SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1803.02999 · 2018

On First-Order Meta-Learning Algorithms

Joshua Achiam, John Schulman, Alex Nichol

arXiv · PDF · Open in the Atlas

Code that ran

27 of 42 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 2 violated it and 4 have no contract verdict; 20 failures of our instrument; 15 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/1803.02999.json · badge: /api/ran/1803.02999.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
peisungtsai/Reptile-Pytorch-Implementation community
Papers with Code archive
8 ran · 2 no recorded run
openai/supervised-reptile marked official in Papers with Code
Papers with Code archive
8 ran · 2 no recorded run
hfahrudin/reptile_implement_tf2 community
Papers with Code archive
3 ran · 4 no recorded run
sanowar-raihan/nerf-meta community
Papers with Code archive
3 ran · 2 no recorded run
gabrielhuang/reptile-pytorch community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
radrumond/Chameleon community
Papers with Code archive
1 ran · 2 no recorded run
MaximeVandegar/Papers-in-100-Lines-of-Code community
Papers with Code archive
1 ran · 2 no recorded run
learnables/learn2learn community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
MiniimagenetModel Ran
ran on a synthesized input · constructed an object; no result was computed
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("8d746ce18ec482a7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

SineModel Ran
ran on a synthesized input · constructed an object; no result was computed
hfahrudin/reptile_implement_tf2/Reptile.py
code served (permissive licence) · get_code("4c7e734776a0037f")
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, SineModel: 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.keras as keras
import numpy as np

# SineModel.__init__ takes no parameters beyond self
ARGS = {}
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 SineModel — 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 SineModel 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
Variable_ Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("e2090b8a7ed2f304")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

_mini_batches Ran
ran; our contract draft was wrong, not the code
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("a9fca842e7dce3dc")
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 generator. 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
import random

torch.manual_seed(0)
random.seed(0)

# Create sample data - list of (input, label) tuples
samples = [
    (torch.randn(4), torch.tensor(0)),
    (torch.randn(4), torch.tensor(1)),
    (torch.randn(4), torch.tensor(0)),
    (torch.randn(4), torch.tensor(1)),
    (torch.randn(4), torch.tensor(0)),
    (torch.randn(4), torch.tensor(1)),
]

ARGS = {
    "samples": samples,
    "batch_size": 2,
    "num_batches": 3,
    "replacement": False,
}
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 generator — a generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a generator 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
_sample_mini_dataset Ran
ran; our contract draft was wrong, not the code
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("8a4d810f710c7156")
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 generator. 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 random

class MockClass:
    """Stand-in for a class object that can be sampled from."""
    def __init__(self, class_id, num_samples=10):
        self.class_id = class_id
        self.samples_list = [f"sample_{class_id}_{i}" for i in range(num_samples)]
    
    def sample(self, num_shots):
        """Return num_shots samples from this class."""
        return random.sample(self.samples_list, min(num_shots, len(self.samples_list)))

# Create a mock dataset as a list of MockClass objects
dataset = [
    MockClass(class_id=0, num_samples=10),
    MockClass(class_id=1, num_samples=10),
    MockClass(class_id=2, num_samples=10),
    MockClass(class_id=3, num_samples=10),
    MockClass(class_id=4, num_samples=10),
]

ARGS = {
    "dataset": dataset,
    "num_classes": 3,
    "num_shots": 2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type generator — a generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a generator 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
_split_train_test Ran
ran; our contract draft was wrong, not the code
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("d91fc29b93983a36")
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 torch
import numpy as np

torch.manual_seed(0)

# The function expects an iterable of (input, label) pairs
# It doesn't actually use the input values, only the labels for splitting
# So we can use simple tensors or any objects as inputs

# Create a few-shot task with multiple classes
# Let's have 3 classes, with 3 examples each
samples = [
    (torch.randn(4), 0),
    (torch.randn(4), 0),
    (torch.randn(4), 0),
    (torch.randn(4), 1),
    (torch.randn(4), 1),
    (torch.randn(4), 1),
    (torch.randn(4), 2),
    (torch.randn(4), 2),
    (torch.randn(4), 2),
]

ARGS = {
    "samples": samples,
    "test_shots": 1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — 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
add_vars Ran
ran; our contract draft was wrong, not the code
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("4e38232a76ab2c21")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# Create two variable sequences (lists of tensors)
# Each sequence contains tensors that will be element-wise added
var_seq_1 = [
    torch.randn(2, 4),
    torch.randn(2, 3),
    torch.randn(5, 4)
]

var_seq_2 = [
    torch.randn(2, 4),
    torch.randn(2, 3),
    torch.randn(5, 4)
]

