SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2408.14368 · 2024

GR-MG: Leveraging Partially Annotated Data via Multi-Modal Goal-Conditioned Policy

arXiv · PDF · Open in the Atlas

Code that ran

8 of 10 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; 3 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 3 failures of our instrument; 2 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/2408.14368.json · badge: /api/ran/2408.14368.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
bytedance/GR-MG canonical (classifier label)
classifier basis: census
8 ran · 2 no recorded run
FunctionStatusWhere it lives
alpha2rotm Ran
ran, honoured the contract we drafted
bytedance/GR-MG/evaluate/utils/utils.py
code served (permissive licence) · get_code("5fb220ab292f1b53")
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 (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 numpy as np

# The function alpha2rotm takes a single parameter 'a' which is used in trigonometric operations
# (np.cos(a) and np.sin(a)), so it expects a scalar angle value in radians

ARGS = {
    'a': 0.5  # A scalar angle in radians
}
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 (3, 3) — shape (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
beta2rotm Ran
ran, honoured the contract we drafted
bytedance/GR-MG/evaluate/utils/utils.py
code served (permissive licence) · get_code("73992eb6c4ba4d3a")
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 (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 numpy as np

# The function beta2rotm takes a single parameter 'b' which is a beta euler angle
# It performs trigonometric operations on it (np.cos, np.sin)
# So 'b' should be a numeric value (float, int, or numpy scalar)

ARGS = {
    'b': 0.5  # A simple float representing the beta euler angle in radians
}
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 (3, 3) — shape (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
count_success Ran
ran; our contract draft was wrong, not the code
bytedance/GR-MG/evaluate/utils/utils.py
code served (permissive licence) · get_code("91e0e52c98282d15")
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.

from collections import Counter

torch_seed = 0
import torch
torch.manual_seed(torch_seed)

# Analyze the function:
# - count_success takes 'results' parameter
# - results is used with Counter(results) - so it needs to be iterable
# - the counter is indexed with integers from range(i, 6) where i goes from 1 to 5
# - so results should contain integers in range 0-5
# - len(results) is called, so it needs __len__
# - the function loops through results to count successes at different steps

# Create a valid results list with integers representing success counts at different steps
# Values should be in range 0-5 to match the logic in the loop
results = [1, 2, 3, 4, 5, 2, 3, 4, 5, 1]

ARGS = {
    "results": results
}
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
get_1d_sincos_pos_embed_from_grid Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
bytedance/GR-MG/policy/model/vision_transformer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from Zhangwenyao1/DreamVLA, is pointer only (licence: NONE) · get_code("5c8ea92d43ead77d")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from Zhangwenyao1/DreamVLA, 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.

get_2d_sincos_pos_embed Ran
ran, honoured the contract we drafted
bytedance/GR-MG/policy/model/vision_transformer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from oliverrensu/d-igpt, is pointer only (licence: NONE) · get_code("5a37690a13a6157d")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from oliverrensu/d-igpt, 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.

reduce_mean Ran
ran, violated the contract we drafted
bytedance/GR-MG/goal_gen/utils/dist_train.py
code served (permissive licence) · get_code("31528b0c209c6069")
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 1, dtype family float, shape (4,) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 os

torch.manual_seed(0)

# Case 1: tensor input (most common case)
tensor_input = torch.tensor([1.0, 2.0, 3.0, 4.0])

# Case 2: scalar input
scalar_input = 5.0

# Case 3: nprocs parameter
nprocs_value = 2

ARGS = {
    'tensor': tensor_input,
    'nprocs': nprocs_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 1, dtype family float, shape (4,) — shape (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
reduce_sum Ran
ran, violated the contract we drafted
bytedance/GR-MG/goal_gen/utils/dist_train.py
code served (permissive licence) · get_code("d8f7af433a3feb75")
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 1, dtype family float, shape (4,) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 os
import torch
import torch.distributed as dist

torch.manual_seed(0)

# Set up non-distributed environment (is_dist() will return False)
if 'WORLD_SIZE' in os.environ:
    del os.environ['WORLD_SIZE']

# Create test tensors
tensor_arg = torch.tensor([1.5, 2.5, 3.5, 4.5])
scalar_arg = 5.0

ARGS = {
    'tensor': tensor_arg
}
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 1, dtype family float, shape (4,) — shape (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
reparameterize Ran
ran; our contract draft was wrong, not the code
bytedance/GR-MG/policy/model/model.py
code served (permissive licence) · get_code("7546ec17a25b5f91")
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
from torch.autograd import Variable

torch.manual_seed(0)

# Analyze the function:
# - mu: used with std * eps, so needs to be a tensor
# - logvar: used with .div(2).exp(), so needs to be a tensor with these methods
# Both should be compatible for element-wise operations

