SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2209.13508 · NeurIPS · 2022

Motion Transformer with Global Intention Localization and Local Movement Refinement

Dengxin Dai, Bernt Schiele, Shaoshuai Shi, Li Jiang

arXiv · PDF · Open in the Atlas

Code that ran

6 of 15 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 3 failures of our instrument; 9 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/2209.13508.json · badge: /api/ran/2209.13508.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
sshaoshuai/mtr canonical (classifier label)
classifier basis: classifier_2025
3 ran · 6 no recorded run
nikeke19/safe-mult-rl community (classifier label)
classifier basis not recorded on the link
3 ran · 3 no recorded run
FunctionStatusWhere it lives
KNNBatchMlogK Ran
ran on a synthesized input · constructed an object; no result was computed
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("bf37b15e57f20352")
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 an instance of its own class, KNNBatchMlogK: the run constructed the object and did not call it, so no result was computed. 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 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 small tensors for testing
# xyz: (n, 3) float - point cloud data
xyz = torch.randn(4, 3, dtype=torch.float32).contiguous()

# query_xyz: (m, 3) float - query points
query_xyz = torch.randn(6, 3, dtype=torch.float32).contiguous()

# batch_idxs: (n) int - batch index for each point in xyz
batch_idxs = torch.tensor([0, 0, 1, 1], dtype=torch.int32).contiguous()

# query_batch_offsets: (B+1) int - offsets for query points per batch
# B=2 batches, so 3 elements: [0, 3, 6] where last = m=6
query_batch_offsets = torch.tensor([0, 3, 6], dtype=torch.int32).contiguous()

# k: int - number of neighbors
k = 2

ARGS = {
    'ctx': None,
    'xyz': xyz,
    'query_xyz': query_xyz,
    'batch_idxs': batch_idxs,
    'query_batch_offsets': query_batch_offsets,
    'k': k,
}
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 KNNBatchMlogK — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a KNNBatchMlogK 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
build_logdir Ran
ran on a synthesized input
nikeke19/safe-mult-rl/builders/log_dir_builder.py
code served (permissive licence) · get_code("d58bf6e2a10e7518")
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.

from omegaconf import DictConfig, OmegaConf

# Create a minimal config that satisfies the function's requirements
cfg_dict = {
    'env_name': 'lunar_lander',
    'model': {
        'name': 'PPO',
        'safe_rl': {
            'safe_mult': False,
            'safe_lagrange': False,
            'advantage_mode': 'V1a'
        }
    }
}

cfg = OmegaConf.create(cfg_dict)

ARGS = {
    'cfg': cfg
}
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 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
compute_critic_targets Ran
ran on a synthesized input
nikeke19/safe-mult-rl/models/common/buffers.py
code served (permissive licence) · get_code("dd07c43e5f8af3a4")
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, ndim 3, 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 as th
import numpy as np

th.manual_seed(0)

buffer_size = 5
n_envs = 2

ARGS = {
    "last_values": th.tensor([[0.5], [0.3]], dtype=th.float32),
    "dones": np.array([0, 1], dtype=np.float32),
    "buffer_size": buffer_size,
    "values": np.random.randn(buffer_size, n_envs).astype(np.float32),
    "episode_starts": np.array([[0], [0], [1], [0], [0]], dtype=np.float32),
    "safe_mult": True,
    "rewards": np.array([[1.0, 2.0], [0.5, -0.5], [1.5, 0.8], [0.2, 0.3], [0.1, 0.4]], dtype=np.float32),
    "gamma": 0.99,
    "n_envs": n_envs,
    "gae_lambda": 0.95,
    "min_reward": -1.0,
}
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, ndim 3, dtype family float — a tuple 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
gen_sineembed_for_position Ran
ran; our contract draft was wrong, not the code
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("782add5288aad78d")
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, 256) 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 math

torch.manual_seed(0)

# Create a position tensor with shape (batch_size, num_positions, num_coords)
# For testing, use batch_size=2, num_positions=3, num_coords=2 (x, y coordinates)
pos_tensor_2d = torch.randn(2, 3, 2)

# Also create a 4D variant (x, y, w, h) for testing the other branch
pos_tensor_4d = torch.randn(2, 3, 4)

# Test with 2D positions (default case)
ARGS = {
    "pos_tensor": pos_tensor_2d,
    "hidden_dim": 256
}
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, 256) — shape (2, 3, 256) 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_batch_offsets Ran
ran; our contract draft was wrong, not the code
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("dd2ec4aa98275a9c")
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 int; no output shape is stated. 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 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

torch.manual_seed(0)

