SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.17268 · NeurIPS · 2021

Fast Certified Robust Training with Short Warmup

Huan Zhang, Cho-Jui Hsieh, A Research, Zhouxing Shi, Cmu, Yihan Wang, Jinfeng Yi

arXiv · PDF · Open in the Atlas

Code that ran

13 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; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 13 failures of our instrument; 5 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2103.17268.json · badge: /api/ran/2103.17268.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
shizhouxing/Fast-Certified-Robust-Training canonical (classifier label)
classifier basis: census
13 ran · 5 no recorded run
FunctionStatusWhere it lives
PreActResNet18 Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/resnet.py
code served (permissive licence) · get_code("abc02d2c96c29b28")
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 PreActResNet. 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 = {
    "in_ch": 3,
    "in_dim": 32,
}
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 PreActResNet — a PreActResNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PreActResNet 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
ResNeXt_2_32 Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/resnext_bn.py
code served (permissive licence) · get_code("fb928ba8123f9d4b")
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 ResNeXt. 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 Flatten(nn.Module):
    def forward(self, x):
        return x.view(x.size(0), -1)

ARGS = {
    "in_ch": 3,
    "in_dim": 32,
    "bn": 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 ResNeXt — a ResNeXt output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNeXt 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
ResNeXt_4_20 Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/resnext_bn.py
code served (permissive licence) · get_code("6c49fd440f42c595")
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 ResNeXt. 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 Flatten(nn.Module):
    def forward(self, x):
        return x.view(x.size(0), -1)

ARGS = {
    "in_ch": 3,
    "in_dim": 32,
}
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 ResNeXt — a ResNeXt output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNeXt 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
ResNeXt_bn_imagenet64 Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/resnext_bn_imagenet64.py
code served (permissive licence) · get_code("9436f818c378093b")
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 ResNeXt. 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 = {
    'in_ch': 3,
    'in_dim': 64,
}
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 ResNeXt — a ResNeXt output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNeXt 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
ResNeXt_imagenet64 Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/resnext_imagenet64.py
code served (permissive licence) · get_code("1a8ed1fd88054090")
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 ResNeXt. 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)

# The function ResNeXt_imagenet64 takes two parameters: in_ch and in_dim
# Looking at the function body, it doesn't actually use in_ch or in_dim
# It just ignores them and creates a ResNeXt with fixed hyperparameters
# So we can pass any integer values

ARGS = {
    "in_ch": 3,
    "in_dim": 64,
}
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 ResNeXt — a ResNeXt output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNeXt 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
conv3x3 Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/wide_resnet.py
pointer only (licence: BSD-3-Clause); the stored copy get_code serves, from DaikiTanak/manifold_mixup, is pointer only (licence: NONE) · get_code("00e569acd6b45ef0")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-3-Clause), but get_code serves the stored copy, first harvested from DaikiTanak/manifold_mixup, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_crown_loss Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/certified.py
code served (permissive licence) · get_code("7a95e6aff8614eaf")
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)

# Define a minimal args class with the required attributes
class Args:
    def __init__(self):
        self.ls = 0.1  # label smoothing parameter
        
# Create the args object
args = Args()

# Create test tensors
batch_size = 2
num_classes = 4

# lb should be a tensor of shape (batch_size, num_classes - 1)
# since it gets concatenated with zeros to form logits
lb = torch.randn(batch_size, num_classes - 1)

# eps for the threshold calculation
eps = 0.05

# Define ce_loss as a simple wrapper (this is used in the function)
def ce_loss(input, target):
    """Cross entropy loss"""
    return nn.functional.cross_entropy(input, target)

ARGS = {
    'args': args,
    'lb': lb,
    'eps': eps
}
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
get_params Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/manual_init.py
code served (permissive licence) · get_code("caeec8b2bd7730d3")
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
import torch.nn as nn

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
        self.bias_param = nn.Parameter(torch.randn(2))
        self.other_param = nn.Parameter(torch.randn(5))
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        return x

model = SimpleModel()

ARGS = {
    'model': model
}
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
resnext Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/resnext.py
code served (permissive licence) · get_code("3db74ea51c141a2d")
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 ResNeXt. 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 = {
    "in_ch": 3,
    "in_dim": 32,
}
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 ResNeXt — a ResNeXt output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNeXt 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
resnext Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/resnext_bn.py
code served (permissive licence) · get_code("14eb1c8929b00788")
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 ResNeXt. 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 Flatten(nn.Module):
    def forward(self, x):
        return x.view(x.size(0), -1)

ARGS = {
    "in_ch": 3,
    "in_dim": 32,
}
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 ResNeXt — a ResNeXt output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ResNeXt 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
update_dict Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/config.py
code served (permissive licence) · get_code("364eb5f6eebd2d6d")
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)