ARGS = {
    "var_seq_1": var_seq_1,
    "var_seq_2": var_seq_2
}
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 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
average_vars Ran
ran, violated the contract we drafted
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("59c0d53cfcbb2398")
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 2, 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 sample variable sequences
# The function expects var_seqs to be an iterable of sequences
# Each sequence will be zipped together, so we need multiple sequences
# with the same number of elements

# Create 3 sequences, each containing 4 variable arrays
seq1 = [np.array([1.0, 2.0, 3.0]), np.array([4.0, 5.0, 6.0]), 
        np.array([7.0, 8.0, 9.0]), np.array([10.0, 11.0, 12.0])]

seq2 = [np.array([2.0, 3.0, 4.0]), np.array([5.0, 6.0, 7.0]), 
        np.array([8.0, 9.0, 10.0]), np.array([11.0, 12.0, 13.0])]

seq3 = [np.array([3.0, 4.0, 5.0]), np.array([6.0, 7.0, 8.0]), 
        np.array([9.0, 10.0, 11.0]), np.array([12.0, 13.0, 14.0])]

var_seqs = [seq1, seq2, seq3]

ARGS = {
    "var_seqs": var_seqs
}
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 2, dtype family float — 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
average_vars Ran
ran; our contract draft was wrong, not the code
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("634c4ae3d42d183d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

clone_model Ran
ran; our contract draft was wrong, not the code
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("5756a43251ee78a0")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

copy_model Ran
ran; our contract draft was wrong, not the code
hfahrudin/reptile_implement_tf2/Reptile.py
code served (permissive licence) · get_code("5b32ff71531a63f4")
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 SineModel. 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 tensorflow.keras as keras
import numpy as np

# Create a source model with weights
source_model = keras.Sequential([
    keras.layers.Dense(40, input_shape=(1,)),
    keras.layers.Dense(40),
    keras.layers.Dense(1)
])

# Initialize weights by doing a forward pass
dummy_input = tf.constant([[1.0], [2.0]], dtype=tf.float32)
source_model(dummy_input)

# Create input example for the copy_model function
x = np.array([[1.0], [2.0]], dtype=np.float32)

ARGS = {
    "model": source_model,
    "x": x
}
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 SineModel — a SineModel output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SineModel 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
find_latest_file Ran
ran on a synthesized input
gabrielhuang/reptile-pytorch/utils.py
code served (permissive licence) · get_code("b63cb25f52bfa440")
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 str. 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 os
import tempfile
import shutil

# Create a temporary directory with test files
temp_dir = tempfile.mkdtemp()

# Create some test files with numbers in their names
test_files = [
    "checkpoint_42.pt",
    "model_10.pth",
    "data_999.pkl",
    "file_5.txt",
    "no_numbers.txt",
    "multi_1_2_3.pt"  # This has multiple numbers, should be skipped
]

for fname in test_files:
    open(os.path.join(temp_dir, fname), 'w').close()

ARGS = {
    "folder": temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("1aec86f12ed5bdbf")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

get_optimizer Ran
ran; our contract draft was wrong, not the code
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("4119f639d71237d9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

get_rays_shapenet Ran
ran; our contract draft was wrong, not the code
sanowar-raihan/nerf-meta/shapenet_train.py
code served (permissive licence) · get_code("874a035aef959ab0")
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 5, 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 torch

torch.manual_seed(0)

# Create hwf: (H, W, focal)
H = 4
W = 5
focal = torch.tensor(10.0)
hwf = torch.tensor([H, W], dtype=torch.float32), focal

# Actually, hwf should be a tuple or tensor of 3 elements
# Let's make it a proper 3-element structure
hwf = (torch.tensor(4.0), torch.tensor(5.0), torch.tensor(10.0))

# Create poses: (N, 4, 4) where N is number of images
# Standard 4x4 pose matrices (homogeneous transformation)
N = 2
poses = torch.eye(4).unsqueeze(0).repeat(N, 1, 1)
# Add slight variation to second pose
poses[1, :3, 3] = torch.tensor([1.0, 2.0, 3.0])

ARGS = {
    "hwf": hwf,
    "poses": poses,
}
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 tuple, ndim 5, 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
interpolate_vars Ran
ran; our contract draft was wrong, not the code
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("b6a6811d1b4f3cb1")
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 torch

torch.manual_seed(0)

