SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2008.08465 · 2020

CosyPose: Consistent multi-view multi-object 6D pose estimation

arXiv · PDF · Open in the Atlas

Code that ran

10 of 13 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; 2 honoured the contract we drafted, 0 violated it and 7 have no contract verdict; 1 failure of our instrument; 3 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/2008.08465.json · badge: /api/ran/2008.08465.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
princeton-vl/coupled-iterative-refinement community (classifier label)
classifier basis not recorded on the link
10 ran · 3 no recorded run
FunctionStatusWhere it lives
apply_imagespace_predictions Ran
ran on a synthesized input
princeton-vl/coupled-iterative-refinement/pose_models/efficientnet.py
code served (permissive licence) · get_code("37dd8850d4b151ee")
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 3, dtype family float, shape (2, 4, 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
import numpy as np

torch.manual_seed(0)

# Create batch size 2 with distinct dimensions
batch_size = 2

# TCO: shape (batch, 4, 4) - transformation matrix
TCO = torch.randn(batch_size, 4, 4)

# K: shape (batch, 3, 3) - camera intrinsic matrix
K = torch.randn(batch_size, 3, 3)

# vxvyvz: shape (batch, 3) - velocity/scale factors
vxvyvz = torch.randn(batch_size, 3)

# dRCO: shape (batch, 3, 3) - rotation delta
dRCO = torch.randn(batch_size, 3, 3)

ARGS = {
    "TCO": TCO,
    "K": K,
    "vxvyvz": vxvyvz,
    "dRCO": dRCO,
}
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 3, dtype family float, shape (2, 4, 4) — shape (2, 4, 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
boxes_from_uv Ran
ran on a synthesized input
princeton-vl/coupled-iterative-refinement/crops/camera_geometry.py
code served (permissive licence) · get_code("4b9267025c1f4992")
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, 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)

# The function expects uv to be a tensor with shape (..., 2)
# It performs min/max operations along dimension 1, so we need at least 2 dimensions
# The function expects batch dimension and point coordinates
# Shape: (batch_size, num_points, 2) where 2 represents (u, v) coordinates

uv = torch.randn(2, 4, 2)  # batch=2, num_points=4, coordinates=2

ARGS = {
    "uv": uv
}
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, 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
compute_rotation_matrix_from_ortho6d Ran
ran on a synthesized input
princeton-vl/coupled-iterative-refinement/pose_models/efficientnet.py
code served (permissive licence) · get_code("d76737aa2cc06279")
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 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 torch

torch.manual_seed(0)

# Create a small batch of ortho6d poses
# Shape: (batch_size, 6) where 6 represents the ortho6d representation
poses = torch.randn(2, 6, dtype=torch.float32)

# Normalize the first 3 elements to ensure valid input
poses[..., 0:3] = poses[..., 0:3] / torch.norm(poses[..., 0:3], p=2, dim=-1, keepdim=True)
poses[..., 3:6] = poses[..., 3:6] / torch.norm(poses[..., 3:6], p=2, dim=-1, keepdim=True)

ARGS = {
    "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 Tensor, 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
deepim_boxes Ran
ran on a synthesized input
princeton-vl/coupled-iterative-refinement/crops/cropping.py
code served (permissive licence) · get_code("72e2c4d465007d00")
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, 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)

# rend_center_uv: shape (..., 1, 2) - center UV coordinates
rend_center_uv = torch.randn(2, 1, 2)

# obs_boxes: N x 4 - observed boxes in [x1, y1, x2, y2] format
obs_boxes = torch.tensor([
    [10.0, 20.0, 50.0, 80.0],
    [30.0, 40.0, 100.0, 120.0]
], dtype=torch.float32)

# rend_boxes: N x 4 - rendered boxes in [x1, y1, x2, y2] format
rend_boxes = torch.tensor([
    [15.0, 25.0, 55.0, 85.0],
    [35.0, 45.0, 105.0, 125.0]
], dtype=torch.float32)