# Construct valid arguments for update_dict function
# The function takes:
# - d: dict to be updated
# - u: dict with updates
# - show_warning: bool flag for warnings

# Create a base dictionary with nested structure
base_dict = {
    'learning_rate': 0.001,
    'batch_size': 32,
    'model': {
        'hidden_dim': 128,
        'num_layers': 2
    },
    'training': {
        'epochs': 10,
        'optimizer': 'adam'
    }
}

# Create an update dictionary with some overlapping and new keys
update_dict_arg = {
    'learning_rate': 0.002,
    'batch_size': 64,
    'model': {
        'hidden_dim': 256,
        'dropout': 0.5
    },
    'new_key': 'new_value'
}

# Construct ARGS dictionary with all parameters
ARGS = {
    'd': base_dict,
    'u': update_dict_arg,
    'show_warning': True
}
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
wide_resnet_cifar_10 Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/wide_resnet.py
code served (permissive licence) · get_code("5778bd34b859bc9e")
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 Wide_ResNet. 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 = {
    'in_ch': 3,
    'in_dim': 32,
}
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 Wide_ResNet — a Wide_ResNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Wide_ResNet 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
wide_resnet_cifar_bn Ran
ran; our contract draft was wrong, not the code
shizhouxing/Fast-Certified-Robust-Training/models/wide_resnet.py
code served (permissive licence) · get_code("0a1ce4a266e25e0d")
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 Wide_ResNet. 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 = {
    'in_ch': 3,
    'in_dim': 32,
}
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 Wide_ResNet — a Wide_ResNet output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Wide_ResNet 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
cnn_7layer Not yet run
unverified (harvested, no recorded run)
shizhouxing/Fast-Certified-Robust-Training/models/feedforward.py
code served (permissive licence) · get_code("d2cf808aa64b27e8")
How it ran

no invocation recorded.

cnn_7layer_bn Not yet run
unverified (harvested, no recorded run)
shizhouxing/Fast-Certified-Robust-Training/models/feedforward.py
code served (permissive licence) · get_code("b2280f08abb758fd")
How it ran

no invocation recorded.

cnn_7layer_bn2 Not yet run
unverified (harvested, no recorded run)
shizhouxing/Fast-Certified-Robust-Training/models/feedforward.py
code served (permissive licence) · get_code("e37ed79862d93970")
How it ran

no invocation recorded.

get_C Not yet run
unverified (harvested, no recorded run)
shizhouxing/Fast-Certified-Robust-Training/certified.py
code served (permissive licence) · get_code("5164505350577eaa")
How it ran

no invocation recorded.

load_data Not yet run
unverified (harvested, no recorded run)
shizhouxing/Fast-Certified-Robust-Training/datasets.py
code served (permissive licence) · get_code("af20db67ef9b7df2")
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, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 2 later) and 16 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
conv3x3 74 other papers
22 dated earlier, 1 the same month, 51 later · 5 only as an assembly brick
1605.07146 (2016-05)
Wide Residual Networks
1707.01629 (2017-07)
Dual Path Networks
1802.05668 (2018-02)
Model compression via distillation and quantization
and 71 more not named here
get_params 1 other paper
1 later
2210.04871 (2022-10)
Certified Training: Small Boxes are All You Need

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

Recently, bound propagation based certified robust training methods have been proposed for training neural networks with certifiable robustness guarantees. Despite that state-of-the-art (SOTA) methods including interval bound propagation (IBP) and CROWN-IBP have per-batch training complexity similar to standard neural network training, they usually use a long warmup schedule with hundreds or thousands epochs to reach SOTA performance and are thus still costly. In this paper, we identify two important issues in existing methods, namely exploded bounds at initialization, and the imbalance in ReLU activation states and improve IBP training. These two issues make certified training difficult and unstable, and thereby long warmup schedules were needed in prior works. To mitigate these issues and conduct faster certified training with shorter warmup, we propose three improvements based on IBP training: 1) We derive a new weight initialization method for IBP training; 2) We propose to fully add Batch Normalization (BN) to each layer in the model, since we find BN can reduce the imbalance in ReLU activation states; 3) We also design regularization to explicitly tighten certified bounds and balance ReLU activation states during wamrup. We are able to obtain 65.03% verified error on CIFAR-10 ( = 8 255 ) and 82.36% verified error on Tiny-ImageNet ( = 1 255 ) using very short training schedules (160 and 80 total epochs, respectively), outperforming literature SOTA trained with hundreds or thousands epochs under the same network architecture. The code is available at https: //github.com/shizhouxing/Fast-Certified-Robust-Training.

For agents

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

get_harvested_code_for_paper("2103.17268")
get_code_for_paper("2103.17268")
have("2103.17268")

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

curl https://syntology.ai/api/ran/2103.17268.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.17268.svg)](https://syntology.ai/paper/2103.17268)

Connect an agent — have() is free.