SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.14024 · ICCV · 2021

PlenOctrees for Real-time Rendering of Neural Radiance Fields

U Berkeley, Hao Li, Ruilong Li, Angjoo Kanazawa, Ren Ng, Alex Yu, Matthew Tancik

arXiv · PDF · Open in the Atlas

Code that ran

10 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 6 have no contract verdict; 4 failures of our instrument; 3 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/2103.14024.json · badge: /api/ran/2103.14024.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
sxyu/plenoctree canonical (classifier label)
classifier basis: census
5 ran · 2 no recorded run
suoxinkey/PlenOctrees_NeRF-SH community
Papers with Code archive
5 ran · 1 no recorded run
FunctionStatusWhere it lives
ComputeSH Ran
ran; our contract draft was wrong, not the code
suoxinkey/PlenOctrees_NeRF-SH/modeling/spacenet.py
code served (permissive licence) · get_code("961e4858abd50485")
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 2, dtype family float, shape (2, 25) 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 dirs to be a tensor of shape (b, 3)
# where b is the batch size and each row contains [x, y, z] direction components
# The function normalizes these to unit vectors conceptually and computes spherical harmonics

# Create a small batch of direction vectors
batch_size = 2
dirs = torch.randn(batch_size, 3)

ARGS = {
    "dirs": dirs
}
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 2, dtype family float, shape (2, 25) — shape (2, 25) 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
DenseLayer Ran
ran on a synthesized input · constructed an object; no result was computed
suoxinkey/PlenOctrees_NeRF-SH/modeling/spacenet.py
code served (permissive licence) · get_code("118360ca96b82409")
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, DenseLayer: the run constructed the object and did not call it, so no result was computed. 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 import nn

torch.manual_seed(0)

ARGS = {
    'in_dim': 4,
    'out_dim': 3,
    'activation': 'relu'
}
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 DenseLayer — 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 DenseLayer 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
SpaceNet Ran
ran on a synthesized input · constructed an object; no result was computed
suoxinkey/PlenOctrees_NeRF-SH/modeling/spacenet.py
code served (permissive licence) · get_code("985be58f00d5d5dc")
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, SpaceNet: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'c_pos': 3,
    'include_input': True,
    'use_sh': True,
}
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 SpaceNet — 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 SpaceNet 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
Trigonometric_kernel Ran
ran on a synthesized input · constructed an object; no result was computed
suoxinkey/PlenOctrees_NeRF-SH/modeling/spacenet.py
code served (permissive licence) · get_code("5c0eb0bb07f7c4d5")
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, Trigonometric_kernel: the run constructed the object and did not call it, so no result was computed. 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 import nn

torch.manual_seed(0)

ARGS = {
    'L': 10,
    'include_input': True,
}
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 Trigonometric_kernel — 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 Trigonometric_kernel 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
cast_rays Ran
ran on a synthesized input
sxyu/plenoctree/nerf_sh/nerf/model_utils.py
code served (permissive licence) · get_code("63673f823b637957")
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, 5, 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 torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - z_vals: indexed with [Ellipsis, None] so needs to be a tensor with batch dimensions
# - origins: indexed with [Ellipsis, None, :] so needs shape (..., feature_dim)
# - directions: indexed with [Ellipsis, None, :] so needs shape (..., feature_dim)
# The function casts rays by: origin + z * direction for each z value

# Create small tensors with distinct sizes for different dimensions
batch_size = 2
num_rays = 3
num_samples = 5
feature_dim = 4

# z_vals: shape (batch_size, num_rays, num_samples)
z_vals = torch.linspace(0.1, 5.0, batch_size * num_rays * num_samples).reshape(batch_size, num_rays, num_samples)

# origins: shape (batch_size, num_rays, feature_dim)
origins = torch.randn(batch_size, num_rays, feature_dim)

# directions: shape (batch_size, num_rays, feature_dim)
directions = torch.randn(batch_size, num_rays, feature_dim)

ARGS = {
    "z_vals": z_vals,
    "origins": origins,
    "directions": directions
}
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, 5, 4) — shape (2, 3, 5, 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
dense_layer Ran
ran; our contract draft was wrong, not the code
sxyu/plenoctree/octree/nerf/model_utils.py
code served (permissive licence) · get_code("6a06392b5d8a30ca")
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 Linear. 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

torch.manual_seed(0)

ARGS = {
    "in_features": 4,
    "out_features": 3,
}
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 Linear — a Linear output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Linear 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
get_embedder Ran
ran; our contract draft was wrong, not the code
suoxinkey/PlenOctrees_NeRF-SH/modeling/spacenet.py
code served (permissive licence) · get_code("4600902bb4439578")
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 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 torch
from torch import nn

torch.manual_seed(0)

ARGS = {
    'multires': 4,
    'i': 0,
    'include_input': True
}
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, ndim 1, 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
posenc Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
sxyu/plenoctree/nerf_sh/nerf/model_utils.py
code served (permissive licence) · get_code("c527f39512491ca6")
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 ArrayImpl, ndim 2, dtype family float, shape (2, 28) 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 jax.numpy as jnp

