SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2212.07555 · 2022

IMos: Intent-Driven Full-Body Motion Synthesis for Human-Object Interactions

arXiv · PDF · Open in the Atlas

Code that ran

9 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; 0 honoured the contract we drafted, 0 violated it and 8 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/2212.07555.json · badge: /api/ran/2212.07555.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
anindita127/IMoS canonical (classifier label)
classifier basis not recorded on the link
9 ran · 1 no recorded run
FunctionStatusWhere it lives
calculate_accuracy Ran
ran on a synthesized input
anindita127/IMoS/src/evaluate/evaluate_statistical_metrics.py
code served (permissive licence) · get_code("1a51f5bfac79407e")
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 float64, ndim 0, 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 torch.nn as nn
import numpy as np

torch.manual_seed(0)

# Define a minimal classifier that implements predict method
class MinimalClassifier(nn.Module):
    def __init__(self, input_dim, num_labels):
        super().__init__()
        self.linear = nn.Linear(input_dim, num_labels)
    
    def predict(self, x):
        # x is expected to be a tensor
        # Returns logits and some dummy second value
        logits = self.linear(x)
        return logits, None

# Setup parameters
num_labels = 3
input_dim = 4
batch_size = 2
num_batches = 2

# Create classifier
classifier = MinimalClassifier(input_dim, num_labels)

# Create joints_3d_vec as list of batches
# Each batch is a tensor of shape (batch_size, input_dim)
joints_3d_vec = [
    torch.randn(batch_size, input_dim),
    torch.randn(batch_size, input_dim)
]

# Create gt_labels as list of one-hot encoded labels
# Each batch is a tensor of shape (batch_size, num_labels)
gt_labels = [
    torch.nn.functional.one_hot(torch.tensor([0, 1]), num_classes=num_labels).float(),
    torch.nn.functional.one_hot(torch.tensor([2, 0]), num_classes=num_labels).float()
]

ARGS = {
    "joints_3d_vec": joints_3d_vec,
    "num_labels": num_labels,
    "classifier": classifier,
    "gt_labels": gt_labels
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
calculate_activation_statistics Ran
ran on a synthesized input
anindita127/IMoS/src/evaluate/evaluate_statistical_metrics.py
code served (permissive licence) · get_code("9f7df29fb72555e6")
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 torch
import numpy as np

torch.manual_seed(0)

# The function expects activations that have:
# - .cpu() method (torch tensor)
# - .numpy() method (after .cpu())
# - Support for np.mean(axis=0)
# - Support for np.cov(rowvar=False)

# Create a small torch tensor with shape (batch_size, feature_dim)
# Using distinct sizes: 2 for batch, 4 for features
activations = torch.randn(2, 4)

ARGS = {
    "activations": activations
}
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 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
collate_fn Ran
ran; our contract draft was wrong, not the code
anindita127/IMoS/src/data_loader/dataloader.py
code served (permissive licence) · get_code("1683815ad34578ea")
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. 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 torch.utils import data

torch.manual_seed(0)

# The collate_fn function takes a single parameter 'data' which should be
# a batch of samples. Typically this is a list of samples from a dataset.
# The function simply returns the data as-is.

# Create a small batch of sample data
# Each sample could be a tuple of (input, target) or just a single tensor
batch_data = [
    (torch.randn(4, 3), torch.tensor(0)),
    (torch.randn(4, 3), torch.tensor(1))
]

ARGS = {
    "data": batch_data
}
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 — a list 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
evaluate_fid Ran
ran on a synthesized input
anindita127/IMoS/src/evaluate/evaluate_statistical_metrics.py
code served (permissive licence) · get_code("275e1779753787cd")
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 float64, ndim 0, 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 torch.nn as nn
import numpy as np

torch.manual_seed(0)
np.random.seed(0)

# Create a minimal classifier that has a predict method
class MinimalClassifier(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 3)
    
    def predict(self, x):
        # x should be a tensor
        # Returns (predictions, activations)
        activations = self.fc(x)
        predictions = torch.argmax(activations, dim=1)
        return predictions, activations

# Create minimal data
classifier = MinimalClassifier()

# Create gt_joint and out_jt as lists of tensors
# Each element should be a tensor that the classifier can process
# gt_joint has 2 items (num_labels), each is a small batch of 4-dimensional data
gt_joint = [
    torch.randn(2, 4),  # batch of 2 samples with 4 features
    torch.randn(2, 4),
]

# out_jt has same structure
out_jt = [
    torch.randn(2, 4),
    torch.randn(2, 4),
]

# gt_labels is used only in the function signature but not in the body
gt_labels = torch.tensor([0, 1])

num_labels = 2

ARGS = {
    "gt_joint": gt_joint,
    "out_jt": out_jt,
    "num_labels": num_labels,
    "classifier": classifier,
    "gt_labels": gt_labels,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
fullpose2smplx_params_aa Ran
ran on a synthesized input
anindita127/IMoS/src/models/model_utils.py
code served (permissive licence) · get_code("a08c7f93b2a233e4")
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

torch.manual_seed(0)

