SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2404.15523 · 2024

Understanding Hyperbolic Metric Learning through Hard Negative Sampling

arXiv · PDF · Open in the Atlas

Code that ran

10 of 12 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, 1 violated it and 0 have no contract verdict; 9 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/2404.15523.json · badge: /api/ran/2404.15523.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
yunyuny/hypmix canonical (classifier label)
classifier basis: census
10 ran · 2 no recorded run
FunctionStatusWhere it lives
arsinh Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/hyptorch/pmath.py
code served (permissive licence) · get_code("127282b695693958")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (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

torch.manual_seed(0)

# The function `arsinh` takes a single parameter `x` which should be a tensor
# The function applies the Arsinh custom autograd function to it
# Looking at the forward method, x needs to support:
# - torch.sqrt_() 
# - pow(2)
# - clamp_min_()
# - log_()
# These are all standard tensor operations, so x should be a regular tensor

# Create a small tensor for testing
x = torch.randn(2, 4, dtype=torch.float32)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 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
artanh Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/hyptorch/pmath.py
code served (permissive licence) · get_code("daf212010121f13d")
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

torch.manual_seed(0)

# The artanh function takes a single parameter x which should be a tensor
# The function clamps x to the range [-1 + 1e-05, 1 - 1e-05]
# and computes the inverse hyperbolic tangent

# Create a small tensor with values in the valid range for artanh
x = torch.randn(2, 4) * 0.5  # Scale down to keep values in reasonable range for artanh

ARGS = {
    "x": x
}
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 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
batched_delta_hyp Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/hyptorch/delta.py
code served (permissive licence) · get_code("5bba1b067e614c37")
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 numpy as np
from scipy.spatial import distance_matrix

np.random.seed(0)

# Create a small point cloud in 4D space
# The function uses X with len(X) to sample indices
# and passes X[idx] to distance_matrix, so X should be a 2D array
X = np.random.randn(50, 4).astype(np.float64)

ARGS = {
    "X": X,
    "n_tries": 2,
    "batch_size": 10,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, tqdm — 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
binarize Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/proxy_anchor/losses.py
code served (permissive licence) · get_code("9501bf152737299c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 numpy as np

torch.manual_seed(0)

# T should be a tensor containing class labels (integers from 0 to nb_classes-1)
# The function will convert it to numpy, then binarize it
T = torch.tensor([0, 1, 2, 1, 0, 2], dtype=torch.long)

# nb_classes should be an integer specifying the number of classes
nb_classes = 3

ARGS = {
    "T": T,
    "nb_classes": nb_classes,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, 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 — no output shape is stated: observe_signature.py could not observe it (call:AssertionError)
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
delta_hyp Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/delta.py
code served (permissive licence) · get_code("d52cb5aa99ad1ebe")
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

torch.manual_seed(0)

# Create a small distance matrix
# The function indexes it with [p, :], [:, p], and performs operations
# A square symmetric matrix makes sense for a distance matrix
# Using shape (4, 4) for a small valid distance matrix
dismat = torch.tensor([
    [0.0, 1.0, 2.0, 3.0],
    [1.0, 0.0, 1.5, 2.5],
    [2.0, 1.5, 0.0, 1.0],
    [3.0, 2.5, 1.0, 0.0]
], dtype=torch.float32)

ARGS = {
    "dismat": dismat
}
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
delta_hyp Ran
ran, violated the contract we drafted
yunyuny/hypmix/hyptorch/delta.py
code served (permissive licence) · get_code("f3a60764ea302fb1")
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 float64, ndim 0, 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 numpy as np

# Create a small distance matrix
# The function expects a square distance matrix where dismat[i,j] represents
# the distance between points i and j
# Using a 5x5 matrix to have enough data for the computation
dismat = np.array([
    [0.0, 1.0, 2.0, 3.0, 4.0],
    [1.0, 0.0, 1.5, 2.5, 3.5],
    [2.0, 1.5, 0.0, 1.0, 2.0],
    [3.0, 2.5, 1.0, 0.0, 1.5],
    [4.0, 3.5, 2.0, 1.5, 0.0]
], dtype=np.float64)

ARGS = {
    "dismat": dismat
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
l2_norm Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/proxy_anchor/losses.py
code served (permissive licence) · get_code("45837478fdb29f78")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (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

torch.manual_seed(0)

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

ARGS = {
    "input": input_tensor
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (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
safe_random_choice Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yunyuny/hypmix/sampler.py
code served (permissive licence) · get_code("7686967dd97570c7")
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 list, ndim 1, dtype family int; 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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Construct arguments for safe_random_choice
input_data = np.array([1, 2, 3, 4, 5, 6, 7, 8, 9, 10])
size = 5

ARGS = {
    "input_data": input_data,
    "size": size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family int — a list 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
tanh Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/hyptorch/pmath.py
code served (permissive licence) · get_code("be5af45dd86c2cc1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (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 torch.nn as nn

torch.manual_seed(0)