# Create small tensor sequences to represent variables
# Using distinct sizes: 2 tensors of shape (2, 4) and (2, 3)
old_vars = [torch.randn(2, 4), torch.randn(2, 3)]
new_vars = [torch.randn(2, 4), torch.randn(2, 3)]
epsilon = 0.5

ARGS = {
    "old_vars": old_vars,
    "new_vars": new_vars,
    "epsilon": epsilon,
}
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 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
interpolate_vars Ran
ran; our contract draft was wrong, not the code
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("e1d757cc1d97469e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

list_dir Ran
ran; our contract draft was wrong, not the code
gabrielhuang/reptile-pytorch/utils.py
code served (permissive licence) · get_code("9bc2cdc80cede326")
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 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 os
import tempfile
import shutil

# Create a temporary directory with some subdirectories for testing
temp_dir = tempfile.mkdtemp()
os.makedirs(os.path.join(temp_dir, "dir1"), exist_ok=True)
os.makedirs(os.path.join(temp_dir, "dir2"), exist_ok=True)
os.makedirs(os.path.join(temp_dir, "dir3"), exist_ok=True)
# Create a file to ensure filter works correctly
with open(os.path.join(temp_dir, "file.txt"), "w") as f:
    f.write("test")

ARGS = {
    "root": temp_dir,
    "prefix": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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 float — 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
list_files Ran
ran; our contract draft was wrong, not the code
gabrielhuang/reptile-pytorch/utils.py
code served (permissive licence) · get_code("fdee3ec4c52c3c4e")
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 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 os
import tempfile
import shutil

# Create a temporary directory with some test files
temp_dir = tempfile.mkdtemp()

# Create some test files with different suffixes
open(os.path.join(temp_dir, "file1.png"), "w").close()
open(os.path.join(temp_dir, "file2.png"), "w").close()
open(os.path.join(temp_dir, "file3.jpg"), "w").close()
open(os.path.join(temp_dir, "file4.txt"), "w").close()
open(os.path.join(temp_dir, "file5.jpg"), "w").close()

ARGS = {
    "root": temp_dir,
    "suffix": ".png",
    "prefix": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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 float — 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
np_to_tensor Ran
ran; our contract draft was wrong, not the code
hfahrudin/reptile_implement_tf2/Reptile.py
code served (permissive licence) · get_code("6898b2b8e89e013c")
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 generator. 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
import tensorflow as tf

# Create a list of small numpy arrays to pass to np_to_tensor
numpy_obj_1 = np.array([[1, 2, 3, 4], [5, 6, 7, 8]], dtype=np.float32)
numpy_obj_2 = np.array([[9, 10, 11, 12]], dtype=np.int32)
numpy_obj_3 = np.array([[[1, 2], [3, 4]], [[5, 6], [7, 8]]], dtype=np.float32)

list_of_numpy_objs = [numpy_obj_1, numpy_obj_2, numpy_obj_3]

ARGS = {
    "list_of_numpy_objs": list_of_numpy_objs
}
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 generator — a generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a generator 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
padCols Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
radrumond/Chameleon/code/reptile.py
pointer only (licence: NONE) · get_code("663789fdc8b48fc7")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, radrumond/Chameleon 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.

predict_label Ran
ran, honoured the contract we drafted
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("1e9509d17fafd962")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

sample_points Ran
ran; our contract draft was wrong, not the code
sanowar-raihan/nerf-meta/shapenet_train.py
code served (permissive licence) · get_code("ee6a21f2e9ccd76e")
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 torch

torch.manual_seed(0)

# Small tensors with distinct sizes for different dimensions
num_rays = 2
ray_dim = 3
num_samples = 4

rays_o = torch.randn(num_rays, ray_dim)
rays_d = torch.randn(num_rays, ray_dim)
near = 0.1
far = 10.0
num_samples = num_samples
perturb = False

ARGS = {
    "rays_o": rays_o,
    "rays_d": rays_d,
    "near": near,
    "far": far,
    "num_samples": num_samples,
    "perturb": perturb,
}
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 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
sample_task Ran
ran on a synthesized input
MaximeVandegar/Papers-in-100-Lines-of-Code/On_First_Order_Meta_Learning_Algorithms/reptile.py
code served (permissive licence) · get_code("38ba8a7c486030e7")
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 torch
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

