SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.01970 · EMNLP · 2022

Evaluate & Evaluation on the Hub: Better Best Practices for Data and Model Measurements

Alexandra Luccioni, Douwe Kiela, Aleksandra Piktus, Alexander Rush, Nazneen Rajani, Margaret Mitchell, Quentin Lhoest, Thomas Wolf, Tristan Thrush, Leandro Von Werra, Mario Šaško, Lewis Tunstall, and 7 more

arXiv · PDF · Open in the Atlas

Code that ran

5 of 9 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 5 have no contract verdict; 0 failures of our instrument; 4 no recorded run), measured 2026-11-22; 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/2210.01970.json · badge: /api/ran/2210.01970.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
huggingface/evaluate canonical (classifier label)
classifier basis not recorded on the link
5 ran · 4 no recorded run
FunctionStatusWhere it lives
camelcase_to_snakecase Ran
ran on a synthesized input
huggingface/evaluate/src/evaluate/naming.py
code served (permissive licence) · get_code("a4f4bd8c6eadcade")
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 str. 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 re

_uppercase_uppercase_re = re.compile('([A-Z]+)([A-Z][a-z])')

_lowercase_uppercase_re = re.compile('([a-z\\d])([A-Z])')

def camelcase_to_snakecase(name):
    """Convert camel-case string to snake-case."""
    name = _uppercase_uppercase_re.sub('\\1_\\2', name)
    name = _lowercase_uppercase_re.sub('\\1_\\2', name)
    return name.lower()

ARGS = {
    "name": "camelCaseExample"
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
filename_prefix_for_name Ran
ran on a synthesized input
huggingface/evaluate/src/evaluate/naming.py
code served (permissive licence) · get_code("7b8d8799c8657b8d")
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 str. 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 os
import re

_uppercase_uppercase_re = re.compile('([A-Z]+)([A-Z][a-z])')
_lowercase_uppercase_re = re.compile('([a-z\\d])([A-Z])')

def camelcase_to_snakecase(name):
    """Convert camel-case string to snake-case."""
    name = _uppercase_uppercase_re.sub('\\1_\\2', name)
    name = _lowercase_uppercase_re.sub('\\1_\\2', name)
    return name.lower()

def filename_prefix_for_name(name):
    if os.path.basename(name) != name:
        raise ValueError(f'Should be a dataset name, not a path: {name}')
    return camelcase_to_snakecase(name)

ARGS = {
    'name': 'MyDatasetName'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
files_to_hash Ran
ran on a synthesized input
huggingface/evaluate/src/evaluate/loading.py
code served (permissive licence) · get_code("932c565562fa4501")
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 str. 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 os
import tempfile
from pathlib import Path
from typing import List

# Create temporary directory with test Python files
temp_dir = tempfile.mkdtemp()

# Create a test Python file
test_file_1 = os.path.join(temp_dir, "script1.py")
with open(test_file_1, 'w', encoding='utf-8') as f:
    f.write("# Test script 1\n")
    f.write("def hello():\n")
    f.write("    pass\n")

# Create another test Python file
test_file_2 = os.path.join(temp_dir, "script2.py")
with open(test_file_2, 'w', encoding='utf-8') as f:
    f.write("# Test script 2\n")
    f.write("def world():\n")
    f.write("    return 42\n")

# Create a subdirectory with a Python file
sub_dir = os.path.join(temp_dir, "subdir")
os.makedirs(sub_dir, exist_ok=True)
test_file_3 = os.path.join(sub_dir, "nested.py")
with open(test_file_3, 'w', encoding='utf-8') as f:
    f.write("# Nested script\n")
    f.write("x = 1\n")

ARGS = {
    "file_paths": [test_file_1, test_file_2]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
datasets — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
snakecase_to_camelcase Ran
ran on a synthesized input
huggingface/evaluate/src/evaluate/naming.py
code served (permissive licence) · get_code("5416ab50555cbdeb")
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 str. 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 itertools
import re

_single_underscore_re = re.compile('(?<!_)_(?!_)')
_multiple_underscores_re = re.compile('(_{2,})')

def snakecase_to_camelcase(name):
    """Convert snake-case string to camel-case string."""
    name = _single_underscore_re.split(name)
    name = [_multiple_underscores_re.split(n) for n in name]
    return ''.join((n.capitalize() for n in itertools.chain.from_iterable(name) if n != ''))

ARGS = {
    'name': 'hello_world_example'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
summarize_if_long_list Ran
ran on a synthesized input
huggingface/evaluate/src/evaluate/module.py
code served (permissive licence) · get_code("0ef8bf8d8d3546da")
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 str. 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)

# Test case 1: Short list (should return as-is)
short_list = [1, 2, 3]

# Test case 2: Long list (should be summarized)
long_list = list(range(20))

# Test case 3: Empty list
empty_list = []

# Test case 4: List with exactly 6 elements (boundary case, should return as-is)
boundary_list = [10, 20, 30, 40, 50, 60]

# Test case 5: List with 7 elements (should be summarized)
long_list_7 = [1, 2, 3, 4, 5, 6, 7]

# Test case 6: Non-list object (should return string representation)
non_list_obj = "hello"

# The function checks `type(obj) is not list or len(obj) <= 6`
# So we need to test both list and non-list cases
ARGS = {
    'obj': long_list
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
combine Not yet run
unverified (harvested, no recorded run)
huggingface/evaluate/src/evaluate/module.py
code served (permissive licence) · get_code("340b2e2c769b120d")
How it ran

no invocation recorded.

list_evaluation_modules Not yet run
unverified (harvested, no recorded run)
huggingface/evaluate/src/evaluate/inspect.py
code served (permissive licence) · get_code("786671ac6dc0003d")
How it ran

no invocation recorded.

radar_plot Not yet run
unverified (harvested, no recorded run)
huggingface/evaluate/src/evaluate/visualization.py
code served (permissive licence) · get_code("cae2ae1c6da8e3ad")
How it ran

no invocation recorded.

save Not yet run
unverified (harvested, no recorded run)
huggingface/evaluate/src/evaluate/saving.py
code served (permissive licence) · get_code("e89abecf63da16c6")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 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.

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

Evaluation is a key part of machine learning (ML), yet there is a lack of support and tooling to enable its informed and systematic practice. We introduce Evaluate and Evaluation on the Hub-a set of tools to facilitate the evaluation of models and datasets in ML. Evaluate is a library to support best practices for measurements, metrics, and comparisons of data and models. Its goal is to support reproducibility of evaluation, centralize and document the evaluation process, and broaden evaluation to cover more facets of model performance. It includes over 50 efficient canonical implementations for a variety of domains and scenarios, interactive documentation, and the ability to easily share implementations and outcomes. The library is available at https://github.com/huggingface/evaluate. In addition, we introduce Evaluation on the Hub, a platform that enables the large-scale evaluation of over 75,000 models and 11,000 datasets on the Hugging Face Hub, for free, at the click of a button. Evaluation on the Hub is available at https://huggingface.co/autoevaluate. Demo screencast: youtu.be/6rU177zRj8Q * Equal contribution.

For agents

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

get_harvested_code_for_paper("2210.01970")
get_code_for_paper("2210.01970")
have("2210.01970")

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

curl https://syntology.ai/api/ran/2210.01970.json

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

[![Syntology run record](https://syntology.ai/api/ran/2210.01970.svg)](https://syntology.ai/paper/2210.01970)

Connect an agent — have() is free.