SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2408.00677 · ECCV · 2024

Scaling Backwards: Minimal Synthetic Pre-training?

Yuki Asano, Christian Rupprecht, Iro Laina, Nakamasa Inoue, Rio Yokota, Hirokatsu Kataoka, Ryo Nakamura, Ryu Tadokoro, Ryosuke Yamada

arXiv · PDF · Open in the Atlas

Code that ran

3 of 5 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 0 have no contract verdict; 2 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/2408.00677.json · badge: /api/ran/2408.00677.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
SUPER-TADORY/1p-frac canonical (classifier label)
classifier basis: census
3 ran · 2 no recorded run
FunctionStatusWhere it lives
calc_p Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
SUPER-TADORY/1p-frac/1p-frac_generater/src/render_from_oneparam_mp.py
code served (permissive licence) · get_code("b291fa6a41b0e305")
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, 7) 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

# Create a valid params array with shape (batch_size, 2, 3)
# The function accesses params[:, 0, 0], [:, 0, 1], [:, 1, 0], [:, 1, 1], [:, 0, 2], [:, 1, 2]
# So we need at least 2 rows, 2 columns, 3 depth
# batch_size = 2, rows = 2, cols = 3

params = np.array([
    [[1.0, 2.0, 3.0],
     [4.0, 5.0, 6.0]],
    [[7.0, 8.0, 9.0],
     [10.0, 11.0, 12.0]]
], dtype=np.float64)

ARGS = {
    "params": params
}
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 (2, 7) — shape (2, 7) 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
reshape Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
SUPER-TADORY/1p-frac/1p-frac_generater/src/render_from_oneparam_mp.py
code served (permissive licence) · get_code("6b023b60d9e509b4")
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, 6) 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

# The function expects params to be a numpy array with shape (batch_size, 2, 3)
# It indexes into dimensions 0 (batch), 1 (rows 0-1), and 2 (cols 0-2)
# The output will have shape (batch_size, 6)

# Create a small batch of parameters with shape (2, 2, 3)
# batch_size=2, 2 rows, 3 columns
params = np.array([
    [[1.0, 2.0, 3.0],
     [4.0, 5.0, 6.0]],
    [[7.0, 8.0, 9.0],
     [10.0, 11.0, 12.0]]
])

ARGS = {
    "params": params
}
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 (2, 6) — shape (2, 6) 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
sample_svs Ran
ran, honoured the contract we drafted
SUPER-TADORY/1p-frac/1p-frac_generater/src/utils.py
code served (permissive licence) · get_code("a1c6321806632355")
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 (5, 2) 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

# Set random seed for reproducibility
np.random.seed(42)

# Define arguments for sample_svs function
n = 5  # number of pairs of singular values
a = 10.0  # constraint on weighted sum, must be in range (0, 3*n) = (0, 15)
rng = np.random.default_rng(seed=42)  # random number generator

ARGS = {
    "n": n,
    "a": a,
    "rng": rng,
}
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 (5, 2) — shape (5, 2) 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
create_dataloader Not yet run
unverified (harvested, no recorded run)
SUPER-TADORY/1p-frac/train/data.py
code served (permissive licence) · get_code("66801f43fe5c3707")
How it ran

no invocation recorded.

sample_system Not yet run
unverified (harvested, no recorded run)
SUPER-TADORY/1p-frac/1p-frac_generater/src/utils.py
code served (permissive licence) · get_code("3a41df871027ed4c")
How it ran

no invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 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
sample_svs 1 other paper
1 dated earlier
2110.03091 (2021-10)
Improving Fractal Pre-training

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

Pre-training and transfer learning are an important building block of current computer vision systems. While pre-training is usually performed on large real-world image datasets, in this paper we ask whether this is truly necessary. To this end, we search for a minimal, purely synthetic pre-training dataset that allows us to achieve performance similar to the 1 million images of ImageNet-1k. We construct such a dataset from a single fractal with perturbations. With this, we contribute three main findings. (i) We show that pre-training is effective even with minimal synthetic images, with performance on par with largescale pre-training datasets like ImageNet-1k for full fine-tuning. (ii) We investigate the single parameter with which we construct artificial categories for our dataset. We find that while the shape differences can be indistinguishable to humans, they are crucial for obtaining strong performances. (iii) Finally, we investigate the minimal requirements for successful pre-training. Surprisingly, we find that a substantial reduction of synthetic images from 1k to 1 can even lead to an increase in pre-training performance, a motivation to further investigate "scaling backwards". Finally, we extend our method from synthetic images to real images to see if a single real image can show similar pre-training effect through shape augmentation. We find that the use of grayscale images and affine transformations allows even real images to "scale backwards". The code is available at https://github.com/SUPER-TADORY/1p-frac.

For agents

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

get_harvested_code_for_paper("2408.00677")
get_code_for_paper("2408.00677")
have("2408.00677")

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

curl https://syntology.ai/api/ran/2408.00677.json

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

[![Syntology run record](https://syntology.ai/api/ran/2408.00677.svg)](https://syntology.ai/paper/2408.00677)

Connect an agent — have() is free.