# The function sample_task() takes no parameters
# It's a self-contained function that generates random task data internally
ARGS = {}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — 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
scale_vars Ran
ran, violated the contract we drafted
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("546a058c2eeefb61")
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 2, 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 torch
import numpy as np

torch.manual_seed(0)

# Construct a simple variable sequence
# The function iterates over var_seq and multiplies each element by scale
# So var_seq should be a list of numeric values or tensors
var_seq = [torch.tensor([1.0, 2.0, 3.0, 4.0]), 
           torch.tensor([5.0, 6.0, 7.0, 8.0]),
           torch.tensor([9.0, 10.0, 11.0, 12.0])]

# Scale factor - can be a numeric value
scale = 2.5

ARGS = {
    "var_seq": var_seq,
    "scale": scale
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family float — 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
subtract_vars Ran
ran; our contract draft was wrong, not the code
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("3880e32be4fd7c7f")
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 torch
import torch.nn as nn

torch.manual_seed(0)

# The function subtract_vars takes two sequences of variables and subtracts
# corresponding elements. Looking at the implementation, it uses:
# - zip() to pair elements
# - subtraction operator (v1 - v2) on each pair
# This works with any objects that support subtraction, but typically tensors.

# Create two sequences of tensors with distinct dimensions for probing
var_seq_1 = [
    torch.randn(2, 4),  # batch 2, dim 4
    torch.randn(2, 3),  # batch 2, dim 3
    torch.randn(2, 5),  # batch 2, dim 5
]

var_seq_2 = [
    torch.randn(2, 4),  # batch 2, dim 4
    torch.randn(2, 3),  # batch 2, dim 3
    torch.randn(2, 5),  # batch 2, dim 5
]

ARGS = {
    "var_seq_1": var_seq_1,
    "var_seq_2": var_seq_2,
}
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 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
volume_render Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
sanowar-raihan/nerf-meta/shapenet_train.py
code served (permissive licence) · get_code("ea9eecfceada7f03")
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, 3) 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)

num_rays = 2
num_samples = 4

rgbs = torch.rand(num_rays, num_samples, 3)
sigmas = torch.rand(num_rays, num_samples)
t_vals = torch.linspace(0, 1, num_samples).unsqueeze(0).expand(num_rays, -1)
white_bkgd = False

ARGS = {
    "rgbs": rgbs,
    "sigmas": sigmas,
    "t_vals": t_vals,
    "white_bkgd": white_bkgd,
}
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, 3) — shape (2, 3) 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
GBML Not yet run
unverified (harvested, no recorded run)
learnables/learn2learn/learn2learn/algorithms/gbml.py
code served (permissive licence) · get_code("c2a3170df7df4870")
How it ran

no invocation recorded.

Reptile Not yet run
unverified (harvested, no recorded run)
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("12c044306a7432d7")
How it ran

no invocation recorded.

Reptile Not yet run
unverified (harvested, no recorded run)
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("d215ca39171d7e6a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

Reptile Not yet run
unverified (harvested, no recorded run)
radrumond/Chameleon/code/reptile.py
pointer only (licence: NONE) · get_code("862a81450c933234")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, radrumond/Chameleon 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.

VariableState Not yet run
unverified (harvested, no recorded run)
openai/supervised-reptile/supervised_reptile/reptile.py
code served (permissive licence) · get_code("fc567dbf68a97f9c")
How it ran

no invocation recorded.

argument_parser Not yet run
unverified (harvested, no recorded run)
peisungtsai/Reptile-Pytorch-Implementation/reptile.py
pointer only (licence: NONE) · get_code("e5b4fcff02414010")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, peisungtsai/Reptile-Pytorch-Implementation 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.

compute_loss Not yet run
unverified (harvested, no recorded run)
hfahrudin/reptile_implement_tf2/Reptile.py
code served (permissive licence) · get_code("ac866a5f88b92797")
How it ran

no invocation recorded.

inner_loop Not yet run
unverified (harvested, no recorded run)
sanowar-raihan/nerf-meta/shapenet_train.py
code served (permissive licence) · get_code("649f40959e38546d")
How it ran

no invocation recorded.

loss_function Not yet run
unverified (harvested, no recorded run)
hfahrudin/reptile_implement_tf2/Reptile.py
code served (permissive licence) · get_code("97d6ee1a0025a4d4")
How it ran

no invocation recorded.

