SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.07773 · CVPR · 2024

SemCity: Semantic Scene Generation with Triplane Diffusion

Kaist, Sung-Eui Yoon, Sebin Lee, Woobin Im, Jumin Lee, Changho Jo, Juhyeong Seon

arXiv · PDF · Open in the Atlas

Code that ran

18 of 22 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, 0 violated it and 3 have no contract verdict; 12 failures of our instrument; 4 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/2403.07773.json · badge: /api/ran/2403.07773.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
zoomin-lee/semcity canonical (classifier label)
classifier basis: census
14 ran · 1 no recorded run
zoomin-lee/SemCity community
Papers with Code archive
4 ran · 3 no recorded run
FunctionStatusWhere it lives
TriplaneConv Ran
ran on a synthesized input
zoomin-lee/SemCity/diffusion/unet_triplane.py
code served (permissive licence) · get_code("b3036001ee3ed95e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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 as th
import torch.nn as nn

class TriplaneConv(nn.Module):

    def __init__(self, channels, out_channels, kernel_size, padding, is_rollout=True) -> None:
        super().__init__()
        in_channels = channels * 3 if is_rollout else channels
        self.is_rollout = is_rollout
        self.conv_xy = nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding)
        self.conv_xz = nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding)
        self.conv_yz = nn.Conv2d(in_channels, out_channels, kernel_size, padding=padding)

    def forward(self, featmaps):
        tpl_xy, tpl_xz, tpl_yz = featmaps
        H, W = tpl_xy.shape[-2:]
        D = tpl_xz.shape[-1]
        if self.is_rollout:
            tpl_xy_h = th.cat([tpl_xy, th.mean(tpl_yz, dim=-1, keepdim=True).transpose(-1, -2).expand_as(tpl_xy), th.mean(tpl_xz, dim=-1, keepdim=True).expand_as(tpl_xy)], dim=1)
            tpl_xz_h = th.cat([tpl_xz, th.mean(tpl_xy, dim=-1, keepdim=True).expand_as(tpl_xz), th.mean(tpl_yz, dim=-2, keepdim=True).expand_as(tpl_xz)], dim=1)
            tpl_yz_h = th.cat([tpl_yz, th.mean(tpl_xy, dim=-2, keepdim=True).transpose(-1, -2).expand_as(tpl_yz), th.mean(tpl_xz, dim=-2, keepdim=True).expand_as(tpl_yz)], dim=1)
        else:
            tpl_xy_h = tpl_xy
            tpl_xz_h = tpl_xz
            tpl_yz_h = tpl_yz
        assert tpl_xy_h.shape[-2] == H and tpl_xy_h.shape[-1] == W
        assert tpl_xz_h.shape[-2] == H and tpl_xz_h.shape[-1] == D
        assert tpl_yz_h.shape[-2] == W and tpl_yz_h.shape[-1] == D
        if tpl_xy_h.dtype != [param.dtype for param in self.conv_xy.parameters()][0]:
            if tpl_xy_h.dtype == th.float16:
                tpl_xy_h = self.conv_xy(tpl_xy_h.float())
                tpl_xz_h = self.conv_xz(tpl_xz_h.float())
                tpl_yz_h = self.conv_yz(tpl_yz_h.float())
            else:
                tpl_xy_h = self.conv_xy(tpl_xy_h.half())
                tpl_xz_h = self.conv_xz(tpl_xz_h.half())
                tpl_yz_h = self.conv_yz(tpl_yz_h.half())
        else:
            tpl_xy_h = self.conv_xy(tpl_xy_h)
            tpl_xz_h = self.conv_xz(tpl_xz_h)
            tpl_yz_h = self.conv_yz(tpl_yz_h)
        return (tpl_xy_h, tpl_xz_h, tpl_yz_h)


INIT_ARGS = {
    'channels': 4,
    'out_channels': 8,
    'kernel_size': 3,
    'padding': 1,
    'is_rollout': True,
}

