SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1611.06607 · 2016

A Hierarchical Approach for Generating Descriptive Image Paragraphs

arXiv · PDF · Open in the Atlas

Code that ran

5 of 9 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, 1 violated it and 3 have no contract verdict; 1 failure of our instrument; 4 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/1611.06607.json · badge: /api/ran/1611.06607.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
InnerPeace-Wu/im2p-tensorflow community (classifier label)
classifier basis not recorded on the link
5 ran · 4 no recorded run
FunctionStatusWhere it lives
bbox_transform Ran
ran, violated the contract we drafted
InnerPeace-Wu/im2p-tensorflow/lib/fast_rcnn/bbox_transform.py
code served (permissive licence) · get_code("84dcf978d5570762")
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 (2, 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 numpy as np

# bbox_transform expects two numpy arrays of bounding boxes
# Each bbox is [x1, y1, x2, y2] format
# The function computes regression targets from example rois to ground truth rois

# Create small example ROIs (batch of 2 boxes, 4 coords each)
ex_rois = np.array([
    [10.0, 20.0, 50.0, 80.0],
    [5.0, 15.0, 45.0, 75.0]
], dtype=np.float32)

# Create small ground truth ROIs (batch of 2 boxes, 4 coords each)
gt_rois = np.array([
    [12.0, 22.0, 52.0, 82.0],
    [7.0, 17.0, 47.0, 77.0]
], dtype=np.float32)

ARGS = {
    "ex_rois": ex_rois,
    "gt_rois": gt_rois
}
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 (2, 4) — shape (2, 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
bbox_transform_inv Ran
ran on a synthesized input
InnerPeace-Wu/im2p-tensorflow/lib/fast_rcnn/bbox_transform.py
code served (permissive licence) · get_code("6b2ace662cea2a36")
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 (2, 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 numpy as np

# Set up arguments for bbox_transform_inv
# The function expects:
# - boxes: numpy array of shape (N, 4) with bounding box coordinates [x1, y1, x2, y2]
# - deltas: numpy array of shape (N, 4*K) with delta predictions for K classes

# Create small sample data
np.random.seed(0)

# boxes: 2 boxes with coordinates [x1, y1, x2, y2]
boxes = np.array([
    [10.0, 20.0, 50.0, 60.0],
    [100.0, 150.0, 200.0, 250.0]
], dtype=np.float32)

# deltas: 2 boxes, 4 delta values per box (could represent multiple classes)
# Format: [dx, dy, dw, dh] repeated for each class
deltas = np.array([
    [0.1, 0.2, -0.1, 0.05, 0.0, 0.0, 0.0, 0.0],
    [0.05, -0.1, 0.1, -0.05, 0.0, 0.0, 0.0, 0.0]
], dtype=np.float32)

ARGS = {
    "boxes": boxes,
    "deltas": deltas
}
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 (2, 8) — shape (2, 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
box2xywh Ran
ran on a synthesized input
InnerPeace-Wu/im2p-tensorflow/lib/dense_cap/vis_whtml.py
code served (permissive licence) · get_code("76b2a70a88afbf78")
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
from six.moves import xrange

torch.manual_seed(0)

# The function box2xywh takes a box parameter
# It accesses box[:2] and box[i + 2] and box[i] where i is 0 or 1
# This means it needs a sequence-like object with at least 4 elements
# Typically a bounding box is [x1, y1, x2, y2] format

# Create a simple box as a list
box = [10, 20, 50, 80]

ARGS = {
    'box': box
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
six, 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
clip_boxes Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
InnerPeace-Wu/im2p-tensorflow/lib/fast_rcnn/bbox_transform.py
code served (permissive licence) · get_code("d24c99284beca7a6")
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 (2, 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 numpy as np

# Create sample boxes array
# Shape: (num_boxes, num_coords) where coords are in format [x1, y1, x2, y2, ...]
# Using 2 boxes with 4 coordinates each (x1, y1, x2, y2)
boxes = np.array([
    [10.5, 20.3, 100.7, 150.2],
    [5.1, 15.8, 200.4, 300.9]
], dtype=np.float32)

# Image shape as (height, width)
im_shape = (256, 512)

ARGS = {
    "boxes": boxes,
    "im_shape": im_shape
}
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 (2, 4) — shape (2, 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
vis_whtml Ran
ran on a synthesized input
InnerPeace-Wu/im2p-tensorflow/lib/dense_cap/vis_whtml.py
code served (permissive licence) · get_code("1ea786fdfcc714c3")
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 cv2
import numpy as np
import os
import tempfile

