SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2010.09063 · NeurIPS · 2021

Enabling Fast Differentially Private SGD via Just-in-Time Compilation and Vectorization *

Pranav Subramani, Nicholas Vadivelu, Gautam Kamath

arXiv · PDF · Open in the Atlas

Code that ran

5 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; 0 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 0 failures of our instrument; 6 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/2010.09063.json · badge: /api/ran/2010.09063.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
TheSalon/fast-dpsgd canonical (classifier label)
classifier basis: census
5 ran · 6 no recorded run
FunctionStatusWhere it lives
get_ffnn_model Ran
ran on a synthesized input
TheSalon/fast-dpsgd/tf2dp.py
code served (permissive licence) · get_code("1ec7af9e54653150")
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 Sequential. 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
import numpy as np

# Create a small numpy array to serve as features
# Shape: (batch_size=2, feature_dim=4)
features = np.random.randn(2, 4).astype(np.float32)

# Define the arguments dictionary
ARGS = {
    "features": features,
    "batch_size": 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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential 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
get_logreg_model Ran
ran on a synthesized input
TheSalon/fast-dpsgd/tf2dp.py
code served (permissive licence) · get_code("9f2833868c634062")
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 Sequential. 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
import numpy as np

# Create a small feature array with shape (batch_size, num_features)
# Using batch_size=2 and 4 features as per requirements
features = np.random.randn(2, 4).astype(np.float32)

# Define ARGS dict with all parameters
ARGS = {
    "features": features,
    "batch_size": 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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential 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
get_parser Ran
ran on a synthesized input
TheSalon/fast-dpsgd/utils.py
code served (permissive licence) · get_code("95ad6d77417195b7")
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 ArgumentParser. 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 argparse
import torch
import torch.nn as nn

torch.manual_seed(0)

# Define experiments list
experiments = ['experiment_1', 'experiment_2', 'experiment_3']

# Create ARGS dict with all parameters for get_parser
ARGS = {
    'experiments': experiments
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
load_dummy_cifar10 Ran
ran on a synthesized input
TheSalon/fast-dpsgd/data.py
code served (permissive licence) · get_code("32038231373d2d7f")
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 numpy as np

ARGS = {
    'num_examples': 2,
    'format': 'NHWC'
}
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 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
make_broadcastable Ran
ran on a synthesized input
TheSalon/fast-dpsgd/backpackdp.py
code served (permissive licence) · get_code("4b2b5048bdb44d1e")
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 3, dtype family float, shape (2, 1, 1) 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)

# Construct test tensors
# v should be 1-dimensional with shape [N]
v = torch.randn(2)

# X should be a tensor with shape [N, ..., ] where first dimension matches v
# Using shape [2, 3, 4] to have distinct dimensions
X = torch.randn(2, 3, 4)

ARGS = {
    "v": v,
    "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 3, dtype family float, shape (2, 1, 1) — shape (2, 1, 1) 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
ffnn_model Not yet run
unverified (harvested, no recorded run)
TheSalon/fast-dpsgd/jaxdp.py
code served (permissive licence) · get_code("8e6323221fce65a0")
How it ran

no invocation recorded.

get_mnist_model Not yet run
unverified (harvested, no recorded run)
TheSalon/fast-dpsgd/tf2dp.py
code served (permissive licence) · get_code("813d74ecc59ea4de")
How it ran

no invocation recorded.

load_cifar10 Not yet run
unverified (harvested, no recorded run)
TheSalon/fast-dpsgd/data.py
code served (permissive licence) · get_code("b61d063e75caff81")
How it ran

no invocation recorded.

load_mnist Not yet run
unverified (harvested, no recorded run)
TheSalon/fast-dpsgd/data.py
code served (permissive licence) · get_code("923ce6d32026c59f")
How it ran

no invocation recorded.

logistic_model Not yet run
unverified (harvested, no recorded run)
TheSalon/fast-dpsgd/jaxdp.py
code served (permissive licence) · get_code("f0ae747648db4abb")
How it ran

no invocation recorded.

mnist_model Not yet run
unverified (harvested, no recorded run)
TheSalon/fast-dpsgd/jaxdp.py
code served (permissive licence) · get_code("10e1c3a30493b84e")
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, 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 11 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

A common pain point in differentially private machine learning is the significant runtime overhead incurred when executing Differentially Private Stochastic Gradient Descent (DPSGD), which may be as large as two orders of magnitude. We thoroughly demonstrate that by exploiting powerful language primitives, including vectorization, just-in-time compilation, and static graph optimization, one can dramatically reduce these overheads, in many cases nearly matching the best non-private running times. These gains are realized in two frameworks: one is JAX, which provides rich support for these primitives through the XLA compiler. We also rebuild core parts of TensorFlow Privacy, integrating more effective vectorization as well as XLA compilation, granting significant memory and runtime improvements over previous release versions. Our proposed approaches allow us to achieve up to 50x speedups compared to the best alternatives. Our code is available at https://github.com/TheSalon/fast-dpsgd.

For agents

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

get_harvested_code_for_paper("2010.09063")
get_code_for_paper("2010.09063")
have("2010.09063")

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

curl https://syntology.ai/api/ran/2010.09063.json

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

[![Syntology run record](https://syntology.ai/api/ran/2010.09063.svg)](https://syntology.ai/paper/2010.09063)

Connect an agent — have() is free.