SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2001.03671 · 2020

Retouchdown: Adding Touchdown to StreetLearn as a Shareable Resource for Language Grounding Tasks in Street View

arXiv · PDF · Open in the Atlas

Code that ran

8 of 8 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 7 have no contract verdict; 1 failure of our instrument; 0 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/2001.03671.json · badge: /api/ran/2001.03671.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
lil-lab/touchdown canonical (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
VegB/VLN-Transformer community (classifier label)
classifier basis not recorded on the link
7 ran · 0 no recorded run
FunctionStatusWhere it lives
attention_bias_ignore_padding Ran
ran on a synthesized input
VegB/VLN-Transformer/texar/torch/utils/transformer_attentions.py
code served (permissive licence) · get_code("139f055daa7f443e")
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 4, dtype family float, shape (2, 1, 1, 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 torch

torch.manual_seed(0)

# Create a small memory_padding tensor with shape [batch, memory_length]
# batch_size = 2, memory_length = 4
memory_padding = torch.tensor([
    [0.0, 1.0, 0.0, 1.0],
    [1.0, 0.0, 1.0, 0.0]
], dtype=torch.float32)

bias_value = -1e18

ARGS = {
    "memory_padding": memory_padding,
    "bias_value": bias_value
}
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 4, dtype family float, shape (2, 1, 1, 4) — shape (2, 1, 1, 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
attention_bias_local Ran
ran on a synthesized input
VegB/VLN-Transformer/texar/torch/utils/transformer_attentions.py
code served (permissive licence) · get_code("d82c65c3c345bed8")
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 4, dtype family float, shape (1, 1, 8, 8) 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
import numpy as np

torch.manual_seed(0)

ARGS = {
    "length": 8,
    "max_backward": 2,
    "max_forward": 3,
    "bias_value": -1e18
}
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 Tensor, ndim 4, dtype family float, shape (1, 1, 8, 8) — shape (1, 1, 8, 8) 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
attention_bias_lower_triangle Ran
ran on a synthesized input
VegB/VLN-Transformer/texar/torch/utils/transformer_attentions.py
code served (permissive licence) · get_code("f3911d17a0ab9c14")
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 4, dtype family float, shape (1, 1, 8, 8) 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
import numpy as np

torch.manual_seed(0)

ARGS = {
    "length": 8,
    "bias_value": -1e18
}
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 Tensor, ndim 4, dtype family float, shape (1, 1, 8, 8) — shape (1, 1, 8, 8) 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
bleu_transformer_tokenize Ran
ran on a synthesized input
VegB/VLN-Transformer/texar/torch/evals/bleu_transformer.py
code served (permissive licence) · get_code("27f380a9d9f1c0be")
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 re
import sys
import unicodedata

class UnicodeRegex:
    """Ad-hoc hack to recognize all punctuation and symbols."""

    def __init__(self):
        punctuation = self.property_chars('P')
        self.nondigit_punct_re = re.compile('([^\\d])([' + punctuation + '])')
        self.punct_nondigit_re = re.compile('([' + punctuation + '])([^\\d])')
        self.symbol_re = re.compile('([' + self.property_chars('S') + '])')

    @staticmethod
    def property_chars(prefix):
        return ''.join((chr(x) for x in range(sys.maxunicode) if unicodedata.category(chr(x)).startswith(prefix)))

uregex = UnicodeRegex()

def bleu_transformer_tokenize(string: str) -> list:
    """Tokenize a string following the official BLEU implementation.

    The BLEU scores from `multi-bleu.perl` depend on your `tokenizer`, which is
    unlikely to be reproducible from your experiment or consistent across
    different users. This function provides a standard tokenization following
    `mteval-v14.pl`.

    See
    `https://github.com/moses-smt/mosesdecoder/blob/master/scripts/generic/mteval-v14.pl#L954-L983`.
    In our case, the input string is expected to be just one line
    and no HTML entities de-escaping is needed.
    So we just tokenize on punctuation and symbols,
    except when a punctuation is preceded and followed by a digit
    (e.g. a comma/dot as a thousand/decimal separator).

    Note that a number (e.g. a year) followed by a dot at the end of sentence
    is NOT tokenized,
    i.e. the dot stays with the number because `s/(\\p{P})(\\P{N})/ $1 $2/g`
    does not match this case (unless we add a space after each sentence).
    However, this error is already in the original `mteval-v14.pl`
    and we want to be consistent with it.

