]
}
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 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
process_single_t Ran
ran, violated the contract we drafted
swyoon/diffusion-by-maxentirl/models/DxMI/trainer.py
code served (permissive licence) · get_code("64a941eba4950213")
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) and returned Tensor, ndim 1, dtype family int, shape (2,) 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)

# Create a small tensor x with batch size 2 and feature dimension 4
x = torch.randn(2, 4)

# Create a scalar integer t
t = 5

ARGS = {
    "x": x,
    "t": t,
}
Setup block drafted by
a language model, Claude Haiku 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 Tensor, ndim 1, dtype family int, shape (2,) — shape (2,) 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
remove_module Ran
ran; our contract draft was wrong, not the code
swyoon/Diffusion-by-MaxEntIRL/models/utils.py
code served (permissive licence) · get_code("90f2942b737cb482")
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 OrderedDict. 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
from collections import OrderedDict

torch.manual_seed(0)

# The function remove_module takes a dict 'd' and removes the 'module.' prefix
# from all keys. It expects a dict-like object with items() method.
# A common use case is removing 'module.' prefix from state_dicts of DataParallel models.

# Create a state dict with 'module.' prefix (as would come from nn.DataParallel)
state_dict = OrderedDict({
    'module.layer1.weight': torch.randn(4, 3),
    'module.layer1.bias': torch.randn(4),
    'module.layer2.weight': torch.randn(5, 4),
    'module.layer2.bias': torch.randn(5),
})

ARGS = {
    'd': state_dict
}
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 OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict 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
rescale Ran
ran, honoured the contract we drafted
swyoon/Diffusion-by-MaxEntIRL/generate_cifar10.py
code served (permissive licence) · get_code("4e07246a19ae09b7")
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) and returned Tensor, 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 torch
import numpy as np

torch.manual_seed(0)

# The function rescale takes X and an optional batch parameter
# It performs: (X - -1) / 2, which simplifies to (X + 1) / 2
# The batch parameter is not used in the function body, but we should provide it

# Create a small tensor for X
X = torch.randn(2, 4)

ARGS = {
    "X": X,
    "batch": True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 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
weight_norm Ran
ran, honoured the contract we drafted
swyoon/Diffusion-by-MaxEntIRL/models/utils.py
code served (permissive licence) · get_code("ffd7b0f9be1df615")
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 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
import torch.nn as nn

torch.manual_seed(0)

class SimpleNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

net = SimpleNet()

ARGS = {
    "net": net
}
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 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
DxMI_Trainer Not yet run
unverified (harvested, no recorded run)
swyoon/diffusion-by-maxentirl/models/DxMI/trainer.py
code served (permissive licence) · get_code("2feb70fa82de3bb5")
How it ran

no invocation recorded.

extract Not yet run
unverified (harvested, no recorded run)
swyoon/diffusion-by-maxentirl/models/DxMI/trainer.py
code served (permissive licence) · get_code("30e33c6c36337446")
How it ran

no invocation recorded.

fix_legacy_dict Not yet run
unverified (harvested, no recorded run)
swyoon/Diffusion-by-MaxEntIRL/models/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from inspire-group/proxy-distributions, is pointer only (licence: NONE) · get_code("995510dd78052fbc")
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 inspire-group/proxy-distributions, 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. No invocation recorded.

get_activations Not yet run
unverified (harvested, no recorded run)
swyoon/Diffusion-by-MaxEntIRL/pytorch_fid/fid_score.py
pointer only (licence: MIT); the stored copy get_code serves, from NVlabs/denoising-diffusion-gan, is pointer only (licence: NOASSERTION) · get_code("26edbd4467e04bec")
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 NVlabs/denoising-diffusion-gan, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

get_activations_from_tensor Not yet run
unverified (harvested, no recorded run)
swyoon/Diffusion-by-MaxEntIRL/pytorch_fid/fid_score.py
code served (permissive licence) · get_code("9226d84ff74e8d9f")
How it ran

no invocation recorded.

get_sigmas_karras Not yet run
unverified (harvested, no recorded run)
swyoon/Diffusion-by-MaxEntIRL/models/cm/karras_diffusion.py
pointer only (licence: MIT); the stored copy get_code serves, from guthasaibharathchandra/MAP-GA, is pointer only (licence: NONE) · get_code("00611e9feba0f93d")
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 guthasaibharathchandra/MAP-GA, 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. No invocation recorded.

karras_sample Not yet run
unverified (harvested, no recorded run)
swyoon/Diffusion-by-MaxEntIRL/models/cm/karras_diffusion.py
code served (permissive licence) · get_code("0caab2b1bad90636")
How it ran

no invocation recorded.