ARGS = {
    'featmaps': (
        th.randn(2, 4, 6, 5),
        th.randn(2, 4, 6, 7),
        th.randn(2, 4, 5, 7),
    ),
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
TriplaneDownsample2x Ran
ran on a synthesized input
zoomin-lee/SemCity/diffusion/unet_triplane.py
code served (permissive licence) · get_code("ff17b7bc4046558e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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 torch.nn as nn

class TriplaneDownsample2x(nn.Module):

    def __init__(self, tri_z_down, conv_down, channels=None) -> None:
        super().__init__()
        self.tri_z_down = tri_z_down
        self.conv_down = conv_down
        if conv_down:
            if self.tri_z_down:
                self.conv_xy = nn.Conv2d(channels, channels, kernel_size=3, padding=1, stride=2, padding_mode='replicate')
                self.conv_xz = nn.Conv2d(channels, channels, kernel_size=3, padding=1, stride=2, padding_mode='replicate')
                self.conv_yz = nn.Conv2d(channels, channels, kernel_size=3, padding=1, stride=2, padding_mode='replicate')
            else:
                self.conv_xy = nn.Conv2d(channels, channels, kernel_size=3, padding=1, stride=2, padding_mode='replicate')
                self.conv_xz = nn.Conv2d(channels, channels, kernel_size=3, padding=1, stride=(2, 1), padding_mode='replicate')
                self.conv_yz = nn.Conv2d(channels, channels, kernel_size=3, padding=1, stride=(2, 1), padding_mode='replicate')

    def forward(self, featmaps):
        import torch.nn.functional as F
        tpl_xy, tpl_xz, tpl_yz = featmaps
        H, W = tpl_xy.shape[-2:]
        D = tpl_xz.shape[-1]
        if self.conv_down:
            tpl_xy = self.conv_xy(tpl_xy)
            tpl_xz = self.conv_xz(tpl_xz)
            tpl_yz = self.conv_yz(tpl_yz)
        else:
            tpl_xy = F.avg_pool2d(tpl_xy, kernel_size=2, stride=2)
            if self.tri_z_down:
                tpl_xz = F.avg_pool2d(tpl_xz, kernel_size=2, stride=2)
                tpl_yz = F.avg_pool2d(tpl_yz, kernel_size=2, stride=2)
            else:
                tpl_xz = F.avg_pool2d(tpl_xz, kernel_size=(2, 1), stride=(2, 1))
                tpl_yz = F.avg_pool2d(tpl_yz, kernel_size=(2, 1), stride=(2, 1))
        return (tpl_xy, tpl_xz, tpl_yz)


INIT_ARGS = {
    'tri_z_down': True,
    'conv_down': True,
    'channels': 4
}

ARGS = {
    'featmaps': (
        torch.randn(2, 4, 8, 6),
        torch.randn(2, 4, 8, 5),
        torch.randn(2, 4, 8, 5)
    )
}

if __name__ == '__main__':
    model = TriplaneDownsample2x(**INIT_ARGS)
    output = model(**ARGS)
    print("Output shapes:")
    print(f"  tpl_xy: {output[0].shape}")
    print(f"  tpl_xz: {output[1].shape}")
    print(f"  tpl_yz: {output[2].shape}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
TriplaneUpsample2x Ran
ran on a synthesized input
zoomin-lee/SemCity/diffusion/unet_triplane.py
code served (permissive licence) · get_code("98c1fed464e645bb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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 torch.nn as nn
import torch.nn.functional as F

class TriplaneUpsample2x(nn.Module):

