SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 0705.2011 · 2007

Multi-Dimensional Recurrent Neural Networks

arXiv · PDF · Open in the Atlas

Code that ran

16 of 23 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 15 have no contract verdict; 1 failure of our instrument; 7 no recorded run), measured 2026-11-21; 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/0705.2011.json · badge: /api/ran/0705.2011.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
jzbjyb/rri_match community (classifier label)
classifier basis not recorded on the link
15 ran · 6 no recorded run
philipperemy/tensorflow-multi-dimensional-lstm reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
T-Almeida/tensorflow-keras-multidimensional-rnn community (classifier label)
classifier basis not recorded on the link
0 ran · 1 no recorded run
FunctionStatusWhere it lives
batch_slice Ran
ran on a synthesized input
jzbjyb/rri_match/rri.py
code served (permissive licence) · get_code("3f6f8ca566974e84")
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 EagerTensor, ndim 3, dtype family float, shape (2, 3, 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 tensorflow as tf
import numpy as np

# Create a small batch tensor with shape [batch_size, sequence_length, feature_dim]
batch = tf.constant(np.random.randn(2, 4, 3).astype(np.float32))

# Create start indices for each batch element
start = tf.constant([0, 1], dtype=tf.int32)

# Create offset values (how many elements to slice)
offset = tf.constant([2, 3], dtype=tf.int32)

# Pad value for constant padding
pad_values = 0.0

ARGS = {
    "batch": batch,
    "start": start,
    "offset": offset,
    "pad_values": pad_values,
}
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 EagerTensor, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 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
batch_slice_slow Ran
ran on a synthesized input
jzbjyb/rri_match/rri.py
code served (permissive licence) · get_code("33b35e906cc63cae")
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 EagerTensor, ndim 3, dtype family float, shape (2, 3, 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 tensorflow as tf
import numpy as np

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

# Create a small batch tensor with shape (batch_size=2, sequence_length=4, feature_dim=3)
batch = tf.constant([
    [[1.0, 2.0, 3.0], [4.0, 5.0, 6.0], [7.0, 8.0, 9.0], [10.0, 11.0, 12.0]],
    [[13.0, 14.0, 15.0], [16.0, 17.0, 18.0], [19.0, 20.0, 21.0], [22.0, 23.0, 24.0]]
], dtype=tf.float32)

# Create start indices for each batch element
start = tf.constant([0, 1], dtype=tf.int32)

# Create offset values for each batch element
offset = tf.constant([2, 3], dtype=tf.int32)

# Create pad values (scalar)
pad_values = 0.0

ARGS = {
    "batch": batch,
    "start": start,
    "offset": offset,
    "pad_values": pad_values
}
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 EagerTensor, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 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
batch_where Ran
ran on a synthesized input
jzbjyb/rri_match/rri.py
code served (permissive licence) · get_code("e93059a43aa4dbfa")
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 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

# Create a condition tensor (batch_size=2, dim=4)
cond = tf.constant([[True, False, True, False],
                     [False, True, False, True]])

# Create xs as a list of tensors (when not equal to 1)
xs = [
    tf.constant([[1.0, 2.0, 3.0, 4.0],
                 [5.0, 6.0, 7.0, 8.0]], dtype=tf.float32),
    tf.constant([[9.0, 10.0, 11.0, 12.0],
                 [13.0, 14.0, 15.0, 16.0]], dtype=tf.float32)
]

# Create ys as a list of tensors with matching shape
ys = [
    tf.constant([[17.0, 18.0, 19.0, 20.0],
                 [21.0, 22.0, 23.0, 24.0]], dtype=tf.float32),
    tf.constant([[25.0, 26.0, 27.0, 28.0],
                 [29.0, 30.0, 31.0, 32.0]], dtype=tf.float32)
]

ARGS = {
    "cond": cond,
    "xs": xs,
    "ys": ys
}
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 list, ndim 3, 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
dcg Ran
ran on a synthesized input
jzbjyb/rri_match/metric.py
code served (permissive licence) · get_code("34a2a38a33c779ed")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct arguments for dcg function
query = "query_1"

doc_list = ["doc_0", "doc_1", "doc_2", "doc_3", "doc_4"]