    Args:
        string: the input string

    Returns:
        a list of tokens
    """
    string = uregex.nondigit_punct_re.sub('\\1 \\2 ', string)
    string = uregex.punct_nondigit_re.sub(' \\1 \\2', string)
    string = uregex.symbol_re.sub(' \\1 ', string)
    return string.split()

# Test strings with various punctuation and symbols
test_string_1 = "Hello, world! This is a test."
test_string_2 = "The price is $100.50 for 3 items."
test_string_3 = "Visit https://example.com (if you want)."

ARGS = {
    "string": "The year 2024. The price is 3.14, right? Yes!@#$"
}
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 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
clones Ran
ran; our contract draft was wrong, not the code
lil-lab/touchdown/sdr/model.py
pointer only (licence: CC-BY-4.0) · get_code("ba08314f96790e8e")
How it ran

The body of this sample is withheld: CC-BY-4.0, as GitHub detects it on this paper's repository, lil-lab/touchdown, is not in the permissive set Syntology serves inline. 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.

corpus_bleu_transformer Ran
ran on a synthesized input
VegB/VLN-Transformer/texar/torch/evals/bleu_transformer.py
code served (permissive licence) · get_code("b1a0c87b5ab99b19")
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 float32, 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 collections
import math
import numpy as np

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

# Create sample tokenized reference corpus
reference_corpus = [
    ["the", "cat", "is", "on", "the", "mat"],
    ["a", "dog", "runs", "fast"],
]

# Create sample tokenized translation corpus
translation_corpus = [
    ["the", "cat", "is", "on", "mat"],
    ["a", "dog", "runs", "quickly"],
]

ARGS = {
    "reference_corpus": reference_corpus,
    "translation_corpus": translation_corpus,
    "max_order": 4,
    "use_bp": 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 float32, 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
default_transformer_poswise_net_hparams Ran
ran on a synthesized input
VegB/VLN-Transformer/texar/torch/modules/encoders/transformer_encoder.py
code served (permissive licence) · get_code("39c364923a957e63")
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 dict. 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)

ARGS = {
    "input_dim": 64,
    "output_dim": 512,
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_features Ran
ran on a synthesized input
VegB/VLN-Transformer/touchdown/env.py
code served (permissive licence) · get_code("51021c67dc837367")
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 tempfile
import os
import numpy as np

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

# Create some dummy feature files
np.save(os.path.join(temp_dir, 'feature1.npy'), np.random.rand(10, 20))
np.save(os.path.join(temp_dir, 'feature2.npy'), np.random.rand(15, 25))
np.save(os.path.join(temp_dir, 'feature3.npy'), np.random.rand(12, 18))

ARGS = {
    'feature_store': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 7 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
clones 79 other papers
18 dated earlier, 61 later · 11 only as an assembly brick
1411.4555 (2014-11)
Show and Tell: A Neural Image Caption Generator
1612.00563 (2016-12)
Self-critical Sequence Training for Image Captioning
1706.03762 (2017-06)
Attention Is All You Need
and 76 more not named here

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

The Touchdown dataset (Chen et al., 2019) provides instructions by human annotators for navigation through New York City streets and for resolving spatial descriptions at a given location. To enable the wider research community to work effectively with the Touchdown tasks, we are publicly releasing the 29k raw Street View panoramas needed for Touchdown. We follow the process used for the StreetLearn data release (Mirowski et al., 2019) to check panoramas for personally identifiable information and blur them as necessary. These have been added to the StreetLearn dataset and can be obtained via the same process as used previously for StreetLearn. We also provide a reference implementation for both of the Touchdown tasks: vision and language navigation (VLN) and spatial description resolution (SDR). We compare our model results to those given in Chen et al. (2019) and show that the panoramas we have added to StreetLearn fully support both Touchdown tasks and can be used effectively for further research and comparison.

For agents

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

get_harvested_code_for_paper("2001.03671")
get_code_for_paper("2001.03671")
have("2001.03671")

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

curl https://syntology.ai/api/ran/2001.03671.json

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

[![Syntology run record](https://syntology.ai/api/ran/2001.03671.svg)](https://syntology.ai/paper/2001.03671)

Connect an agent — have() is free.