Also present in other papers' code

Of 18 samples lifted from this paper's linked repositories, 18 carry a function identity and 0 do not. Of the 18 distinct functions with an identity, 9 are also present in other papers whose code we hold (9 with at least one paper dated earlier, 2 the same month, 3 later) and 9 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
calculate_frechet_distance 6 other papers
4 dated earlier, 2 later
2106.00132 (2021-06)
On Fast Sampling of Diffusion Probabilistic Models
2303.05456 (2023-03)
Restoration based Generative Models
2311.17101 (2023-11)
A High-Quality Robust Diffusion Framework for Corrupted Dataset
and 3 more not named here
get_activations 5 other papers
3 dated earlier, 2 later
2106.00132 (2021-06)
On Fast Sampling of Diffusion Probabilistic Models
2303.05456 (2023-03)
Restoration based Generative Models
2312.12431 (2023-12)
On Inference Stability for Diffusion Models
and 2 more not named here
extract
brick of this paper's assembly: helper
3 other papers
3 dated earlier
2206.02262 (2022-06)
Diffusion-GAN: Training GANs with Diffusion
2304.11582 (2023-04)
DiffTraj: Generating GPS Trajectory with Diffusion Probabilistic Model
2405.09539 (2024-05)
MMFusion: Multi-modality Diffusion Model for Lymph Node Metastasis Diagnosis in Esophageal Cancer
get_sigmas_karras 3 other papers
2 dated earlier, 1 the same month
2303.01469 (2023-03)
Consistency Models
2305.02463 (2023-05)
Shap-E: Generating Conditional 3D Implicit Functions
2407.20784 (2024-07)
Inverse Problems with Diffusion Models: A MAP Estimation Perspective
get_weightings 3 other papers
1 dated earlier, 1 the same month, 1 later
2303.01469 (2023-03)
Consistency Models
2407.20784 (2024-07)
Inverse Problems with Diffusion Models: A MAP Estimation Perspective
2411.08378 (2024-11)
Physics Informed Distillation for Diffusion Models
fix_legacy_dict 1 other paper
1 dated earlier
2104.09425 (2021-04)
Robust Learning Meets Generative Models: Can Proxy Distributions Improve Adversarial Robustness?
karras_sample 1 other paper
1 dated earlier
2303.01469 (2023-03)
Consistency Models
remove_module 1 other paper
1 dated earlier
2401.03140 (2024-01)
Fair Sampling in Diffusion Models through Switching Mechanism
rescale 1 other paper
1 dated earlier
2301.13362 (2023-01)
Optimizing DDPM Sampling with Shortcut Fine-Tuning

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 present a maximum entropy inverse reinforcement learning (IRL) approach for improving the sample quality of diffusion generative models, especially when the number of generation time steps is small. Similar to how IRL trains a policy based on the reward function learned from expert demonstrations, we train (or fine-tune) a diffusion model using the log probability density estimated from training data. Since we employ an energy-based model (EBM) to represent the log density, our approach boils down to the joint training of a diffusion model and an EBM. Our IRL formulation, named Diffusion by Maximum Entropy IRL (DxMI), is a minimax problem that reaches equilibrium when both models converge to the data distribution. The entropy maximization plays a key role in DxMI, facilitating the exploration of the diffusion model and ensuring the convergence of the EBM. We also propose Diffusion by Dynamic Programming (DxDP), a novel reinforcement learning algorithm for diffusion models, as a subroutine in DxMI. DxDP makes the diffusion model update in DxMI efficient by transforming the original problem into an optimal control formulation where value functions replace back-propagation in time. Our empirical studies show that diffusion models fine-tuned using DxMI can generate high-quality samples in as few as 4 and 10 steps. Additionally, DxMI enables the training of an EBM without MCMC, stabilizing EBM training dynamics and enhancing anomaly detection performance.

For agents

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

get_harvested_code_for_paper("2407.00626")
get_code_for_paper("2407.00626")
have("2407.00626")

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

curl https://syntology.ai/api/ran/2407.00626.json

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

[![Syntology run record](https://syntology.ai/api/ran/2407.00626.svg)](https://syntology.ai/paper/2407.00626)

Connect an agent — have() is free.