# lamb: scaling parameter
lamb = 1.4

# im_size: (height, width) tuple
im_size = (240, 320)

# clamp: whether to clamp boxes
clamp = False

ARGS = {
    "rend_center_uv": rend_center_uv,
    "obs_boxes": obs_boxes,
    "rend_boxes": rend_boxes,
    "lamb": lamb,
    "im_size": im_size,
    "clamp": clamp
}
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, 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
drop_connect Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
princeton-vl/coupled-iterative-refinement/pose_models/efficientnet_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from yoctta/multiple-attention, is pointer only (licence: NONE) · get_code("2dd5dfa94542911e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from yoctta/multiple-attention, and GitHub detected no licence in that repository 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.

keep_bop19 Ran
ran on a synthesized input
princeton-vl/coupled-iterative-refinement/datasets/bop.py
code served (permissive licence) · get_code("df7cc5396e64965b")
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 DataFrame. 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.

from pathlib import Path
import pandas as pd
import tempfile
import json

# Create a temporary directory to simulate ds_dir
temp_dir = tempfile.mkdtemp()
ds_dir = Path(temp_dir)

# Create test_targets_bop19.json
bop19_data = [
    {'im_id': 1, 'obj_id': 1, 'scene_id': 'scene_1'},
    {'im_id': 2, 'obj_id': 2, 'scene_id': 'scene_1'},
    {'im_id': 3, 'obj_id': 3, 'scene_id': 'scene_2'},
    {'im_id': 4, 'obj_id': 1, 'scene_id': 'scene_2'},
]

with open(ds_dir / 'test_targets_bop19.json', 'w') as f:
    json.dump(bop19_data, f)

# Create a frame_index DataFrame that will be merged with targets
# It needs 'scene_id' and 'view_id' columns, plus potentially other columns
frame_index = pd.DataFrame({
    'scene_id': ['scene_1', 'scene_1', 'scene_2', 'scene_2'],
    'view_id': [1, 2, 3, 4],
    'image_path': ['img1.jpg', 'img2.jpg', 'img3.jpg', 'img4.jpg'],
})

ARGS = {
    'frame_index': frame_index,
    'ds_dir': ds_dir,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
project_points_robust Ran
ran on a synthesized input
princeton-vl/coupled-iterative-refinement/crops/camera_geometry.py
code served (permissive licence) · get_code("0b09f5c918855150")
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 3, dtype family float, shape (2, 4, 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 torch

torch.manual_seed(0)

# Small batch of 3D points (batch_size=2, n_points=4, 3D coordinates)
points_3d = torch.randn(2, 4, 3)

# Camera intrinsic matrix (3x3)
K = torch.tensor([
    [500.0, 0.0, 320.0],
    [0.0, 500.0, 240.0],
    [0.0, 0.0, 1.0]
])
K = K.unsqueeze(0).expand(2, -1, -1)  # batch_size=2

# Transformation matrix from camera to object (4x4 homogeneous transform)
TCO = torch.eye(4).unsqueeze(0).expand(2, -1, -1)  # batch_size=2
TCO[0, 0, 3] = 2.0  # Add some translation in x
TCO[1, 1, 3] = 1.5  # Add some translation in y

# z_min parameter
z_min = 0.1

ARGS = {
    "points_3d": points_3d,
    "K": K,
    "TCO": TCO,
    "z_min": z_min,
}
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 3, dtype family float, shape (2, 4, 2) — shape (2, 4, 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
remap_bop_targets Ran
ran on a synthesized input
princeton-vl/coupled-iterative-refinement/datasets/bop.py
code served (permissive licence) · get_code("48f8f57abcc61932")
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 DataFrame. 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 pandas as pd
import numpy as np

# Create a small DataFrame with the required columns
data = {
    'im_id': [1, 2, 3],
    'obj_id': [42, 100, 7],
    'other_col': ['a', 'b', 'c']
}
targets_df = pd.DataFrame(data)