# Create small tensors with distinct dimensions
batch_size = 2
latent_dim = 4

mu = torch.randn(batch_size, latent_dim)
logvar = torch.randn(batch_size, latent_dim)

ARGS = {
    "mu": mu,
    "logvar": logvar,
}
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
get_2d_sincos_pos_embed_from_grid Not yet run
unverified (harvested, no recorded run)
bytedance/GR-MG/policy/model/vision_transformer.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from Zhangwenyao1/DreamVLA, is pointer only (licence: NONE) · get_code("2884deadbb637c3e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from Zhangwenyao1/DreamVLA, 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. No invocation recorded.

load_tf_weights_in_gpt2 Not yet run
unverified (harvested, no recorded run)
bytedance/GR-MG/policy/model/gpt2.py
code served (permissive licence) · get_code("3665fea6427d655d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 7 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 2 the same month, 6 later) and 3 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
get_2d_sincos_pos_embed_from_grid 190 other papers
104 dated earlier, 6 the same month, 79 later, 1 same year, order not determinable · 65 only as an assembly brick
1807.10221 (2018-07)
Unified Perceptual Parsing for Scene Understanding
2105.12723 (2021-05)
Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
2111.06377 (2021-11)
Masked Autoencoders Are Scalable Vision Learners
and 187 more not named here
get_2d_sincos_pos_embed 95 other papers
62 dated earlier, 3 the same month, 29 later, 1 same year, order not determinable · 33 only as an assembly brick
1807.10221 (2018-07)
Unified Perceptual Parsing for Scene Understanding
2105.12723 (2021-05)
Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
2111.06377 (2021-11)
Masked Autoencoders Are Scalable Vision Learners
and 92 more not named here
get_1d_sincos_pos_embed_from_grid 25 other papers
13 dated earlier, 12 later · 16 only as an assembly brick
2010.11929 (2020-10)
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
2111.06377 (2021-11)
Masked Autoencoders Are Scalable Vision Learners
2301.01296 (2023-01)
TinyMIM: An Empirical Study of Distilling MIM Pre-trained Models
and 22 more not named here
load_tf_weights_in_gpt2 11 other papers
7 dated earlier, 4 later
2310.09753 (2023-10)
When can transformers reason with abstract symbols?
2312.16682 (2023-12)
Some things are more CRINGE than others: Iterative Preference Optimization with the Pairwise Cringe Loss
2402.04852 (2024-02)
Multi-Patch Prediction: Adapting LLMs for Time Series Representation Learning
and 8 more not named here
count_success 2 other papers
2 dated earlier
2112.03227 (2021-12)
CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks
2407.05996 (2024-07)
Multimodal Diffusion Transformer: Learning Versatile Behavior from Multimodal Goals
alpha2rotm 1 other paper
1 later
2506.03079 (2025-06)
ORV: 4D Occupancy-centric Robot Video Generation
beta2rotm 1 other paper
1 later
2506.03079 (2025-06)
ORV: 4D Occupancy-centric Robot Video Generation

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

The robotics community has consistently aimed to achieve generalizable robot manipulation with flexible natural language instructions. One primary challenge is that obtaining robot trajectories fully annotated with both actions and texts is time-consuming and labor-intensive. However, partially-annotated data, such as human activity videos without action labels and robot trajectories without text labels, are much easier to collect. Can we leverage these data to enhance the generalization capabilities of robots? In this paper, we propose GR-MG, a novel method which supports conditioning on a text instruction and a goal image. During training, GR-MG samples goal images from trajectories and conditions on both the text and the goal image or solely on the image when text is not available. During inference, where only the text is provided, GR-MG generates the goal image via a diffusion-based image-editing model and conditions on both the text and the generated image. This approach enables GR-MG to leverage large amounts of partially-annotated data while still using languages to flexibly specify tasks. To generate accurate goal images, we propose a novel progress-guided goal image generation model which injects task progress information into the generation process. In simulation experiments, GR-MG improves the average number of tasks completed in a row of 5 from 3.35 to 4.04. In real-robot experiments, GR-MG is able to perform 58 different tasks and improves the success rate from 68.7\% to 78.1\% and 44.4\% to 60.6\% in simple and generalization settings, respectively. It also outperforms comparing baseline methods in few-shot learning of novel skills. Video demos, code, and checkpoints are available on the project page: https://gr-mg.github.io/.

For agents

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

get_harvested_code_for_paper("2408.14368")
get_code_for_paper("2408.14368")
have("2408.14368")

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

curl https://syntology.ai/api/ran/2408.14368.json

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

[![Syntology run record](https://syntology.ai/api/ran/2408.14368.svg)](https://syntology.ai/paper/2408.14368)

Connect an agent — have() is free.