SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1802.03268 · 2018

Efficient Neural Architecture Search via Parameter Sharing

Quoc Le, Barret Zoph, Hieu Pham, Jeff Dean, Melody Guan

arXiv · PDF · Open in the Atlas

Code that ran

18 of 38 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (10 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 11 have no contract verdict; 6 failures of our instrument; 20 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/1802.03268.json · badge: /api/ran/1802.03268.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
zbyte64/pytorch-dagsearch community
Papers with Code archive
5 ran · 1 no recorded run
google-research/google-research community
Papers with Code archive
3 ran · 1 no recorded run
invisibleForce/ENAS-Pytorch community
Papers with Code archive
3 ran · 1 no recorded run
kaileymonn/Quantized-ENAS-ConvNets community
Papers with Code archive
2 ran · 6 no recorded run
ahundt/renas community
Papers with Code archive
2 ran · 3 no recorded run
MengTianjian/enas-pytorch community
Papers with Code archive
1 ran · 0 no recorded run
rutgerswiselab/autolossgen community
Papers with Code archive
1 ran · 0 no recorded run
RualPerez/AutoML community
Papers with Code archive
1 ran · 0 no recorded run
MINGUKKANG/ENAS-Tensorflow community
Papers with Code archive
0 ran · 4 no recorded run
MINGUKKANG/PNU_Termproject_ENAS community
Papers with Code archive
0 ran · 1 no recorded run
ahundt/enas community
Papers with Code archive
0 ran · 1 no recorded run
distrue/enas_tensorflow community
Papers with Code archive
0 ran · 1 no recorded run
melodyguan/enas community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
CNN Ran
ran on a synthesized input · constructed an object; no result was computed
MengTianjian/enas-pytorch/micro_child.py
code served (permissive licence) · get_code("eab22bd84e3559c2")
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 an instance of its own class, CNN: the run constructed the object and did not call it, so no result was computed. 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
from torch import nn

torch.manual_seed(0)

class Args:
    def __init__(self):
        self.child_num_layers = 9
        self.child_out_filters = 16
        self.child_num_branches = 4
        self.child_num_cells = 3
        self.child_use_aux_heads = True

ARGS = {"args": Args()}
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 CNN — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a CNN 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
Connector Ran
ran on a synthesized input · constructed an object; no result was computed
zbyte64/pytorch-dagsearch/dagsearch/dag.py
pointer only (licence: NONE) · get_code("f4716de831df0059")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zbyte64/pytorch-dagsearch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Controller Ran
ran on a synthesized input · constructed an object; no result was computed
google-research/google-research/enas_lm/src/controller.py
code served (permissive licence) · get_code("c8730537c0104a93")
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 an instance of its own class, Controller: the run constructed the object and did not call it, so no result was computed. 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

tf.compat.v1.disable_eager_execution()

# Define FLAGS with required attributes
flags = tf.compat.v1.app.flags
FLAGS = flags.FLAGS

flags.DEFINE_integer('controller_num_layers', 2, 'Number of controller layers')
flags.DEFINE_float('controller_baseline_dec', 0.99, 'Baseline decay')
flags.DEFINE_float('controller_entropy_weight', 0.0001, 'Entropy weight')
flags.DEFINE_float('controller_temperature', 1.0, 'Temperature')
flags.DEFINE_float('controller_tanh_constant', 1.5, 'Tanh constant')
flags.DEFINE_float('controller_learning_rate', 0.001, 'Learning rate')

# Create a minimal HParams object
class HParams:
    def __init__(self):
        self._dict = {}
    
    def add_hparam(self, key, value):
        self._dict[key] = value
        setattr(self, key, value)
    
    def __getattr__(self, key):
        if key.startswith('_'):
            return super().__getattribute__(key)
        if key in self._dict:
            return self._dict[key]
        raise AttributeError(f"'{type(self).__name__}' object has no attribute '{key}'")