# The function expects:
# - pose: a tensor with shape [..., 150] (3+63+3+3+3+45+30 = 150)
# - trans: a tensor with shape [..., 3] for translation

# Create small tensors
batch_size = 2
pose = torch.randn(batch_size, 150)
trans = torch.randn(batch_size, 3)

ARGS = {
    'pose': pose,
    'trans': trans
}
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
fullpose2smplx_params_rotmat Ran
ran on a synthesized input
anindita127/IMoS/src/models/model_utils.py
code served (permissive licence) · get_code("dca39d9f42f81c7b")
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

torch.manual_seed(0)

# Construct pose tensor: batch_size=2, total_joints=55 (55 rotation matrices flattened)
# The function expects pose to have shape [batch_size, 55] for rotation matrices
# It slices: [0:1], [1:22], [22:23], [23:24], [24:25], [25:40], [40:]
pose = torch.randn(2, 55)

# Construct trans tensor: batch_size=2, translation vector (3D)
trans = torch.randn(2, 3)

ARGS = {
    'pose': pose,
    'trans': trans
}
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
masking Ran
ran on a synthesized input
anindita127/IMoS/src/models/model.py
code served (permissive licence) · get_code("c37a84e74446989e")
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 numpy as np

# The function masking takes:
# - frame_len: an integer representing the length of the frame
# - N: an integer representing a period/stride value
# 
# It creates a list of ones with length frame_len, then zeros out
# every N-th element starting from position N-1.

ARGS = {
    "frame_len": 10,
    "N": 3,
}
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 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
one_hot_vectors Ran
ran on a synthesized input
anindita127/IMoS/src/data_loader/dataset_preprocess.py
code served (permissive licence) · get_code("c1434cadd67e5025")
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 1, dtype family int, shape (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

# Construct arguments for one_hot_vectors function
word = "hello"
word_list = ["hello", "world", "python", "code"]
mapping = {"hello": 0, "world": 1, "python": 2, "code": 3}

ARGS = {
    "word": word,
    "word_list": word_list,
    "mapping": mapping
}
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 1, dtype family int, 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
smplx_loc2glob Ran
ran on a synthesized input
anindita127/IMoS/src/models/model_utils.py
code served (permissive licence) · get_code("01e3136f10300a97")
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, 55, 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)

# The function expects local_pose to be a tensor of shape (bs, num_joints * 3, 3)
# where bs is batch size, num_joints is number of joints, and the last two dims are 3x3 rotation matrices
# smplx_parents has 55 elements, so num_joints = 55
# We'll use a small batch size of 2

batch_size = 2
num_joints = 55

# Create a local_pose tensor of shape (batch_size, num_joints * 3, 3)
# This represents 55 joints, each with a 3x3 rotation matrix
local_pose = torch.randn(batch_size, num_joints * 3, 3)

ARGS = {
    "local_pose": local_pose
}
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, 55, 3, 3) — shape (2, 55, 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
one_hot_vectors Not yet run
unverified (harvested, no recorded run)
anindita127/IMoS/src/train/train_action_recognition.py
code served (permissive licence) · get_code("2c679bc46119fd48")
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, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 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
collate_fn 1 other paper
1 later
2312.03631 (2023-12)
Mitigating Open-Vocabulary Caption Hallucinations

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

Can we make virtual characters in a scene interact with their surrounding objects through simple instructions? Is it possible to synthesize such motion plausibly with a diverse set of objects and instructions? Inspired by these questions, we present the first framework to synthesize the full-body motion of virtual human characters performing specified actions with 3D objects placed within their reach. Our system takes textual instructions specifying the objects and the associated intentions of the virtual characters as input and outputs diverse sequences of full-body motions. This contrasts existing works, where full-body action synthesis methods generally do not consider object interactions, and human-object interaction methods focus mainly on synthesizing hand or finger movements for grasping objects. We accomplish our objective by designing an intent-driven fullbody motion generator, which uses a pair of decoupled conditional variational auto-regressors to learn the motion of the body parts in an autoregressive manner. We also optimize the 6-DoF pose of the objects such that they plausibly fit within the hands of the synthesized characters. We compare our proposed method with the existing methods of motion synthesis and establish a new and stronger state-of-the-art for the task of intent-driven motion synthesis.

For agents

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

get_harvested_code_for_paper("2212.07555")
get_code_for_paper("2212.07555")
have("2212.07555")

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

curl https://syntology.ai/api/ran/2212.07555.json

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

[![Syntology run record](https://syntology.ai/api/ran/2212.07555.svg)](https://syntology.ai/paper/2212.07555)

Connect an agent — have() is free.