    def __init__(self, tri_z_down, conv_up, channels=None) -> None:
        super().__init__()
        self.tri_z_down = tri_z_down
        self.conv_up = conv_up
        if conv_up:
            if self.tri_z_down:
                self.conv_xy = nn.ConvTranspose2d(channels, channels, kernel_size=3, padding=1, output_padding=1, stride=2)
                self.conv_xz = nn.ConvTranspose2d(channels, channels, kernel_size=3, padding=1, output_padding=1, stride=2)
                self.conv_yz = nn.ConvTranspose2d(channels, channels, kernel_size=3, padding=1, output_padding=1, stride=2)
            else:
                self.conv_xy = nn.ConvTranspose2d(channels, channels, kernel_size=3, padding=1, output_padding=1, stride=2)
                self.conv_xz = nn.ConvTranspose2d(channels, channels, kernel_size=3, padding=1, output_padding=(1, 0), stride=(2, 1))
                self.conv_yz = nn.ConvTranspose2d(channels, channels, kernel_size=3, padding=1, output_padding=(1, 0), stride=(2, 1))

    def forward(self, featmaps):
        tpl_xy, tpl_xz, tpl_yz = featmaps
        H, W = tpl_xy.shape[-2:]
        D = tpl_xz.shape[-1]
        if self.conv_up:
            tpl_xy = self.conv_xy(tpl_xy)
            tpl_xz = self.conv_xz(tpl_xz)
            tpl_yz = self.conv_yz(tpl_yz)
        else:
            tpl_xy = F.interpolate(tpl_xy, scale_factor=2, mode='bilinear', align_corners=False)
            if self.tri_z_down:
                tpl_xz = F.interpolate(tpl_xz, scale_factor=2, mode='bilinear', align_corners=False)
                tpl_yz = F.interpolate(tpl_yz, scale_factor=2, mode='bilinear', align_corners=False)
            else:
                tpl_xz = F.interpolate(tpl_xz, scale_factor=(2, 1), mode='bilinear', align_corners=False)
                tpl_yz = F.interpolate(tpl_yz, scale_factor=(2, 1), mode='bilinear', align_corners=False)
        return (tpl_xy, tpl_xz, tpl_yz)


INIT_ARGS = {
    'tri_z_down': True,
    'conv_up': True,
    'channels': 4,
}

ARGS = {
    'featmaps': (
        torch.randn(2, 4, 8, 6),
        torch.randn(2, 4, 8, 5),
        torch.randn(2, 4, 7, 6),
    ),
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
approx_standard_normal_cdf Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/diffusion/losses.py
pointer only (licence: MIT); the stored copy get_code serves, from quandao10/slct, is pointer only (licence: NONE) · get_code("cfd76fd0d89574a4")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from quandao10/slct, 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.

betas_for_alpha_bar Ran
ran, honoured the contract we drafted
zoomin-lee/semcity/diffusion/gaussian_diffusion.py
pointer only (licence: MIT); the stored copy get_code serves, from NVlabs/DiffPure, is pointer only (licence: NOASSERTION) · get_code("2ab2316ac6fdd869")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from NVlabs/DiffPure, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

compute_tdf Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/dataset/kitti_dataset.py
code served (permissive licence) · get_code("87ef5d1a6cbdb58b")
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 3, dtype family float, shape (4, 5, 6) 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
from scipy.ndimage import distance_transform_edt

# Create a small 3D voxel label array
# The function expects [X, Y, Z] dimensions
# We'll create a small 4x5x6 array with some labeled and free voxels
voxel_label = np.zeros((4, 5, 6), dtype=np.int32)
# Set some voxels to 1 (occupied) to create regions of 0s (free space)
voxel_label[1:3, 1:3, 1:3] = 1
voxel_label[0, 0, 0] = 1

ARGS = {
    "voxel_label": voxel_label,
    "trunc_distance": 3.0,
    "trunc_value": -1.0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (4, 5, 6) — shape (4, 5, 6) 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
decompose_featmaps Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zoomin-lee/SemCity/diffusion/unet_triplane.py
code served (permissive licence) · get_code("2c048a12482e6f52")
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) and returned tuple, ndim 5, 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

torch.manual_seed(0)

# The function decomposes a composed feature map into three separate feature maps
# by slicing along spatial dimensions. The composed_map is indexed with [..., :H, :W]
# and [..., :H, W:] and [..., H:, :W], so it needs at least 2 spatial dimensions.
# The function also calls .transpose(-1, -2) on one of the outputs.