# Create params object
params = HParams()

ARGS = {
    'params': params,
    'name': 'controller'
}
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 Controller — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Controller 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
Graph Ran
ran on a synthesized input · constructed an object; no result was computed
zbyte64/pytorch-dagsearch/dagsearch/dag.py
pointer only (licence: NONE) · get_code("acc4bf1454c8cc39")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zbyte64/pytorch-dagsearch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

HyperCell Ran
ran on a synthesized input · constructed an object; no result was computed
zbyte64/pytorch-dagsearch/dagsearch/dag.py
pointer only (licence: NONE) · get_code("f356225b3b24f114")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zbyte64/pytorch-dagsearch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Identity Ran
ran on a synthesized input
zbyte64/pytorch-dagsearch/dagsearch/dag.py
pointer only (licence: NONE) · get_code("cefafcb412701ad5")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zbyte64/pytorch-dagsearch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Interpolate Ran
ran on a synthesized input · constructed an object; no result was computed
zbyte64/pytorch-dagsearch/dagsearch/dag.py
pointer only (licence: NONE) · get_code("2517ea4a652117c0")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zbyte64/pytorch-dagsearch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

LayerOp Ran
ran on a synthesized input · constructed an object; no result was computed
invisibleForce/ENAS-Pytorch/child_model.py
pointer only (licence: NONE) · get_code("83ec11e0af4a6100")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, invisibleForce/ENAS-Pytorch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

LossLayer Ran
ran on a synthesized input · constructed an object; no result was computed
rutgerswiselab/autolossgen/src/models/LossFormula.py
pointer only (licence: NONE) · get_code("53b29d9f18830ebd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, rutgerswiselab/autolossgen when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

NasLayer Ran
ran on a synthesized input · constructed an object; no result was computed
invisibleForce/ENAS-Pytorch/child_model.py
pointer only (licence: NONE) · get_code("d733f480a9cc9d6a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, invisibleForce/ENAS-Pytorch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

PolicyNet Ran
ran on a synthesized input · constructed an object; no result was computed
RualPerez/AutoML/policy.py
pointer only (licence: NONE) · get_code("196af5ccc39415a7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, RualPerez/AutoML when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

_lstm Ran
ran; our contract draft was wrong, not the code
google-research/google-research/enas_lm/src/controller.py
code served (permissive licence) · get_code("0c0acd8344bfebad")
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 tensorflow as tf
import numpy as np

tf.compat.v1.disable_eager_execution()

# Set random seed for reproducibility
np.random.seed(0)
tf.compat.v1.set_random_seed(0)

# Define tensor dimensions
batch_size = 2
input_dim = 4
hidden_dim = 3

# Create input tensor x: [batch_size, input_dim]
x = tf.constant(np.random.randn(batch_size, input_dim).astype(np.float32))

# Create previous cell state prev_c: [batch_size, hidden_dim]
prev_c = tf.constant(np.random.randn(batch_size, hidden_dim).astype(np.float32))

# Create previous hidden state prev_h: [batch_size, hidden_dim]
prev_h = tf.constant(np.random.randn(batch_size, hidden_dim).astype(np.float32))

# Create LSTM weight matrix w_lstm: [input_dim + hidden_dim, 4 * hidden_dim]
# The concatenation of [x, prev_h] has dimension input_dim + hidden_dim
# The output is split into 4 parts (i, f, o, g), each of dimension hidden_dim
w_lstm = tf.constant(
    np.random.randn(input_dim + hidden_dim, 4 * hidden_dim).astype(np.float32)
)

