SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.01911 · 2022

Grounding Language with Visual Affordances over Unstructured Data

arXiv · PDF · Open in the Atlas

Code that ran

5 of 6 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; 1 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/2210.01911.json · badge: /api/ran/2210.01911.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
mees/hulc2 canonical (classifier label)
classifier basis not recorded on the link
5 ran · 1 no recorded run
FunctionStatusWhere it lives
error_file_name Ran
ran on a synthesized input
mees/hulc2/hulc2/wrap_training.py
code served (permissive licence) · get_code("e8be9c7ab6d65b3a")
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 str. 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 sys
from pathlib import Path
from unittest.mock import MagicMock

# Mock __file__ to simulate a module location
mock_file = "/home/user/project/module.py"

# Create a temporary directory structure for LOG_DIR
# LOG_DIR = Path(__file__).absolute().parents[2]
# If __file__ is "/home/user/project/module.py", then:
# parents[0] = /home/user/project
# parents[1] = /home/user
# parents[2] = /home
LOG_DIR = Path("/home")

def error_file_name(i):
    path = LOG_DIR / f'stderr_{i}.err'
    return path.as_posix()

ARGS = {
    'i': 0
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
is_multi_gpu_training Ran
ran, violated the contract we drafted
mees/hulc2/hulc2/training.py
code served (permissive licence) · get_code("66bed539eb74bc7e")
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 bool, 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.

from omegaconf import DictConfig, ListConfig, OmegaConf

# Test case 1: int > 1 (multi-gpu)
gpus_int_multi = 2

# Test case 2: int == -1 (all gpus, multi-gpu)
gpus_int_all = -1

# Test case 3: int == 1 (single gpu)
gpus_int_single = 1

# Test case 4: str with length > 1 (multi-gpu)
gpus_str_multi = "0,1"

# Test case 5: str with length 1 (single gpu)
gpus_str_single = "0"

# Test case 6: ListConfig with length > 1 (multi-gpu)
gpus_listconfig_multi = OmegaConf.create([0, 1])

# Test case 7: ListConfig with length 1 (single gpu)
gpus_listconfig_single = OmegaConf.create([0])

# Using one representative case for ARGS
ARGS = {
    "gpus": 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
omegaconf — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, 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
out_file_name Ran
ran on a synthesized input
mees/hulc2/hulc2/wrap_training.py
code served (permissive licence) · get_code("0d846dcc84379a31")
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 str. 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
from pathlib import Path

# Mock LOG_DIR for testing purposes
LOG_DIR = Path.home() / "test_logs"

def out_file_name(i):
    path = LOG_DIR / f'stdout_{i}.out'
    return path.as_posix()

# Define ARGS with the required parameter
ARGS = {
    'i': 0
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
print_cfg Ran
ran on a synthesized input
mees/hulc2/hulc2/affordance/train_affordance.py
code served (permissive licence) · get_code("cef200d9fd3ef989")
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 DictConfig. 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
from omegaconf import OmegaConf

torch.manual_seed(0)

# The function takes a cfg parameter and:
# 1. Converts it to a container
# 2. Pops 'paths' key
# 3. Pops 'trainer' key
# 4. Creates a new OmegaConf from the modified dict
# So cfg must be an OmegaConf object with at least 'paths' and 'trainer' keys

cfg = OmegaConf.create({
    'paths': {
        'data_dir': '/data',
        'model_dir': '/models'
    },
    'trainer': {
        'epochs': 10,
        'batch_size': 32,
        'learning_rate': 0.001
    },
    'model': {
        'hidden_dim': 64,
        'num_layers': 2
    },
    'dataset': {
        'name': 'test_dataset',
        'split': 'train'
    }
})

ARGS = {
    'cfg': cfg
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
omegaconf, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DictConfig — a DictConfig output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DictConfig output the recorded ndim 1 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
setup_callbacks Ran
ran; our contract draft was wrong, not the code
mees/hulc2/hulc2/training.py
code served (permissive licence) · get_code("7282f6517d8b89bf")
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 omegaconf import DictConfig, OmegaConf
from pytorch_lightning import Callback
from pytorch_lightning.callbacks import EarlyStopping, ModelCheckpoint
import torch

torch.manual_seed(0)

# Create minimal callback configurations
# The function uses hydra.utils.instantiate on each callback config,
# so we need DictConfig objects with _target_ fields

callbacks_cfg = OmegaConf.create({
    "early_stopping": {
        "_target_": "pytorch_lightning.callbacks.EarlyStopping",
        "monitor": "val_loss",
        "patience": 3,
    },
    "model_checkpoint": {
        "_target_": "pytorch_lightning.callbacks.ModelCheckpoint",
        "monitor": "val_loss",
        "save_top_k": 1,
    }
})

ARGS = {
    "callbacks_cfg": callbacks_cfg
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
hydra, omegaconf, pytorch_lightning, 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
run_training Not yet run
unverified (harvested, no recorded run)
mees/hulc2/hulc2/wrap_training.py
code served (permissive licence) · get_code("ccfe70a95ba3e3ed")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 2 later) and 4 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
setup_callbacks 3 other papers
2 dated earlier, 1 later
2112.03227 (2021-12)
CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Tasks
2204.06252 (2022-04)
What Matters in Language Conditioned Robotic Imitation Learning over Unstructured Data
2407.05996 (2024-07)
Multimodal Diffusion Transformer: Learning Versatile Behavior from Multimodal Goals
is_multi_gpu_training 2 other papers
2 later
2307.11949 (2023-07)
HIQL: Offline Goal-Conditioned RL with Latent States as Actions
2407.05996 (2024-07)
Multimodal Diffusion Transformer: Learning Versatile Behavior from Multimodal Goals

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 works have shown that Large Language Models (LLMs) can be applied to ground natural language to a wide variety of robot skills. However, in practice, learning multi-task, language-conditioned robotic skills typically requires large-scale data collection and frequent human intervention to reset the environment or help correcting the current policies. In this work, we propose a novel approach to efficiently learn general-purpose language-conditioned robot skills from unstructured, offline and reset-free data in the real world by exploiting a self-supervised visuo-lingual affordance model, which requires annotating as little as 1% of the total data with language. We evaluate our method in extensive experiments both in simulated and real-world robotic tasks, achieving state-of-the-art performance on the challenging CALVIN benchmark and learning over 25 distinct visuomotor manipulation tasks with a single policy in the real world. We find that when paired with LLMs to break down abstract natural language instructions into subgoals via few-shot prompting, our method is capable of completing long-horizon, multi-tier tasks in the real world, while requiring an order of magnitude less data than previous approaches. Code and videos are available at http://hulc2.cs.uni-freiburg.de

For agents

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

get_harvested_code_for_paper("2210.01911")
get_code_for_paper("2210.01911")
have("2210.01911")

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

curl https://syntology.ai/api/ran/2210.01911.json

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

[![Syntology run record](https://syntology.ai/api/ran/2210.01911.svg)](https://syntology.ai/paper/2210.01911)

Connect an agent — have() is free.