perform_k_training_steps Not yet run
unverified (harvested, no recorded run)
MaximeVandegar/Papers-in-100-Lines-of-Code/On_First_Order_Meta_Learning_Algorithms/reptile.py
code served (permissive licence) · get_code("7c5446461cabc053")
How it ran

no invocation recorded.

reptile Not yet run
unverified (harvested, no recorded run)
MaximeVandegar/Papers-in-100-Lines-of-Code/On_First_Order_Meta_Learning_Algorithms/reptile.py
code served (permissive licence) · get_code("a7a5b0bb081fdc90")
How it ran

no invocation recorded.

savePlot Not yet run
unverified (harvested, no recorded run)
radrumond/Chameleon/code/reptile.py
pointer only (licence: NONE) · get_code("a2a852379817fd88")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, radrumond/Chameleon 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.

train_batch Not yet run
unverified (harvested, no recorded run)
hfahrudin/reptile_implement_tf2/Reptile.py
code served (permissive licence) · get_code("6ccf8109b819b534")
How it ran

no invocation recorded.

train_meta Not yet run
unverified (harvested, no recorded run)
sanowar-raihan/nerf-meta/shapenet_train.py
code served (permissive licence) · get_code("9ad33249189e3716")
How it ran

no invocation recorded.

train_reptile Not yet run
unverified (harvested, no recorded run)
hfahrudin/reptile_implement_tf2/Reptile.py
code served (permissive licence) · get_code("7ce7bf6a46593388")
How it ran

no invocation recorded.

Also present in other papers' code

Of 42 samples lifted from this paper's linked repositories, 42 carry a function identity and 0 do not. Of the 42 distinct functions with an identity, 5 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 5 later) and 37 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
list_dir 19 other papers
19 later
1903.12648 (2019-03)
Overcoming Catastrophic Forgetting with Unlabeled Data in the Wild
1906.00189 (2019-06)
Are Anchor Points Really Indispensable in Label-Noise Learning?
2206.02262 (2022-06)
Diffusion-GAN: Training GANs with Diffusion
and 16 more not named here
list_files 16 other papers
16 later
1903.12648 (2019-03)
Overcoming Catastrophic Forgetting with Unlabeled Data in the Wild
1906.00189 (2019-06)
Are Anchor Points Really Indispensable in Label-Noise Learning?
2206.02262 (2022-06)
Diffusion-GAN: Training GANs with Diffusion
and 13 more not named here
average_vars
brick of this paper's assembly: helper
3 other papers
3 later
1904.02957 (2019-04)
Learning to Adapt for Stereo
1906.02717 (2019-06)
Adaptive Gradient-Based Meta-Learning Methods
2010.12777 (2020-10)
Improving Multilingual Models with Language-Clustered Vocabularies
interpolate_vars
brick of this paper's assembly: entry, helper
3 other papers
3 later
1904.02957 (2019-04)
Learning to Adapt for Stereo
1906.02717 (2019-06)
Adaptive Gradient-Based Meta-Learning Methods
2010.12777 (2020-10)
Improving Multilingual Models with Language-Clustered Vocabularies
subtract_vars
brick of this paper's assembly: helper
3 other papers
3 later
1904.02957 (2019-04)
Learning to Adapt for Stereo
1906.02717 (2019-06)
Adaptive Gradient-Based Meta-Learning Methods
2010.12777 (2020-10)
Improving Multilingual Models with Language-Clustered Vocabularies

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

This paper considers meta-learning problems, where there is a distribution of tasks, and we would like to obtain an agent that performs well (i.e., learns quickly) when presented with a previously unseen task sampled from this distribution. We analyze a family of algorithms for learning a parameter initialization that can be fine-tuned quickly on a new task, using only firstorder derivatives for the meta-learning updates. This family includes and generalizes first-order MAML, an approximation to MAML obtained by ignoring second-order derivatives. It also includes Reptile, a new algorithm that we introduce here, which works by repeatedly sampling a task, training on it, and moving the initialization towards the trained weights on that task. We expand on the results from Finn et al. showing that first-order meta-learning algorithms perform well on some well-established benchmarks for few-shot classification, and we provide theoretical analysis aimed at understanding why these algorithms work.

For agents

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

get_harvested_code_for_paper("1803.02999")
get_code_for_paper("1803.02999")
have("1803.02999")

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

curl https://syntology.ai/api/ran/1803.02999.json

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

[![Syntology run record](https://syntology.ai/api/ran/1803.02999.svg)](https://syntology.ai/paper/1803.02999)

Connect an agent — have() is free.