SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2106.04015 · 2021

Uncertainty Baselines: Benchmarks for Uncertainty & Robustness in Deep Learning

arXiv · PDF · Open in the Atlas

Code that ran

9 of 11 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; 2 honoured the contract we drafted, 0 violated it and 4 have no contract verdict; 3 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/2106.04015.json · badge: /api/ran/2106.04015.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
thlarsen13/thlarsen_uncertainty_models community (classifier label)
classifier basis not recorded on the link
9 ran · 2 no recorded run
FunctionStatusWhere it lives
conv_kernel_initializer Ran
ran; our contract draft was wrong, not the code
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/efficientnet_utils.py
code served (permissive licence) · get_code("eff25194ff1a4c16")
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 SymbolicTensor. 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 numpy as np
import tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

# The function expects:
# - shape: a tuple representing (kernel_height, kernel_width, in_channels, out_filters)
# - dtype: optional dtype (can be None or a tf dtype)
# - partition_info: unused parameter

ARGS = {
    "shape": (3, 4, 5, 2),
    "dtype": tf.float32,
    "partition_info": None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
create_config Ran
ran; our contract draft was wrong, not the code
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/bert.py
code served (permissive licence) · get_code("295ee5b09ae2c97c")
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 json
import tempfile
import os
from typing import Any, Dict, Tuple, List

# Create a temporary directory and config file for testing
temp_dir = tempfile.mkdtemp()
config_file_path = os.path.join(temp_dir, "config.json")

# Create a sample BERT config dictionary
sample_config = {
    "attention_probs_dropout_prob": 0.1,
    "directionality": "bidi",
    "hidden_act": "gelu",
    "hidden_dropout_prob": 0.1,
    "hidden_size": 768,
    "initializer_range": 0.02,
    "intermediate_size": 3072,
    "layer_norm_eps": 1e-12,
    "max_position_embeddings": 512,
    "model_type": "bert",
    "num_attention_heads": 12,
    "num_hidden_layers": 12,
    "output_hidden_states": False,
    "output_attentions": False,
    "pad_token_id": 0,
    "position_embedding_type": "absolute",
    "torch_dtype": "float32",
    "transformers_version": "4.0.0",
    "type_vocab_size": 2,
    "use_cache": True,
    "vocab_size": 30522
}

# Write the config to the temporary file
with open(config_file_path, "w") as f:
    json.dump(sample_config, f)

# Define arguments for the function
ARGS = {
    "config_dir": config_file_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
create_feature_and_label Ran
ran; our contract draft was wrong, not the code
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/bert.py
code served (permissive licence) · get_code("40ffce7d1720bb3e")
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 tensorflow as tf

ARGS = {
    'inputs': {
        'features': tf.constant([[101, 2054, 2003, 102], [101, 1045, 2572, 102]], dtype=tf.int32),
        'labels': tf.constant([0, 1], dtype=tf.int32),
        'num_tokens': tf.constant([4, 3], dtype=tf.int32),
    },
    'max_seq_length': 4,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
drop_connect Ran
ran on a synthesized input
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/efficientnet_utils.py
code served (permissive licence) · get_code("48ffc519251345a8")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create a small input tensor with shape [batch_size, height, width, channels]
# Using batch_size=2, height=4, width=3, channels=5 for distinct dimensions
inputs = tf.constant(np.random.randn(2, 4, 3, 5).astype(np.float32))

# is_training: boolean flag
is_training = True

# survival_prob: probability that a unit is kept (typically 0.8)
survival_prob = 0.8

ARGS = {
    "inputs": inputs,
    "is_training": is_training,
    "survival_prob": survival_prob,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
efficientnet_params Ran
ran, honoured the contract we drafted
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/efficientnet.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ravi02512/efficientdet-keras, is pointer only (licence: NONE) · get_code("9e354d875b0e94e6")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from ravi02512/efficientdet-keras, 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.

efficientnet_params Ran
ran on a synthesized input
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/efficientnet_utils.py
code served (permissive licence) · get_code("bc1878c9e235eca0")
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, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 takes a model_name string and returns a tuple of parameters
# Looking at the function body, it simply looks up the model_name in a dictionary
# and returns the corresponding tuple. No complex operations needed.