ARGS = {
    'targets': targets_df
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
round_filters Ran
ran, honoured the contract we drafted
princeton-vl/coupled-iterative-refinement/pose_models/efficientnet_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from yoctta/multiple-attention, is pointer only (licence: NONE) · get_code("826e2e91d6a9ec09")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from yoctta/multiple-attention, and GitHub detected no licence in that repository 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.

round_repeats Ran
ran, honoured the contract we drafted
princeton-vl/coupled-iterative-refinement/pose_models/efficientnet_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from yoctta/multiple-attention, is pointer only (licence: NONE) · get_code("4d8e794b593d34a0")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from yoctta/multiple-attention, and GitHub detected no licence in that repository 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.

create_model_detector Not yet run
unverified (harvested, no recorded run)
princeton-vl/coupled-iterative-refinement/detector/detector_models_cfg.py
code served (permissive licence) · get_code("367b68d1f210f140")
How it ran

no invocation recorded.

deepim_crops_robust Not yet run
unverified (harvested, no recorded run)
princeton-vl/coupled-iterative-refinement/crops/cropping.py
code served (permissive licence) · get_code("afd05bc8d8059b93")
How it ran

no invocation recorded.

project_points Not yet run
unverified (harvested, no recorded run)
princeton-vl/coupled-iterative-refinement/crops/camera_geometry.py
code served (permissive licence) · get_code("97302b107be98065")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 4 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 3 the same month, 4 later) and 9 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
round_filters 31 other papers
6 dated earlier, 1 the same month, 24 later · 3 only as an assembly brick
1808.08127 (2018-08)
Recalibrating Fully Convolutional Networks with Spatial and Channel 'Squeeze & Excitation' Blocks
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
1911.09070 (2019-11)
EfficientDet: Scalable and Efficient Object Detection
and 28 more not named here
round_repeats 30 other papers
6 dated earlier, 1 the same month, 23 later · 3 only as an assembly brick
1808.08127 (2018-08)
Recalibrating Fully Convolutional Networks with Spatial and Channel 'Squeeze & Excitation' Blocks
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
1911.09070 (2019-11)
EfficientDet: Scalable and Efficient Object Detection
and 27 more not named here
drop_connect 15 other papers
4 dated earlier, 1 the same month, 10 later · 2 only as an assembly brick
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
1911.09070 (2019-11)
EfficientDet: Scalable and Efficient Object Detection
1912.06218 (2019-12)
YOLACT++: Better Real-time Instance Segmentation
and 12 more not named here
apply_imagespace_predictions 1 other paper
1 later
2305.10808 (2023-05)
Manifold-Aware Self-Training for Unsupervised Domain Adaptation on Regressing 6D Object Pose

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

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

Abstract

We introduce an approach for recovering the 6D pose of multiple known objects in a scene captured by a set of input images with unknown camera viewpoints. First, we present a single-view single-object 6D pose estimation method, which we use to generate 6D object pose hypotheses. Second, we develop a robust method for matching individual 6D object pose hypotheses across different input images in order to jointly estimate camera viewpoints and 6D poses of all objects in a single consistent scene. Our approach explicitly handles object symmetries, does not require depth measurements, is robust to missing or incorrect object hypotheses, and automatically recovers the number of objects in the scene. Third, we develop a method for global scene refinement given multiple object hypotheses and their correspondences across views. This is achieved by solving an object-level bundle adjustment problem that refines the poses of cameras and objects to minimize the reprojection error in all views. We demonstrate that the proposed method, dubbed CosyPose, outperforms current state-of-the-art results for single-view and multi-view 6D object pose estimation by a large margin on two challenging benchmarks: the YCB-Video and T-LESS datasets. Code and pre-trained models are available on the project webpage https://www.di.ens.fr/willow/research/cosypose/.

For agents

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

get_harvested_code_for_paper("2008.08465")
get_code_for_paper("2008.08465")
have("2008.08465")

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

curl https://syntology.ai/api/ran/2008.08465.json

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

[![Syntology run record](https://syntology.ai/api/ran/2008.08465.svg)](https://syntology.ai/paper/2008.08465)

Connect an agent — have() is free.