# Create a composed_map tensor with shape that allows slicing:
# We need dimensions for: batch, channels, height, width
# tri_size=(128, 128, 16) means H=128, W=128, D=16
# The composed_map should have spatial dimensions that accommodate:
# - feat_xy: [..., :H, :W] = [..., :128, :128]
# - feat_xz: [..., :H, W:] = [..., :128, 128:]
# - feat_yz: [..., H:, :W] = [..., 128:, :128]
# So total width should be at least 2*W=256, total height should be at least H+H=256

# For tiny tensors: batch=2, channels=4
# Height: 256 (to fit H + H where H=128)
# Width: 256 (to fit W + W where W=128)
composed_map = torch.randn(2, 4, 256, 256)

tri_size = (128, 128, 16)
transpose = True

ARGS = {
    "composed_map": composed_map,
    "tri_size": tri_size,
    "transpose": transpose,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 5, 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
discretized_gaussian_log_likelihood Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/diffusion/losses.py
pointer only (licence: MIT); the stored copy get_code serves, from quandao10/slct, is pointer only (licence: NONE) · get_code("cd33283d615fb3d7")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from quandao10/slct, 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.

flip Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/dataset/kitti_dataset.py
code served (permissive licence) · get_code("dda2d38f31cffc71")
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 4, 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

# Create sample voxel data - 3D array with shape (4, 5, 6)
voxel = np.random.rand(4, 5, 6).astype(np.float32)

# Create sample invalid mask - same shape as voxel
invalid = np.random.randint(0, 2, size=(4, 5, 6)).astype(np.bool_)

# Define arguments dictionary
ARGS = {
    "voxel": voxel,
    "invalid": invalid,
    "flip_dim": 0
}
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 tuple, ndim 4, 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
get_named_beta_schedule Ran
ran, honoured the contract we drafted
zoomin-lee/semcity/diffusion/gaussian_diffusion.py
pointer only (licence: MIT); the stored copy get_code serves, from NVlabs/DiffPure, is pointer only (licence: NOASSERTION) · get_code("a086d6286a40b889")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from NVlabs/DiffPure, and GitHub could not classify the licence of that repository (NOASSERTION). 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_param_groups_and_shapes Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/diffusion/fp16_util.py
pointer only (licence: MIT); the stored copy get_code serves, from ZhendongWang6/DIRE, is pointer only (licence: NONE) · get_code("e41367ad14ff58fd")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ZhendongWang6/DIRE, 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_query Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/dataset/kitti_dataset.py
code served (permissive licence) · get_code("9a009f7bfad56ea2")
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 numpy as np
import torch

torch.manual_seed(0)

# Create a small voxel_label array with shape (256, 256, 32)
# containing some labeled regions (1-19) and free space (0)
voxel_label = np.zeros((256, 256, 32), dtype=np.int32)

# Add some labeled voxels for classes 1-5
voxel_label[10:15, 10:15, 5:8] = 1
voxel_label[20:25, 20:25, 10:12] = 2
voxel_label[30:35, 30:35, 15:18] = 3
voxel_label[40:45, 40:45, 20:25] = 4
voxel_label[50:55, 50:55, 25:28] = 5

# Rest is free space (0) by default

ARGS = {
    "voxel_label": voxel_label,
    "num_class": 20,
    "grid_size": (256, 256, 32),
    "max_points": 400000,
}
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 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
make_master_params Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/diffusion/fp16_util.py
pointer only (licence: MIT); the stored copy get_code serves, from ZhendongWang6/DIRE, is pointer only (licence: NONE) · get_code("e20dd5102da3b050")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ZhendongWang6/DIRE, 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.

make_output_format Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/diffusion/logger.py
code served (permissive licence) · get_code("bcd8b4acab199405")
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 HumanOutputFormat. 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 os
import sys
import tempfile