ARGS = {
    "x": x,
    "prev_c": prev_c,
    "prev_h": prev_h,
    "w_lstm": w_lstm,
}
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 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
_set_default_params Ran
ran; our contract draft was wrong, not the code
google-research/google-research/enas_lm/src/controller.py
code served (permissive licence) · get_code("e2997adeaed89759")
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 MockHParams. 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
import sys

tf = tensorflow.compat.v1

# Disable v2 behavior for tf.app.flags compatibility
tf.disable_v2_behavior()

flags = tf.app.flags
FLAGS = flags.FLAGS

# Define the required flags that are referenced in the function
flags.DEFINE_integer('controller_num_layers', 2, 'Number of layers')
flags.DEFINE_float('controller_baseline_dec', 0.99, 'Baseline decay')
flags.DEFINE_float('controller_entropy_weight', 0.0001, 'Entropy weight')
flags.DEFINE_float('controller_temperature', 5.0, 'Temperature')
flags.DEFINE_float('controller_tanh_constant', 2.25, 'Tanh constant')
flags.DEFINE_float('controller_learning_rate', 0.001, 'Learning rate')

# Create a mock HParams object since we're using tf.app.flags
class MockHParams:
    def __init__(self):
        self._hparams = {}
    
    def add_hparam(self, key, value):
        self._hparams[key] = value
        setattr(self, key, value)
    
    def __repr__(self):
        return str(self._hparams)

# Set up arguments for the function
params = MockHParams()

ARGS = {
    'params': params
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MockHParams — a MockHParams output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockHParams 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
count_model_params Ran
ran, honoured the contract we drafted
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("4ad735e14ee6c0d4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

global_avg_pool Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("c122e163d2187416")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

global_avgpool Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
invisibleForce/ENAS-Pytorch/child_model.py
pointer only (licence: NONE) · get_code("5bb25c6ad202549b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, invisibleForce/ENAS-Pytorch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

lstm Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ahundt/renas/enas/cifar10/general_controller.py
code served (permissive licence) · get_code("266925fcd482144b")
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 3, 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 tensorflow as tf

tf.random.set_seed(0)

# Determine sizes from function body:
# - x and prev_h are concatenated, then matmul with w
# - ifog = tf.matmul(tf.concat([x, prev_h], axis=1), w)
# - ifog is split into 4 parts along axis=1
# - Element-wise operations between i, g and f, prev_c, o
# So: x shape (batch, input_dim), prev_h shape (batch, hidden_dim)
# concat gives (batch, input_dim + hidden_dim)
# w shape must be (input_dim + hidden_dim, 4 * hidden_dim)
# ifog shape (batch, 4 * hidden_dim)
# Each of i, f, o, g has shape (batch, hidden_dim)

batch_size = 2
input_dim = 3
hidden_dim = 4

x = tf.constant([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]], dtype=tf.float32)  # shape (2, 3)
prev_c = tf.constant([[0.1, 0.2, 0.3, 0.4], [0.5, 0.6, 0.7, 0.8]], dtype=tf.float32)  # shape (2, 4)
prev_h = tf.constant([[0.01, 0.02, 0.03, 0.04], [0.05, 0.06, 0.07, 0.08]], dtype=tf.float32)  # shape (2, 4)

# w must map from (input_dim + hidden_dim) to (4 * hidden_dim)
# w shape: (7, 16)
w = tf.random.normal((input_dim + hidden_dim, 4 * hidden_dim), dtype=tf.float32)

ARGS = {
    "x": x,
    "prev_c": prev_c,
    "prev_h": prev_h,
    "w": w,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 3, 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
stack_lstm Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ahundt/renas/enas/cifar10/general_controller.py
code served (permissive licence) · get_code("19717b48f9cb0a9a")
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 4, 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 tensorflow as tf
import numpy as np

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

# Parameters
batch_size = 2
input_dim = 4
hidden_dim = 3
num_layers = 2

# Create input x: shape (batch_size, input_dim)
x = tf.constant(np.random.randn(batch_size, input_dim).astype(np.float32))

# Create prev_c (list of cell states for each layer): each shape (batch_size, hidden_dim)
prev_c = [
    tf.constant(np.random.randn(batch_size, hidden_dim).astype(np.float32))
    for _ in range(num_layers)
]

