SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.01351 · 2023

Streaming Motion Forecasting for Autonomous Driving

arXiv · PDF · Open in the Atlas

Code that ran

4 of 8 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 0 have no contract verdict; 4 failures of our instrument; 4 no recorded run), measured 2026-11-21; 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/2310.01351.json · badge: /api/ran/2310.01351.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
ziqipang/streamingforecasting canonical (classifier label)
classifier basis: census
4 ran · 4 no recorded run
FunctionStatusWhere it lives
batch_actors Ran
ran; our contract draft was wrong, not the code
ziqipang/streamingforecasting/streaming_forecasting/forecaster/models/batching.py
code served (permissive licence) · get_code("427aa52acd7fe518")
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

torch.manual_seed(0)

# Create sample actor features
# Each actor tensor has shape: (num_actors, feat_len, feat_dim)
# We'll create 2 samples with different numbers of actors
actor_sample_1 = torch.randn(3, 5, 4)  # 3 actors, 5 timesteps, 4 features
actor_sample_2 = torch.randn(2, 5, 4)  # 2 actors, 5 timesteps, 4 features

actors = [actor_sample_1, actor_sample_2]
device = "cpu"

ARGS = {
    "actors": actors,
    "device": device,
}
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 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
batch_lanes Ran
ran; our contract draft was wrong, not the code
ziqipang/streamingforecasting/streaming_forecasting/forecaster/models/batching.py
code served (permissive licence) · get_code("c279a1970cadbc4f")
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

torch.manual_seed(0)

# Create a minimal graph structure
# Each graph item needs:
# - 'centerlines': shape (num_lanes, seg_num, 2) - coordinates of lane segments
# - 'turns': shape (num_lanes,) - turn information
# - 'intersections': shape (num_lanes,) - intersection information
# - 'controls': shape (num_lanes,) - control information

seg_num = 5  # number of segments per lane
batch_size = 2

# First graph item with 3 lanes
graph_0 = {
    'centerlines': torch.randn(3, seg_num, 2),  # 3 lanes, 5 segments, 2D coords
    'turns': torch.randn(3),
    'intersections': torch.randn(3),
    'controls': torch.randn(3)
}

# Second graph item with 2 lanes
graph_1 = {
    'centerlines': torch.randn(2, seg_num, 2),  # 2 lanes, 5 segments, 2D coords
    'turns': torch.randn(2),
    'intersections': torch.randn(2),
    'controls': torch.randn(2)
}

graph = [graph_0, graph_1]

device = torch.device('cpu')

ARGS = {
    'graph': graph,
    'device': device
}
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 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
gather_actors Ran
ran; our contract draft was wrong, not the code
ziqipang/streamingforecasting/streaming_forecasting/forecaster/models/gather.py
code served (permissive licence) · get_code("8509923d09d0ceb9")
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

torch.manual_seed(0)

# Create sample actor features
# Each actor should be a tensor that can be transposed with (1, 2)
# So shape should be (num_actors_in_sample, feature_dim, seq_len) or similar
# After transpose(1, 2), it becomes (num_actors_in_sample, seq_len, feature_dim)

# Batch of 2 samples, each with different number of actors
actor_sample_1 = torch.randn(3, 4, 5)  # 3 actors, 4 features, 5 sequence length
actor_sample_2 = torch.randn(2, 4, 5)  # 2 actors, 4 features, 5 sequence length

actors = [actor_sample_1, actor_sample_2]
device = torch.device('cpu')

ARGS = {
    'actors': actors,
    'device': device
}
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 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
gather_lanes Ran
ran; our contract draft was wrong, not the code
ziqipang/streamingforecasting/streaming_forecasting/forecaster/models/gather.py
code served (permissive licence) · get_code("9156cc9266b78145")
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)