# The function takes x (a tensor) and clamp (an int)
# It calls x.clamp(-clamp, clamp).tanh()
# So x must be a tensor with clamp and tanh methods

x = torch.randn(2, 4)
clamp_value = 15

ARGS = {
    "x": x,
    "clamp": clamp_value
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (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
uniform_loss Ran
ran; our contract draft was wrong, not the code
yunyuny/hypmix/helpers.py
code served (permissive licence) · get_code("8d3d5f756818e024")
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

torch.manual_seed(0)

# x should be a 2D tensor for pdist to compute pairwise distances
# keeping it small: 2 samples, 4 dimensions
x = torch.randn(2, 4)

# t is a scalar parameter (default 2)
t = 2

ARGS = {
    "x": x,
    "t": t,
}
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_delta Not yet run
unverified (harvested, no recorded run)
yunyuny/hypmix/hyptorch/delta.py
code served (permissive licence) · get_code("cf1928988047f33d")
How it ran

no invocation recorded.

get_labels_to_indices Not yet run
unverified (harvested, no recorded run)
yunyuny/hypmix/sampler.py
code served (permissive licence) · get_code("e66fff5859babb9f")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 8 are also present in other papers whose code we hold (8 with at least one paper dated earlier, 0 the same month, 3 later) and 4 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
artanh 5 other papers
5 dated earlier · 1 only as an assembly brick
1904.02239 (2019-04)
Hyperbolic Image Embeddings
2005.00545 (2020-05)
Low-Dimensional Hyperbolic Knowledge Graph Embeddings
2108.02390 (2021-08)
Fuzzy Logic Based Logical Query Answering on Knowledge Graphs
and 2 more not named here
l2_norm 4 other papers
2 dated earlier, 2 later · 1 only as an assembly brick
2210.05557 (2022-10)
OPERA: Omni-Supervised Representation Learning with Hierarchical Supervisions
2307.10943 (2023-07)
Proxy Anchor-based Unsupervised Learning for Continuous Generalized Category Discovery
2408.13492 (2024-08)
Online Continuous Generalized Category Discovery
and 1 more not named here
tanh 4 other papers
3 dated earlier, 1 later · 1 only as an assembly brick
1904.02239 (2019-04)
Hyperbolic Image Embeddings
2108.02390 (2021-08)
Fuzzy Logic Based Logical Query Answering on Knowledge Graphs
2203.10833 (2022-03)
Hyperbolic Vision Transformers: Combining Improvements in Metric Learning
and 1 more not named here
binarize 3 other papers
1 dated earlier, 2 later · 1 only as an assembly brick
2307.10943 (2023-07)
Proxy Anchor-based Unsupervised Learning for Continuous Generalized Category Discovery
2408.13492 (2024-08)
Online Continuous Generalized Category Discovery
2508.03256 (2025-08)
Beyond Isolated Words: Diffusion Brush for Handwritten Text-Line Generation
uniform_loss 3 other papers
3 dated earlier
2304.04704 (2023-04)
Prompt Pre-Training with Twenty-Thousand Classes for Open-Vocabulary Visual Recognition
2309.06086 (2023-09)
Plasticity-Optimized Complementary Networks for Unsupervised Continual Learning
2403.00642 (2024-03)
Rethinking The Uniformity Metric in Self-Supervised Learning
arsinh 1 other paper
1 dated earlier
1904.02239 (2019-04)
Hyperbolic Image Embeddings
delta_hyp 1 other paper
1 dated earlier
1904.02239 (2019-04)
Hyperbolic Image Embeddings
get_delta 1 other paper
1 dated earlier
1904.02239 (2019-04)
Hyperbolic Image Embeddings

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

In recent years, there has been a growing trend of incorporating hyperbolic geometry methods into computer vision. While these methods have achieved state-of-the-art performance on various metric learning tasks using hyperbolic distance measurements, the underlying theoretical analysis supporting this superior performance remains under-exploited. In this study, we investigate the effects of integrating hyperbolic space into metric learning, particularly when training with contrastive loss. We identify a need for a comprehensive comparison between Euclidean and hyperbolic spaces regarding the temperature effect in the contrastive loss within the existing literature. To address this gap, we conduct an extensive investigation to benchmark the results of Vision Transformers (ViTs) using a hybrid objective function that combines loss from Euclidean and hyperbolic spaces. Additionally, we provide a theoretical analysis of the observed performance improvement. We also reveal that hyperbolic metric learning is highly related to hard negative sampling, providing insights for future work. This work will provide valuable data points and experience in understanding hyperbolic image embeddings. To shed more light on problem-solving and encourage further investigation into our approach, our code is available online (https://github.com/YunYunY/HypMix).

For agents

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

get_harvested_code_for_paper("2404.15523")
get_code_for_paper("2404.15523")
have("2404.15523")

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

curl https://syntology.ai/api/ran/2404.15523.json

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

[![Syntology run record](https://syntology.ai/api/ran/2404.15523.svg)](https://syntology.ai/paper/2404.15523)

Connect an agent — have() is free.