# Create prev_h (list of hidden states for each layer): each shape (batch_size, hidden_dim)
prev_h = [
    tf.constant(np.random.randn(batch_size, hidden_dim).astype(np.float32))
    for _ in range(num_layers)
]

# Create weight matrices for each layer
# For layer 0: input is x (dim input_dim), concatenated with prev_h (dim hidden_dim)
# So w[0] shape is (input_dim + hidden_dim, 4 * hidden_dim)
# For layer 1+: input is prev_h from layer i-1 (dim hidden_dim), concatenated with prev_h (dim hidden_dim)
# So w[i] shape is (hidden_dim + hidden_dim, 4 * hidden_dim) = (2*hidden_dim, 4*hidden_dim)

w = [
    tf.constant(np.random.randn(input_dim + hidden_dim, 4 * hidden_dim).astype(np.float32)),
    tf.constant(np.random.randn(hidden_dim + hidden_dim, 4 * hidden_dim).astype(np.float32))
]

ARGS = {
    'x': x,
    'prev_c': prev_c,
    'prev_h': prev_h,
    'w': w
}
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 tuple, ndim 4, 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
ChildModel Not yet run
unverified (harvested, no recorded run)
invisibleForce/ENAS-Pytorch/child_model.py
pointer only (licence: NONE) · get_code("864f33691502f6c2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, invisibleForce/ENAS-Pytorch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Controller Not yet run
unverified (harvested, no recorded run)
ahundt/renas/enas/cifar10/general_controller.py
code served (permissive licence) · get_code("6d91b9378fc9430c")
How it ran

no invocation recorded.

GeneralController Not yet run
unverified (harvested, no recorded run)
ahundt/renas/enas/cifar10/general_controller.py
code served (permissive licence) · get_code("ececc4762c60d0f9")
How it ran

no invocation recorded.

MicroChild Not yet run
unverified (harvested, no recorded run)
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("9031a7dc27f32e30")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

MicroChild Not yet run
unverified (harvested, no recorded run)
MINGUKKANG/ENAS-Tensorflow/micro_child.py
pointer only (licence: NONE) · get_code("1ce2168d55961e5d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MINGUKKANG/ENAS-Tensorflow when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

MicroController Not yet run
unverified (harvested, no recorded run)
MINGUKKANG/PNU_Termproject_ENAS/micro_controller.py
pointer only (licence: NONE) · get_code("2b3cfe1fcdcd863b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MINGUKKANG/PNU_Termproject_ENAS when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

MicroController Not yet run
unverified (harvested, no recorded run)
ahundt/enas/enas/cifar10/micro_controller.py
code served (permissive licence) · get_code("5bd590a434395d44")
How it ran

no invocation recorded.

MicroController Not yet run
unverified (harvested, no recorded run)
distrue/enas_tensorflow/src/cifar10/micro_controller.py
code served (permissive licence) · get_code("2c0648dbc9adf7c3")
How it ran

no invocation recorded.

MicroController Not yet run
unverified (harvested, no recorded run)
melodyguan/enas/src/cifar10/micro_controller.py
code served (permissive licence) · get_code("86474604afa3121b")
How it ran

no invocation recorded.

Model Not yet run
unverified (harvested, no recorded run)
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("28a9ce542ac9b038")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Model Not yet run
unverified (harvested, no recorded run)
MINGUKKANG/ENAS-Tensorflow/micro_child.py
pointer only (licence: NONE) · get_code("56d5391895065250")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MINGUKKANG/ENAS-Tensorflow when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

View Not yet run
unverified (harvested, no recorded run)
zbyte64/pytorch-dagsearch/dagsearch/dag.py
pointer only (licence: NONE) · get_code("d44cc1b50bd34d5c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, zbyte64/pytorch-dagsearch when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