# Create a minimal graph structure matching what gather_lanes expects
def create_graph_item(num_nodes, num_scales=2):
    """Create a single graph item with required keys"""
    item = {
        'num_nodes': num_nodes,
        'ctrs': torch.randn(num_nodes, 4),  # node centers
        'segs': torch.randn(num_nodes, 3),  # segments
        'turns': torch.randint(0, 2, (num_nodes, 5)),  # turns feature
        'controls': torch.randint(0, 2, (num_nodes, 6)),  # controls feature
        'intersects': torch.randint(0, 2, (num_nodes, 7)),  # intersects feature
        'pre': [],  # predecessors at multiple scales
        'suc': [],  # successors at multiple scales
        'left': {'u': torch.tensor([], dtype=torch.long), 'v': torch.tensor([], dtype=torch.long)},
        'right': {'u': torch.tensor([], dtype=torch.long), 'v': torch.tensor([], dtype=torch.long)},
    }
    
    # Build pre and suc structures for each scale
    for scale in range(num_scales):
        pre_dict = {
            'u': torch.randint(0, num_nodes, (num_nodes * 2,)),
            'v': torch.randint(0, num_nodes, (num_nodes * 2,)),
        }
        suc_dict = {
            'u': torch.randint(0, num_nodes, (num_nodes * 2,)),
            'v': torch.randint(0, num_nodes, (num_nodes * 2,)),
        }
        item['pre'].append(pre_dict)
        item['suc'].append(suc_dict)
    
    return item

# Create a batch of 2 graphs with different sizes
graph = [
    create_graph_item(num_nodes=3, num_scales=2),
    create_graph_item(num_nodes=4, num_scales=2),
]

device = torch.device('cpu')

ARGS = {
    'graph': graph,
    'device': device,
}
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
api_builder Not yet run
unverified (harvested, no recorded run)
ziqipang/streamingforecasting/streaming_forecasting/forecaster/api.py
code served (permissive licence) · get_code("27c2f377bfb3bcee")
How it ran

no invocation recorded.

build_dataset Not yet run
unverified (harvested, no recorded run)
ziqipang/streamingforecasting/streaming_forecasting/forecaster/builder.py
code served (permissive licence) · get_code("ff108e9c703a507f")
How it ran

no invocation recorded.

build_model Not yet run
unverified (harvested, no recorded run)
ziqipang/streamingforecasting/streaming_forecasting/forecaster/builder.py
code served (permissive licence) · get_code("38755f0f9a52f9f2")
How it ran

no invocation recorded.

builder Not yet run
unverified (harvested, no recorded run)
ziqipang/streamingforecasting/streaming_forecasting/forecaster/builder.py
code served (permissive licence) · get_code("0e6b38005f5c6023")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 8 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.

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 2026-11-21.

Repositories linked to this paper

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

Abstract

Trajectory forecasting is a widely-studied problem for autonomous navigation. However, existing benchmarks evaluate forecasting based on independent snapshots of trajectories, which are not representative of real-world applications that operate on a continuous stream of data. To bridge this gap, we introduce a benchmark that continuously queries future trajectories on streaming data and we refer to it as "streaming forecasting." Our benchmark inherently captures the disappearance and re-appearance of agents, presenting the emergent challenge of forecasting for occluded agents, which is a safety-critical problem yet overlooked by snapshot-based benchmarks. Moreover, forecasting in the context of continuous timestamps naturally asks for temporal coherence between predictions from adjacent timestamps. Based on this benchmark, we further provide solutions and analysis for streaming forecasting. We propose a plug-and-play meta-algorithm called "Predictive Streamer" that can adapt any snapshot-based forecaster into a streaming forecaster. Our algorithm estimates the states of occluded agents by propagating their positions with multi-modal trajectories, and leverages differentiable filters to ensure temporal consistency. Both occlusion reasoning and temporal coherence strategies significantly improve forecasting quality, resulting in 25% smaller endpoint errors for occluded agents and 10-20% smaller fluctuations of trajectories. Our work is intended to generate interest within the community by highlighting the importance of addressing motion forecasting in its intrinsic streaming setting. Code is available at https://github.com/ziqipang/StreamingForecasting.

For agents

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

get_harvested_code_for_paper("2310.01351")
get_code_for_paper("2310.01351")
have("2310.01351")

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

curl https://syntology.ai/api/ran/2310.01351.json

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

[![Syntology run record](https://syntology.ai/api/ran/2310.01351.svg)](https://syntology.ai/paper/2310.01351)

Connect an agent — have() is free.