# Create a temporary directory for saving outputs
temp_dir = tempfile.mkdtemp()

# Create a small test image (3 channels, 100x100)
im = np.random.randint(0, 255, (100, 100, 3), dtype=np.uint8)

# Create image path
im_path = '/path/to/test_image.jpg'

# Create captions list - one caption per detection
captions = ['cat', 'dog']

# Create detections array: shape (2, 5) where each row is [x1, y1, x2, y2, score]
# x1, y1 are top-left corner, x2, y2 are bottom-right corner, last column is confidence score
dets = np.array([
    [10.0, 20.0, 50.0, 60.0, 0.9],
    [30.0, 40.0, 80.0, 90.0, 0.7]
], dtype=np.float32)

# Create pre_results dict (optional parameter with default)
pre_results = {}

# Set threshold
thresh = 0.5

# Set save path
save_path = temp_dir

ARGS = {
    'im_path': im_path,
    'im': im,
    'captions': captions,
    'dets': dets,
    'pre_results': pre_results,
    'thresh': thresh,
    'save_path': save_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy, six — 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
get_imdb Not yet run
unverified (harvested, no recorded run)
InnerPeace-Wu/im2p-tensorflow/lib/datasets/factory.py
code served (permissive licence) · get_code("9f3cea1708a8a9db")
How it ran

no invocation recorded.

get_output_dir Not yet run
unverified (harvested, no recorded run)
InnerPeace-Wu/im2p-tensorflow/lib/config.py
code served (permissive licence) · get_code("4786edbc775d299d")
How it ran

no invocation recorded.

get_output_tb_dir Not yet run
unverified (harvested, no recorded run)
InnerPeace-Wu/im2p-tensorflow/lib/config.py
code served (permissive licence) · get_code("97f5f6cd7eab1c78")
How it ran

no invocation recorded.

initialize_vocabulary Not yet run
unverified (harvested, no recorded run)
InnerPeace-Wu/im2p-tensorflow/lib/pre_glove.py
code served (permissive licence) · get_code("9292d54e38364142")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 1 the same month, 3 later) and 5 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
clip_boxes 7 other papers
3 dated earlier, 1 the same month, 3 later
1506.01497 (2015-06)
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
1506.02025 (2015-06)
Spatial Transformer Networks
1609.03605 (2016-09)
Detecting Text in Natural Image with Connectionist Text Proposal Network
and 4 more not named here
bbox_transform 4 other papers
3 dated earlier, 1 later
1506.01497 (2015-06)
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
1506.02025 (2015-06)
Spatial Transformer Networks
1609.03605 (2016-09)
Detecting Text in Natural Image with Connectionist Text Proposal Network
and 1 more not named here
get_imdb 4 other papers
3 dated earlier, 1 later
1506.01497 (2015-06)
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
1506.02025 (2015-06)
Spatial Transformer Networks
1507.00448 (2015-07)
Cross Modal Distillation for Supervision Transfer
and 1 more not named here
get_output_dir 2 other papers
2 dated earlier
1506.01497 (2015-06)
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
1506.02025 (2015-06)
Spatial Transformer Networks

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

Recent progress on image captioning has made it possible to generate novel sentences describing images in natural language, but compressing an image into a single sentence can describe visual content in only coarse detail. While one new captioning approach, dense captioning, can potentially describe images in finer levels of detail by captioning many regions within an image, it in turn is unable to produce a coherent story for an image. In this paper we overcome these limitations by generating entire paragraphs for describing images, which can tell detailed, unified stories. We develop a model that decomposes both images and paragraphs into their constituent parts, detecting semantic regions in images and using a hierarchical recurrent neural network to reason about language. Linguistic analysis confirms the complexity of the paragraph generation task, and thorough experiments on a new dataset of image and paragraph pairs demonstrate the effectiveness of our approach.

For agents

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

get_harvested_code_for_paper("1611.06607")
get_code_for_paper("1611.06607")
have("1611.06607")

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

curl https://syntology.ai/api/ran/1611.06607.json

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

[![Syntology run record](https://syntology.ai/api/ran/1611.06607.svg)](https://syntology.ai/paper/1611.06607)

Connect an agent — have() is free.