SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1802.09419 · 2018

Stochastic Hyperparameter Optimization through Hypernetworks

arXiv · PDF · Open in the Atlas

Code that ran

11 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; 1 honoured the contract we drafted, 0 violated it and 10 have no contract verdict; 0 failures of our instrument; 0 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/1802.09419.json · badge: /api/ran/1802.09419.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
lorraine2/hypernet-hypertraining community (classifier label)
classifier basis not recorded on the link
11 ran · 0 no recorded run
FunctionStatusWhere it lives
adam Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/optimizers.py
code served (permissive licence) · get_code("ee2f9c951a3f0a07")
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 autograd.numpy as np
from autograd.misc.flatten import flatten_func

def simple_grad_fn(params, iter_num):
    """A simple gradient function that returns gradients for the parameters."""
    # Returns gradients with same shape as params
    return np.array([0.1, 0.05, 0.02, 0.01])

def simple_callback(params, iter_num, grads):
    """A simple callback function."""
    pass

# Create initial parameters
init_params = np.array([1.0, 2.0, 3.0, 4.0])

ARGS = {
    "grad": simple_grad_fn,
    "init_params": init_params,
    "callback": simple_callback,
    "num_iters": 5,
    "step_size": 0.001,
    "b1": 0.9,
    "b2": 0.999,
    "eps": 1e-8,
    "m": None,
    "v": None,
    "offset": None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
autograd — 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
best_response Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/loss_manifold.py
code served (permissive licence) · get_code("69d2f57c7e093831")
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

# The function takes a hyperparameter y and returns the best-response weights
# y should be a numeric value (float or numpy scalar)
# The function applies sin(y + pi) to it

y = np.array(0.5)  # A simple hyperparameter value

ARGS = {
    "y": y
}
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
create_figure_and_axs Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/plotting.py
code served (permissive licence) · get_code("9c2a23e40c56bf52")
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 matplotlib
import matplotlib.pyplot as plt

ARGS = {
    'fig_width': 10,
    'fig_height': 10,
    'num_rows': 2,
    'num_cols': 3,
    'ms_size': 28
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib — 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
full_extent Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/plotting.py
code served (permissive licence) · get_code("f4b5a0d4c2bb9340")
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 Bbox. 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 matplotlib.pyplot as plt
import matplotlib.patches as mpatches
from matplotlib.transforms import Bbox

# Create a figure and axis
fig, ax = plt.subplots(figsize=(8, 6))

# Plot some data to make the axis have content
ax.plot([1, 2, 3], [1, 4, 2], label='test')
ax.set_xlabel('X Label')
ax.set_ylabel('Y Label')
ax.set_title('Test Title')

# Draw the canvas so that window extents are calculated
fig.canvas.draw()

# Define arguments for full_extent function
ARGS = {
    'ax': ax,
    'pad': 0.1,
    'do_xticks': True,
    'do_yticks': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Bbox — a Bbox output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Bbox output the recorded ndim 2 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
identity Ran
ran, honoured the contract we drafted
lorraine2/hypernet-hypertraining/code/neural_network.py
code served (permissive licence) · get_code("efd23da01a3364b1")
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 identity function simply returns its input unchanged
# So we need to provide a valid input tensor
x = torch.randn(2, 4)

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
init_random_params Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/neural_network.py
code served (permissive licence) · get_code("6c7099b9362c9e4e")
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. 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

# Create a numpy random state
rs = np.random.RandomState(42)

# Define parameters
var = 0.5  # positive float for variance
layer_sizes = [4, 3, 2, 1]  # layer sizes: input 4, hidden 3, hidden 2, output 1

ARGS = {
    "var": var,
    "layer_sizes": layer_sizes,
    "rs": rs
}
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 — a list 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
log_gaussian Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/learn_vs_true_loss_scatter.py
code served (permissive licence) · get_code("33654d7c183a991c")
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 autograd.numpy as np

# Construct weights: a nested list structure where weights[0][0] is a 2D array
# The function accesses weights[0][0][:, i] for i in range(10), so we need
# at least 10 columns. Rows can be any number (here using 5).
weights_array = np.random.randn(5, 10)
weights = [[weights_array]]

# Construct var: a 1D array with 10 elements (one for each column index)
var = np.random.randn(10)

ARGS = {
    "weights": weights,
    "var": var
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
autograd — 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
opt_params Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/optimizers.py
code served (permissive licence) · get_code("e037e68de30ee470")
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

torch.manual_seed(0)

# The function opt_params takes a single parameter: graph_iters
# which should be an integer > 0 representing the number of iterations
# we can potentially graph for allocating memory

ARGS = {
    'graph_iters': 100
}
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
orthogonal_proj Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/loss_manifold.py
code served (permissive licence) · get_code("76060568553562d9")
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 ndarray, ndim 2, dtype family float, shape (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 numpy as np

ARGS = {
    "zfront": 1.0,
    "zback": 0.1,
}
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 ndarray, ndim 2, dtype family float, shape (4, 4) — shape (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
relu Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/neural_network.py
code served (permissive licence) · get_code("57d85d9dbc47f39e")
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 ndarray, 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 autograd.numpy as np

# The relu function takes a single parameter x
# It applies np.maximum(x, 0) to compute the ReLU activation
# x should be a numeric array-like object

# Create a small test input
x = np.array([[-2.0, -1.0, 0.0, 1.0],
              [2.0, 3.0, -0.5, 0.5]])

ARGS = {
    'x': x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
autograd — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, 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
train_loss_fun Ran
ran on a synthesized input
lorraine2/hypernet-hypertraining/code/loss_manifold.py
code served (permissive licence) · get_code("a3486434b5aed9e0")
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

# The function takes args which should be an array-like with two elements [x, y]
# x and y are used in mathematical operations (logaddexp, sin, subtraction, etc.)
# They should be numpy arrays or scalars

ARGS = {
    'args': np.array([2.5, 1.3])
}
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

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, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 2 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
identity 69 other papers
1 dated earlier, 68 later · 5 only as an assembly brick
1706.07296 (2017-06)
A Minimal Developmental Model Can Increase Evolvability in Soft Robots
1804.06055 (2018-04)
Co-occurrence Feature Learning from Skeleton Data for Action Recognition and Detection with Hierarchical Aggregation
1808.09357 (2018-08)
Rational Recurrences
and 66 more not named here
relu 3 other papers
1 dated earlier, 2 later
1610.02391 (2016-10)
Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization
1811.00145 (2018-11)
Scalable End-to-End Autonomous Vehicle Testing via Rare-event Simulation
2102.10058 (2021-02)
Principled Simplicial Neural Networks for Trajectory Prediction

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

Machine learning models are often tuned by nesting optimization of model weights inside the optimization of hyperparameters. We give a method to collapse this nested optimization into joint stochastic optimization of weights and hyperparameters. Our process trains a neural network to output approximately optimal weights as a function of hyperparameters. We show that our technique converges to locally optimal weights and hyperparameters for sufficiently large hypernetworks. We compare this method to standard hyperparameter optimization strategies and demonstrate its effectiveness for tuning thousands of hyperparameters.

For agents

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

get_harvested_code_for_paper("1802.09419")
get_code_for_paper("1802.09419")
have("1802.09419")

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

curl https://syntology.ai/api/ran/1802.09419.json

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

[![Syntology run record](https://syntology.ai/api/ran/1802.09419.svg)](https://syntology.ai/paper/1802.09419)

Connect an agent — have() is free.