# batch_idxs should be a 1D tensor of integers indicating which batch each element belongs to
# bs is the batch size (number of batches)
# For a small example: if we have 5 elements distributed across 3 batches
# batch_idxs might be [0, 1, 0, 2, 1] meaning:
#   - element 0 belongs to batch 0
#   - element 1 belongs to batch 1
#   - element 2 belongs to batch 0
#   - element 3 belongs to batch 2
#   - element 4 belongs to batch 1

batch_idxs = torch.tensor([0, 1, 0, 2, 1], dtype=torch.int32)
bs = 3

ARGS = {
    "batch_idxs": batch_idxs,
    "bs": bs,
}
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 int — no output shape is stated: observe_signature.py could not observe it (call:AssertionError)
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
heuristic Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
nikeke19/safe-mult-rl/envs/lunar_lander.py
code served (permissive licence) · get_code("858e421dc259e10b")
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 float, shape (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 numpy as np

class MockEnv:
    def __init__(self, continuous=True):
        self.continuous = continuous

# Create a mock environment
env = MockEnv(continuous=True)

# Create a state vector with 8 elements as specified in the docstring
# s[0]: horizontal coordinate
# s[1]: vertical coordinate
# s[2]: horizontal speed
# s[3]: vertical speed
# s[4]: angle
# s[5]: angular speed
# s[6]: 1 if first leg has contact, else 0
# s[7]: 1 if second leg has contact, else 0
state = [0.1, 0.5, 0.05, -0.02, 0.0, 0.01, 0.0, 0.0]

ARGS = {
    "env": env,
    "s": state
}
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 float, shape (2,) — shape (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
MTREncoder Not yet run
unverified (harvested, no recorded run)
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("8a85662bb5ec0dc5")
How it ran

no invocation recorded.

MotionTransformer Not yet run
unverified (harvested, no recorded run)
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("665e8b6ad3555cad")
How it ran

no invocation recorded.

PointNetPolylineEncoder Not yet run
unverified (harvested, no recorded run)
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("50efe8caae662527")
How it ran

no invocation recorded.

TransformerEncoderLayer Not yet run
unverified (harvested, no recorded run)
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("7e91726a274c25bd")
How it ran

no invocation recorded.

build_context_encoder Not yet run
unverified (harvested, no recorded run)
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("3fd13269f02df5ec")
How it ran

no invocation recorded.

build_motion_decoder Not yet run
unverified (harvested, no recorded run)
sshaoshuai/mtr/mtr/models/model.py
code served (permissive licence) · get_code("2645dd2d2d34fdcb")
How it ran

no invocation recorded.

compute_advantage Not yet run
unverified (harvested, no recorded run)
nikeke19/safe-mult-rl/models/common/buffers.py
code served (permissive licence) · get_code("e1bebfaa850ab645")
How it ran

no invocation recorded.

compute_safety_critic_targets Not yet run
unverified (harvested, no recorded run)
nikeke19/safe-mult-rl/models/common/buffers.py
code served (permissive licence) · get_code("55d554f49065505f")
How it ran

no invocation recorded.

demo_heuristic_lander Not yet run
unverified (harvested, no recorded run)
nikeke19/safe-mult-rl/envs/lunar_lander.py
code served (permissive licence) · get_code("d29d29c816b2ad92")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 1 later) and 13 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
heuristic 3 other papers
2 dated earlier, 1 later
1805.08010 (2018-05)
Where Do You Think You're Going?: Inferring Beliefs about Dynamics from Behavior
2206.13326 (2022-06)
A penalisation method for batch multi-objective Bayesian optimisation with application in heat exchanger design
2212.03117 (2022-12)
Q-Pensieve: Boosting Sample Efficiency of Multi-Objective RL Through Memory Sharing of Q-Snapshots
demo_heuristic_lander 1 other paper
1 dated earlier
2206.13326 (2022-06)
A penalisation method for batch multi-objective Bayesian optimisation with application in heat exchanger design

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

Predicting multimodal future behavior of traffic participants is essential for robotic vehicles to make safe decisions. Existing works explore to directly predict future trajectories based on latent features or utilize dense goal candidates to identify agent's destinations, where the former strategy converges slowly since all motion modes are derived from the same feature while the latter strategy has efficiency issue since its performance highly relies on the density of goal candidates. In this paper, we propose the Motion TRansformer (MTR) framework that models motion prediction as the joint optimization of global intention localization and local movement refinement. Instead of using goal candidates, MTR incorporates spatial intention priors by adopting a small set of learnable motion query pairs. Each motion query pair takes charge of trajectory prediction and refinement for a specific motion mode, which stabilizes the training process and facilitates better multimodal predictions. Experiments show that MTR achieves state-of-the-art performance on both the marginal and joint motion prediction challenges, ranking 1 st on the leaderboards of Waymo Open Motion Dataset. The source code is available at https://github.com/sshaoshuai/MTR. 36th Conference on Neural Information Processing Systems (NeurIPS 2022).

For agents

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

get_harvested_code_for_paper("2209.13508")
get_code_for_paper("2209.13508")
have("2209.13508")

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

curl https://syntology.ai/api/ran/2209.13508.json

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

[![Syntology run record](https://syntology.ai/api/ran/2209.13508.svg)](https://syntology.ai/paper/2209.13508)

Connect an agent — have() is free.