# Create test input x - shape (batch=2, features=4), values in [-pi, pi]
x = jnp.array([
    [-3.0, -1.5, 0.0, 1.5],
    [3.0, 1.0, -0.5, 2.0]
], dtype=jnp.float32)

ARGS = {
    "x": x,
    "min_deg": 0,
    "max_deg": 3,
    "legacy_posenc_order": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 2, dtype family float, shape (2, 28) — shape (2, 28) 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
posenc Ran
ran on a synthesized input
sxyu/plenoctree/octree/nerf/model_utils.py
code served (permissive licence) · get_code("85c8efa3dca48076")
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 2, dtype family float, shape (2, 28) 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 small tensor with shape [batch=2, features=4]
# x should be in [-pi, pi] according to the docstring
x = torch.randn(2, 4) * math.pi

# min_deg and max_deg define the range of encoding scales
min_deg = 0
max_deg = 3

# legacy_posenc_order controls the ordering of the encoding
legacy_posenc_order = False

ARGS = {
    "x": x,
    "min_deg": min_deg,
    "max_deg": max_deg,
    "legacy_posenc_order": legacy_posenc_order,
}
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 2, dtype family float, shape (2, 28) — shape (2, 28) 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
sample_along_rays Ran
ran on a synthesized input
sxyu/plenoctree/nerf_sh/nerf/model_utils.py
code served (permissive licence) · get_code("f126e5167d10b39d")
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 jax
import jax.numpy as jnp
from jax import random

# Set seed for reproducibility
key = random.PRNGKey(0)

# Define small test arguments
batch_size = 2
num_samples = 4
ray_dim = 3

# Create ray origins: [batch_size, 3]
origins = jnp.array([
    [0.0, 0.0, 0.0],
    [1.0, 1.0, 1.0]
], dtype=jnp.float32)

# Create ray directions: [batch_size, 3]
directions = jnp.array([
    [0.0, 0.0, 1.0],
    [0.707, 0.0, 0.707]
], dtype=jnp.float32)

# Define clipping planes
near = 0.1
far = 10.0

# Number of samples along each ray
num_samples_val = 4

# Randomization flags
randomized = True
lindisp = False

ARGS = {
    "key": key,
    "origins": origins,
    "directions": directions,
    "num_samples": num_samples_val,
    "near": near,
    "far": far,
    "randomized": randomized,
    "lindisp": lindisp,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — 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
auto_scale Not yet run
unverified (harvested, no recorded run)
sxyu/plenoctree/octree/extraction.py
code served (permissive licence) · get_code("73b45a7ea6ac09e9")
How it ran

no invocation recorded.

computeRGB Not yet run
unverified (harvested, no recorded run)
suoxinkey/PlenOctrees_NeRF-SH/modeling/spacenet.py
code served (permissive licence) · get_code("e51ca9b39922f7d2")
How it ran

no invocation recorded.

restore_model_state Not yet run
unverified (harvested, no recorded run)
sxyu/plenoctree/octree/nerf/models.py
code served (permissive licence) · get_code("f7d4878a6f883a8d")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 10 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
DenseLayer
brick of this paper's assembly: model
1 other paper
1 later
2104.00677 (2021-04)
Putting NeRF on a Diet: Semantically Consistent Few-Shot View Synthesis
dense_layer 1 other paper
1 later
2206.08929 (2022-06)
TAVA: Template-free Animatable Volumetric Actors
posenc 1 other paper
1 later
2407.15420 (2024-07)
Local All-Pair Correspondence for Point Tracking

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

We introduce a method to render Neural Radiance Fields (NeRFs) in real time using PlenOctrees, an octree-based 3D representation which supports view-dependent effects. Our method can render 800×800 images at more than 150 FPS, which is over 3000 times faster than conventional NeRFs. We do so without sacrificing quality while preserving the ability of NeRFs to perform free-viewpoint rendering of scenes with arbitrary geometry and view-dependent effects. Real-time performance is achieved by pre-tabulating the NeRF into a PlenOctree. In order to preserve viewdependent effects such as specularities, we factorize the appearance via closed-form spherical basis functions. Specifically, we show that it is possible to train NeRFs to predict a spherical harmonic representation of radiance, removing the viewing direction as an input to the neural network. Furthermore, we show that PlenOctrees can be directly optimized to further minimize the reconstruction loss, which leads to equal or better quality compared to competing methods. Moreover, this octree optimization step can be used to reduce the training time, as we no longer need to wait for the NeRF training to converge fully. Our real-time neural rendering approach may potentially enable new applications such as 6-DOF industrial and product visualizations, as well as next generation AR/VR systems. PlenOctrees are amenable to in-browser rendering as well; please visit the project page for the interactive online demo, as well as video and code: https://alexyu. net/plenoctrees.

For agents

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

get_harvested_code_for_paper("2103.14024")
get_code_for_paper("2103.14024")
have("2103.14024")

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

curl https://syntology.ai/api/ran/2103.14024.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.14024.svg)](https://syntology.ai/paper/2103.14024)

Connect an agent — have() is free.