# Create a temporary directory for output
temp_dir = tempfile.mkdtemp()

ARGS = {
    'format': 'log',
    'ev_dir': temp_dir,
    'log_suffix': '_test'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type HumanOutputFormat — a HumanOutputFormat output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a HumanOutputFormat 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
mask_img Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/diffusion/nn.py
code served (permissive licence) · get_code("892ae51921341521")
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, 3, 256, 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 numpy as np

torch.manual_seed(0)

# Create tiny tensors with distinct dimensions for clarity
# Batch size: 2, Channels: 3, Height: 256, Width: 256
batch_size = 2
channels = 3
height = 256
width = 256

img = torch.randn(batch_size, channels, height, width)
cond = torch.randn(batch_size, channels, height, width)

# Test with tuple mode
mode_tuple = (100, 200, 80, 160)

# Test with string mode
mode_string = 'downright'

# Overlap parameter - can be integer or string
overlap_value = 32

# H parameter - list with single element
H_value = [128]

ARGS = {
    'img': img,
    'cond': cond,
    'mode': mode_string,
    'overlap': overlap_value,
    'H': H_value
}
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 Tensor, ndim 4, dtype family float, shape (2, 3, 256, 256) — shape (2, 3, 256, 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
mpi_weighted_mean Ran
ran; our contract draft was wrong, not the code
zoomin-lee/semcity/diffusion/logger.py
pointer only (licence: MIT); the stored copy get_code serves, from HeadLiuYun/NeuroDiff, is pointer only (licence: NONE) · get_code("e515a67f7f32e76d")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from HeadLiuYun/NeuroDiff, 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.

normal_kl Ran
ran, honoured the contract we drafted
zoomin-lee/semcity/diffusion/losses.py
pointer only (licence: MIT); the stored copy get_code serves, from facebookresearch/agrol, is pointer only (licence: NOASSERTION) · get_code("cf2798b666b231ca")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from facebookresearch/agrol, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

unflatten_master_params Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zoomin-lee/semcity/diffusion/fp16_util.py
pointer only (licence: MIT); the stored copy get_code serves, from ZhendongWang6/DIRE, is pointer only (licence: NONE) · get_code("64fff1e30802b815")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from ZhendongWang6/DIRE, 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.

TriplaneResBlock Not yet run
unverified (harvested, no recorded run)
zoomin-lee/SemCity/diffusion/unet_triplane.py
code served (permissive licence) · get_code("fe0e214e83341048")
How it ran

no invocation recorded.

TriplaneUNetModel Not yet run
unverified (harvested, no recorded run)
zoomin-lee/SemCity/diffusion/unet_triplane.py
code served (permissive licence) · get_code("8e7981a79272ed82")
How it ran

no invocation recorded.

compose_featmaps Not yet run
unverified (harvested, no recorded run)
zoomin-lee/SemCity/diffusion/unet_triplane.py
code served (permissive licence) · get_code("58f62059adb7a36e")
How it ran

no invocation recorded.