# qd_judge is a nested dict mapping query -> doc -> relevance score
# The function accesses qd_judge[query][doc_list[i]] for each i in range(real_top_k)
qd_judge = {
    "query_1": {
        "doc_0": 3,
        "doc_1": 2,
        "doc_2": 1,
        "doc_3": 0,
        "doc_4": 2,
    }
}

top_k = 5

ARGS = {
    "query": query,
    "doc_list": doc_list,
    "qd_judge": qd_judge,
    "top_k": top_k,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
dynamic_split_and_pad_map_fn Ran
ran on a synthesized input
jzbjyb/rri_match/represent.py
code served (permissive licence) · get_code("39edbcf27e85fbea")
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 EagerTensor, ndim 3, dtype family float, shape (3, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tensorflow as tf

# Create small tensors for testing
x = tf.constant([[1.0, 2.0, 3.0, 4.0],
                 [5.0, 6.0, 7.0, 8.0],
                 [9.0, 10.0, 11.0, 12.0],
                 [13.0, 14.0, 15.0, 16.0],
                 [17.0, 18.0, 19.0, 20.0],
                 [21.0, 22.0, 23.0, 24.0]])

# sp represents split offsets - how many elements to take for each piece
sp = tf.constant([2, 3, 1], dtype=tf.int32)

ARGS = {
    "x": x,
    "sp": sp
}
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 EagerTensor, ndim 3, dtype family float, shape (3, 3, 4) — shape (3, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
dynamic_split_and_pad_map_fn_ Ran
ran on a synthesized input
jzbjyb/rri_match/represent.py
code served (permissive licence) · get_code("4cfeb52c620297d9")
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 EagerTensor, ndim 3, dtype family float, shape (2, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tensorflow as tf
import numpy as np

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

# Create a small tensor to split and pad
# Shape: (10, 4) - 10 elements along axis 0, 4 features
x = tf.constant(np.arange(40, dtype=np.float32).reshape(10, 4))

# Create start indices - where each piece begins
# Using 2 pieces for this example
start = tf.constant([0, 5], dtype=tf.int32)

# Create offset lengths - how long each piece is
# First piece: 3 elements, Second piece: 2 elements
offset = tf.constant([3, 2], dtype=tf.int32)

ARGS = {
    'x': x,
    'start': start,
    'offset': offset
}
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 EagerTensor, ndim 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
eval_mm Ran
ran on a synthesized input
jzbjyb/rri_match/ir_match_matrix.py
code served (permissive licence) · get_code("96826e3e74c40415")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 1, dtype family float, shape (2,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct a small match matrix (bs=2, h=3, w=4)
# Values represent match scores between 0 and 1
mms = np.array([
    [[0.5, 0.2, 0.8, 0.1],
     [0.3, 0.9, 0.0, 0.4],
     [0.7, 0.1, 0.6, 0.2]],
    [[0.0, 0.0, 0.0, 0.0],
     [1.0, 0.0, 1.0, 0.0],
     [0.0, 0.5, 0.0, 0.5]]
], dtype=np.float32)

features = 'tf_proximity'

ARGS = {
    'mms': mms,
    'features': features
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (2,) — shape (2,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
fft_ind_gen Ran
ran on a synthesized input
jzbjyb/rri_match/data_gen.py
code served (permissive licence) · get_code("5c573a033c9bdc92")
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 int; 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

ARGS = {
    "n": 8
}
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, ndim 1, dtype family int — 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
gaussian_random_field Ran
ran on a synthesized input
jzbjyb/rri_match/data_gen.py
code served (permissive licence) · get_code("939949c527bd2577")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (64, 48) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

def custom_pk(k):
    """Custom power spectrum function"""
    return k ** (-2.5)

ARGS = {
    "pk": custom_pk,
    "size1": 64,
    "size2": 48,
    "anisotropy": True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (64, 48) — shape (64, 48) 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
get_arguments Ran
ran; our contract draft was wrong, not the code
philipperemy/tensorflow-multi-dimensional-lstm/trainer.py
code served (permissive licence) · get_code("bf26c8d187534a2d")
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 Namespace. 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

# Create a minimal ArgumentParser with some basic arguments
parser = argparse.ArgumentParser(
    description="Test parser for get_arguments function"
)
parser.add_argument('--name', type=str, default='test', help='A test argument')
parser.add_argument('--value', type=int, default=42, help='A numeric argument')
parser.add_argument('--flag', action='store_true', help='A boolean flag')