_build_train_op Not yet run
unverified (harvested, no recorded run)
google-research/google-research/enas_lm/src/controller.py
code served (permissive licence) · get_code("848fb3ac5d7601b9")
How it ran

no invocation recorded.

batch_norm Not yet run
unverified (harvested, no recorded run)
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("5b9ee0a559cb458e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

create_weight Not yet run
unverified (harvested, no recorded run)
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("6abbd474b4d1304b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

create_weight Not yet run
unverified (harvested, no recorded run)
MINGUKKANG/ENAS-Tensorflow/micro_child.py
pointer only (licence: NONE) · get_code("f2a563259f2890ae")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MINGUKKANG/ENAS-Tensorflow when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

drop_path Not yet run
unverified (harvested, no recorded run)
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("04864e18c26d7b4a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

get_train_ops Not yet run
unverified (harvested, no recorded run)
ahundt/renas/enas/cifar10/general_controller.py
code served (permissive licence) · get_code("c6454e6006e3888e")
How it ran

no invocation recorded.

get_train_ops Not yet run
unverified (harvested, no recorded run)
kaileymonn/Quantized-ENAS-ConvNets/micro_child.py
pointer only (licence: NONE) · get_code("6176866b222836d8")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, kaileymonn/Quantized-ENAS-ConvNets when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

get_train_ops Not yet run
unverified (harvested, no recorded run)
MINGUKKANG/ENAS-Tensorflow/micro_child.py
pointer only (licence: NONE) · get_code("dbdb4bb390fc9aaf")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, MINGUKKANG/ENAS-Tensorflow when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 38 samples lifted from this paper's linked repositories, 38 carry a function identity and 0 do not. Of the 35 distinct functions with an identity, 4 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 4 later) and 31 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
brick of this paper's assembly: helper
10 other papers
1 dated earlier, 9 later · 8 only as an assembly brick
1511.06434 (2015-11)
Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
2103.02376 (2021-03)
Event-based Synthetic Aperture Imaging with a Hybrid Network
2103.13253 (2021-03)
Learning Versatile Neural Architectures by Propagating Network Codes
and 7 more not named here
create_weight
brick of this paper's assembly: helper
1 other paper
1 later
1904.11088 (2019-04)
D-VAE: A Variational Autoencoder for Directed Acyclic Graphs
lstm
brick of this paper's assembly: helper
1 other paper
1 later
1904.11088 (2019-04)
D-VAE: A Variational Autoencoder for Directed Acyclic Graphs
stack_lstm
brick of this paper's assembly: helper
1 other paper
1 later
1904.11088 (2019-04)
D-VAE: A Variational Autoencoder for Directed Acyclic Graphs

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

We propose Efficient Neural Architecture Search (ENAS), a fast and inexpensive approach for automatic model design. In ENAS, a controller discovers neural network architectures by searching for an optimal subgraph within a large computational graph. The controller is trained with policy gradient to select a subgraph that maximizes the expected reward on a validation set. Meanwhile the model corresponding to the selected subgraph is trained to minimize a canonical cross entropy loss. Sharing parameters among child models allows ENAS to deliver strong empirical performances, while using much fewer GPUhours than existing automatic model design approaches, and notably, 1000x less expensive than standard Neural Architecture Search. On the Penn Treebank dataset, ENAS discovers a novel architecture that achieves a test perplexity of 55.8, establishing a new state-of-the-art among all methods without post-training processing. On the CIFAR-10 dataset, ENAS finds a novel architecture that achieves 2.89% test error, which is on par with the 2.65% test error of NAS-Net (Zoph et al., 2018).

For agents

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

get_harvested_code_for_paper("1802.03268")
get_code_for_paper("1802.03268")
have("1802.03268")

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

curl https://syntology.ai/api/ran/1802.03268.json

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

[![Syntology run record](https://syntology.ai/api/ran/1802.03268.svg)](https://syntology.ai/paper/1802.03268)

Connect an agent — have() is free.