profile Not yet run
unverified (harvested, no recorded run)
zoomin-lee/semcity/diffusion/logger.py
pointer only (licence: MIT); the stored copy get_code serves, from HeadLiuYun/NeuroDiff, is pointer only (licence: NONE) · get_code("0c6607473a4c4c55")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from HeadLiuYun/NeuroDiff, 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.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 22 distinct functions with an identity, 11 are also present in other papers whose code we hold (11 with at least one paper dated earlier, 11 the same month, 11 later) and 11 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
normal_kl 160 other papers
91 dated earlier, 6 the same month, 61 later, 2 same year, order not determinable · 19 only as an assembly brick
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2105.05233 (2021-05)
Diffusion Models Beat GANs on Image Synthesis
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
and 157 more not named here
approx_standard_normal_cdf 154 other papers
86 dated earlier, 6 the same month, 60 later, 2 same year, order not determinable · 17 only as an assembly brick
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2105.05233 (2021-05)
Diffusion Models Beat GANs on Image Synthesis
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
and 151 more not named here
discretized_gaussian_log_likelihood 128 other papers
82 dated earlier, 5 the same month, 40 later, 1 same year, order not determinable · 16 only as an assembly brick
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2105.05233 (2021-05)
Diffusion Models Beat GANs on Image Synthesis
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
and 125 more not named here
betas_for_alpha_bar 92 other papers
65 dated earlier, 1 the same month, 24 later, 2 same year, order not determinable · 3 only as an assembly brick
2006.11239 (2020-06)
Denoising Diffusion Probabilistic Models
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2102.09672 (2021-02)
Improved Denoising Diffusion Probabilistic Models
and 89 more not named here
make_output_format 73 other papers
57 dated earlier, 2 the same month, 13 later, 1 same year, order not determinable
1611.05763 (2016-11)
Learning to reinforcement learn
1804.06459 (2018-04)
On Learning Intrinsic Rewards for Policy Gradient Methods
1806.02473 (2018-06)
Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
and 70 more not named here
unflatten_master_params 72 other papers
42 dated earlier, 5 the same month, 24 later, 1 same year, order not determinable
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
2111.14818 (2021-11)
Blended Diffusion for Text-driven Editing of Natural Images
2112.03145 (2021-12)
Diffusion Models for Implicit Image Segmentation Ensembles
and 69 more not named here
make_master_params 70 other papers
41 dated earlier, 5 the same month, 23 later, 1 same year, order not determinable
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
2111.14818 (2021-11)
Blended Diffusion for Text-driven Editing of Natural Images
2112.03145 (2021-12)
Diffusion Models for Implicit Image Segmentation Ensembles
and 67 more not named here
get_param_groups_and_shapes 68 other papers
39 dated earlier, 5 the same month, 23 later, 1 same year, order not determinable
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
2111.14818 (2021-11)
Blended Diffusion for Text-driven Editing of Natural Images
2112.03145 (2021-12)
Diffusion Models for Implicit Image Segmentation Ensembles
and 65 more not named here
get_named_beta_schedule 63 other papers
40 dated earlier, 1 the same month, 21 later, 1 same year, order not determinable
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2108.02938 (2021-08)
ILVR: Conditioning Method for Denoising Diffusion Probabilistic Models
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
and 60 more not named here
profile 61 other papers
44 dated earlier, 2 the same month, 14 later, 1 same year, order not determinable
1611.05763 (2016-11)
Learning to reinforcement learn
1905.06750 (2019-05)
Random Expert Distillation: Imitation Learning via Expert Policy Support Estimation
1906.05841 (2019-06)
Deep Reinforcement Learning for Industrial Insertion Tasks with Visual Inputs and Natural Rewards
and 58 more not named here
mpi_weighted_mean 52 other papers
36 dated earlier, 2 the same month, 13 later, 1 same year, order not determinable
2010.02502 (2020-10)
Denoising Diffusion Implicit Models
2111.11672 (2021-11)
Few-shot Image Generation with Mixup-based Distance Learning
2111.14818 (2021-11)
Blended Diffusion for Text-driven Editing of Natural Images
and 49 more not named here

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

scene completion refinements. In experimental results, we demonstrate that our triplane diffusion model shows meaningful generation results compared with existing work in a real-outdoor dataset, SemanticKITTI. We also show our triplane manipulation facilitates seamlessly adding, removing, or modifying objects within a scene. Further, it also enables the expansion of scenes toward a city-level scale. Finally, we evaluate our method on semantic scene completion refinements where our diffusion model enhances predictions of semantic scene completion networks by learning scene distribution. Our code is available at https: //github.com/zoomin-lee/SemCity.

For agents

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

get_harvested_code_for_paper("2403.07773")
get_code_for_paper("2403.07773")
have("2403.07773")

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

curl https://syntology.ai/api/ran/2403.07773.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.07773.svg)](https://syntology.ai/paper/2403.07773)

Connect an agent — have() is free.