ARGS = {
    'model_name': 'efficientnet-b0'
}
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, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
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_adjacency_matrix Ran
ran on a synthesized input
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/mpnn.py
code served (permissive licence) · get_code("7b2943bee6656eca")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EagerTensor, ndim 3, dtype family float, shape (2, 4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tensorflow as tf
import numpy as np

# Create a small pairs tensor with shape [batch_size, max_nodes, max_nodes, num_pair_features]
# Using distinct sizes: batch_size=2, max_nodes=4, num_pair_features=5
batch_size = 2
max_nodes = 4
num_pair_features = 5

# Create random edge features tensor
pairs_data = np.random.rand(batch_size, max_nodes, max_nodes, num_pair_features).astype(np.float32)
pairs = tf.constant(pairs_data)

ARGS = {
    "pairs": pairs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 4, 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
round_filters Ran
ran on a synthesized input
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/efficientnet.py
code served (permissive licence) · get_code("ec0ebc77bce35a4a")
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)

# Analyze the function:
# - filters: numeric value (int or float) that gets multiplied
# - width_coefficient: numeric multiplier
# - depth_divisor: numeric divisor used for rounding
# - min_depth: numeric minimum depth value

ARGS = {
    "filters": 32,
    "width_coefficient": 1.0,
    "depth_divisor": 8,
    "min_depth": 1
}
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
round_repeats Ran
ran, honoured the contract we drafted
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/efficientnet.py
code served (permissive licence) · get_code("e316074dac020280")
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 math

def round_repeats(repeats, depth_coefficient):
    """Round number of filters based on depth multiplier."""
    return int(math.ceil(depth_coefficient * repeats))

ARGS = {
    "repeats": 4,
    "depth_coefficient": 1.2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
genomics_cnn Not yet run
unverified (harvested, no recorded run)
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/genomics_cnn.py
code served (permissive licence) · get_code("4979d2806a89ebe3")
How it ran

no invocation recorded.

movielens Not yet run
unverified (harvested, no recorded run)
thlarsen13/thlarsen_uncertainty_models/uncertainty_baselines/models/movielens.py
code served (permissive licence) · get_code("38f23256728e342b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 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, 4 later) and 6 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
round_repeats 4 other papers
3 dated earlier, 1 later
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
1911.09070 (2019-11)
EfficientDet: Scalable and Efficient Object Detection
1911.12448 (2019-11)
Soft Anchor-Point Object Detection
and 1 more not named here
efficientnet_params 3 other papers
1 dated earlier, 2 later
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
2107.02612 (2021-07)
Combining EfficientNet and Vision Transformers for Video Deepfake Detection
2403.01786 (2024-03)
Exposing the Deception: Uncovering More Forgery Clues for Deepfake Detection
conv_kernel_initializer 1 other paper
1 dated earlier
1911.04252 (2019-11)
Self-training with Noisy Student improves ImageNet classification
create_config 1 other paper
1 later
2606.13818 (2026-06)
PAC-Chernoff Bounds: Understanding Generalization in the Interpolation Regime
create_feature_and_label 1 other paper
1 later
2606.13818 (2026-06)
PAC-Chernoff Bounds: Understanding Generalization in the Interpolation Regime

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

High-quality estimates of uncertainty and robustness are crucial for numerous real-world applications, especially for deep learning which underlies many deployed ML systems. The ability to compare techniques for improving these estimates is therefore very important for research and practice alike. Yet, competitive comparisons of methods are often lacking due to a range of reasons, including: compute availability for extensive tuning, incorporation of sufficiently many baselines, and concrete documentation for reproducibility. In this paper we introduce Uncertainty Baselines: high-quality implementations of standard and state-of-the-art deep learning methods on a variety of tasks. As of this writing, the collection spans 19 methods across 9 tasks, each with at least 5 metrics. Each baseline is a self-contained experiment pipeline with easily reusable and extendable components. Our goal is to provide immediate starting points for experimentation with new methods or applications. Additionally we provide model checkpoints, experiment outputs as Python notebooks, and leaderboards for comparing results. Code available at https://github.com/google/uncertainty-baselines.

For agents

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

get_harvested_code_for_paper("2106.04015")
get_code_for_paper("2106.04015")
have("2106.04015")

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

curl https://syntology.ai/api/ran/2106.04015.json

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

[![Syntology run record](https://syntology.ai/api/ran/2106.04015.svg)](https://syntology.ai/paper/2106.04015)

Connect an agent — have() is free.