# Mock sys.argv to avoid parsing actual command line arguments
import sys
original_argv = sys.argv
sys.argv = ['test_script.py']

ARGS = {
    'parser': parser
}

# Restore original argv
sys.argv = original_argv
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 Namespace — a Namespace output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Namespace 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_padding Ran
ran on a synthesized input
jzbjyb/rri_match/represent.py
code served (permissive licence) · get_code("3e2732158df7801d")
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 int; 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 tensorflow as tf

torch.manual_seed(0)

# Create a TensorFlow tensor (the function calls .get_shape() which is TF-style)
x = tf.constant([[1.0, 2.0, 3.0, 4.0],
                  [5.0, 6.0, 7.0, 8.0]])

axis = 1
pad = [1, 2]

ARGS = {
    "x": x,
    "axis": axis,
    "pad": pad
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow, 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 int — 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
ndcg Ran
ran on a synthesized input
jzbjyb/rri_match/metric.py
code served (permissive licence) · get_code("e41abfe79a1abfca")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Build qd_judge: a dict mapping queries to dicts of {doc: relevance_score}
# The function accesses qd_judge[query] and qd_judge[query][doc]
qd_judge = {
    "query1": {
        "doc1": 3,
        "doc2": 2,
        "doc3": 1,
        "doc4": 0,
    },
    "query2": {
        "doc5": 2,
        "doc6": 1,
        "doc7": 0,
    }
}

# Build doc_list: a list of document IDs
doc_list = ["doc1", "doc2", "doc3", "doc4"]

# query: a string key that exists in qd_judge
query = "query1"

# top_k: integer parameter
top_k = 5

ARGS = {
    "query": query,
    "doc_list": doc_list,
    "qd_judge": qd_judge,
    "top_k": top_k,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
next_batch_gaussian Ran
ran on a synthesized input
jzbjyb/rri_match/data_gen.py
code served (permissive licence) · get_code("0eb4737d6f16a402")
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 numpy as np

ARGS = {
    "bs": 2,
    "h": 4,
    "w": 3,
    "anisotropy": True
}
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, 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
parse_sogou_qcl_one_qd Ran
ran on a synthesized input
jzbjyb/rri_match/utils.py
code served (permissive licence) · get_code("3be55b1f27889067")
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 torch
import torch.nn as nn
import numpy as np

# The function parse_sogou_qcl_one_qd expects a list of lines (strings)
# that represent XML-like structured data for search queries and documents.
# Let's construct a realistic example with the nested structure:
# <q>
#   <query>...</query>
#   <query_frequency>...</query_frequency>
#   <query_id>...</query_id>
#   <doc>
#     <url>...</url>
#     <doc_id>...</doc_id>
#     <title>...</title>
#     <content>...</content>
#     <html>...</html>
#     <doc_frequency>...</doc_frequency>
#     <relevance>
#       <TCM>...</TCM>
#       <DBN>...</DBN>
#       <PSCM>...</PSCM>
#       <TACM>...</TACM>
#       <UBM>...</UBM>
#     </relevance>
#   </doc>
# </q>

lines = [
    '<q>',
    '<query>python programming</query>',
    '<query_frequency>100</query_frequency>',
    '<query_id>q001</query_id>',
    '<doc>',
    '<url>https://example.com/page1</url>',
    '<doc_id>d001</doc_id>',
    '<title>Python Guide</title>',
    '<content>Learn Python programming basics</content>',
    '<html><html>Python content</html></html>',
    '<doc_frequency>50</doc_frequency>',
    '<relevance>',
    '<TCM>0.95</TCM>',
    '<DBN>0.87</DBN>',
    '<PSCM>0.92</PSCM>',
    '<TACM>0.89</TACM>',
    '<UBM>0.91</UBM>',
    '</relevance>',
    '</doc>',
    '<doc>',
    '<url>https://example.com/page2</url>',
    '<doc_id>d002</doc_id>',
    '<title>Advanced Python</title>',
    '<content>Advanced Python concepts</content>',
    '<html><html>Advanced Python</html></html>',
    '<doc_frequency>45</doc_frequency>',
    '<relevance>',
    '<TCM>0.88</TCM>',
    '<DBN>0.80</DBN>',
    '<PSCM>0.85</PSCM>',
    '<TACM>0.82</TACM>',
    '<UBM>0.84</UBM>',
    '</relevance>',
    '</doc>',
    '</q>',
]

ARGS = {
    'lines': lines
}
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 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
precision Ran
ran on a synthesized input
jzbjyb/rri_match/metric.py
code served (permissive licence) · get_code("2007c4e9c81aaf2c")
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 float, ndim 0, dtype family float, shape () 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 arguments for the precision function
# The function expects:
# - query: a key to index into qd_judge dict
# - doc_list: a list of document identifiers
# - qd_judge: a nested dict mapping query -> doc -> relevance score
# - top_k: integer parameter (has default value of 5)

query = "q1"
doc_list = ["doc1", "doc2", "doc3", "doc4", "doc5", "doc6"]

# qd_judge is a nested dict: qd_judge[query][doc] should return a relevance score
# Positive values indicate relevant documents
qd_judge = {
    "q1": {
        "doc1": 1,      # relevant
        "doc2": 0,      # not relevant
        "doc3": 2,      # relevant
        "doc4": 0,      # not relevant
        "doc5": 1,      # relevant
        "doc6": 0,      # not relevant
    }
}

top_k = 5

ARGS = {
    "query": query,
    "doc_list": doc_list,
    "qd_judge": qd_judge,
    "top_k": top_k
}
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 float, ndim 0, dtype family float, shape () — shape () 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
word_segment Ran
ran on a synthesized input
jzbjyb/rri_match/utils.py
code served (permissive licence) · get_code("f64b418ef9cd57ca")
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 torch
import numpy as np
import jieba

# The function word_segment takes a single parameter: text
# It uses jieba.cut() to segment the text, which expects a string

ARGS = {
    "text": "今天天气很好,我想去公园散步。这是一个测试句子。"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jieba, 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 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
conv_knrm Not yet run
unverified (harvested, no recorded run)
jzbjyb/rri_match/cnn.py
code served (permissive licence) · get_code("b6525b4986ae0270")
How it ran

no invocation recorded.

knrm Not yet run
unverified (harvested, no recorded run)
jzbjyb/rri_match/cnn.py
code served (permissive licence) · get_code("c7d7c11bd2988e25")
How it ran

no invocation recorded.

ln Not yet run
unverified (harvested, no recorded run)
jzbjyb/rri_match/md_lstm.py
code served (permissive licence) · get_code("03595c4790906612")
How it ran

no invocation recorded.

map_fn_concat Not yet run
unverified (harvested, no recorded run)
jzbjyb/rri_match/basic_op.py
code served (permissive licence) · get_code("3f90df6c4fe61b20")
How it ran

no invocation recorded.

mlp Not yet run
unverified (harvested, no recorded run)
jzbjyb/rri_match/cnn.py
code served (permissive licence) · get_code("6fb3d9e6ce59a987")
How it ran

no invocation recorded.

save_graph Not yet run
unverified (harvested, no recorded run)
T-Almeida/tensorflow-keras-multidimensional-rnn/utils.py
code served (permissive licence) · get_code("9d88c7a8d285c466")
How it ran

no invocation recorded.

tf_jacobian Not yet run
unverified (harvested, no recorded run)
jzbjyb/rri_match/utils.py
code served (permissive licence) · get_code("f4f484080e6ec5d4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 23 samples lifted from this paper's linked repositories, 23 carry a function identity and 0 do not. Of the 23 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 23 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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Recurrent neural networks (RNNs) have proved effective at one dimensional sequence learning tasks, such as speech and online handwriting recognition. Some of the properties that make RNNs suitable for such tasks, for example robustness to input warping, and the ability to access contextual information, are also desirable in multidimensional domains. However, there has so far been no direct way of applying RNNs to data with more than one spatio-temporal dimension. This paper introduces multi-dimensional recurrent neural networks (MDRNNs), thereby extending the potential applicability of RNNs to vision, video processing, medical imaging and many other areas, while avoiding the scaling problems that have plagued other multi-dimensional models. Experimental results are provided for two image segmentation tasks.

For agents

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

get_harvested_code_for_paper("0705.2011")
get_code_for_paper("0705.2011")
have("0705.2011")

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

curl https://syntology.ai/api/ran/0705.2011.json

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

[![Syntology run record](https://syntology.ai/api/ran/0705.2011.svg)](https://syntology.ai/paper/0705.2011)

Connect an agent — have() is free.