SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2004.04730 · 2020

X3D: Expanding Architectures for Efficient Video Recognition

arXiv · PDF · Open in the Atlas

Code that ran

11 of 15 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; 2 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 1 failure 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/2004.04730.json · badge: /api/ran/2004.04730.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
STONE-SAMA/x3d_mindspore community (classifier label)
classifier basis not recorded on the link
5 ran · 2 no recorded run
Chianugoogidi/X3D-tf community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
kkahatapitiya/X3D-Multigrid extension (classifier label)
classifier basis: classifier_2025
2 ran · 1 no recorded run
FunctionStatusWhere it lives
check_is_number Ran
ran on a synthesized input
STONE-SAMA/x3d_mindspore/src/utils/check_param.py
code served (permissive licence) · get_code("293dab24489b41d1")
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 int, ndim 0, dtype family int, shape () 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 math
import numpy as np

# Test case 1: valid integer
arg_value_1 = 42
arg_type_1 = int
arg_name_1 = "bias"
prim_name_1 = "bias_class"

# Test case 2: valid float
arg_value_2 = 3.14
arg_type_2 = float
arg_name_2 = "weight"
prim_name_2 = "weight_layer"

# Test case 3: no optional names
arg_value_3 = 100
arg_type_3 = int

ARGS = {
    "arg_value": 42,
    "arg_type": int,
    "arg_name": "bias",
    "prim_name": "bias_class"
}
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 int, ndim 0, dtype family int, shape () — shape () 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
conv1x1x1 Ran
ran; our contract draft was wrong, not the code
kkahatapitiya/X3D-Multigrid/x3d.py
code served (permissive licence) · get_code("0a3fae1570b6c8c7")
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 Conv3d. 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_planes": 3,
    "out_planes": 64,
    "stride": 2,
}
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 Conv3d — a Conv3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv3d 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_inplanes Ran
ran, honoured the contract we drafted
kkahatapitiya/X3D-Multigrid/x3d.py
code served (permissive licence) · get_code("acfa08610335818d")
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 list, ndim 2, dtype family int; 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 get_inplanes takes a version parameter and returns a list of tuples
# based on the version key ('S', 'M', or 'XL')
# The function only uses the version parameter to index into a dictionary

ARGS = {
    'version': 'S'
}
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 list, ndim 2, dtype family int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
linear_warmup_learning_rate Ran
ran, honoured the contract we drafted
STONE-SAMA/x3d_mindspore/src/schedule/lr_schedule.py
code served (permissive licence) · get_code("d6a99ec714147ac1")
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 float, ndim 0, dtype family float, shape () 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:
# - current_step: used in arithmetic (int or float)
# - warmup_steps: used in division (int or float)
# - base_lr: used in float() conversion and subtraction (int or float)
# - init_lr: used in float() conversion and addition (int or float)

ARGS = {
    "current_step": 500,
    "warmup_steps": 1000,
    "base_lr": 0.1,
    "init_lr": 0.001,
}
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 float, ndim 0, dtype family float, shape () — shape () 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
normalize Ran
ran on a synthesized input
Chianugoogidi/X3D-tf/utils.py
code served (permissive licence) · get_code("0aca91fe0fed88cb")
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 EagerTensor, ndim 5, dtype family float, shape (2, 3, 4, 5, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tensorflow as tf
import numpy as np

tf.random.set_seed(0)

# Create a small video clip tensor
# Shape: (batch=2, frames=3, height=4, width=5, channels=3)
clips = tf.constant(np.random.randint(0, 256, size=(2, 3, 4, 5, 3), dtype=np.uint8), dtype=tf.float32)

# Mean values for RGB channels
mean = [0.485, 0.456, 0.406]

# Standard deviation values for RGB channels
std = [0.229, 0.224, 0.225]

# Normalization value (default 255 for image pixel normalization)
norm_value = 255

ARGS = {
    "clips": clips,
    "mean": mean,
    "std": std,
    "norm_value": norm_value
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 5, dtype family float, shape (2, 3, 4, 5, 3) — shape (2, 3, 4, 5, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
recur_list2tuple Ran
ran on a synthesized input
STONE-SAMA/x3d_mindspore/src/utils/config.py
code served (permissive licence) · get_code("64bc6885f46ca205")
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
import torch.nn as nn

torch.manual_seed(0)

# Construct a nested dictionary with lists at various levels
# The function recursively converts all lists to tuples
test_dict = {
    'simple_list': [1, 2, 3],
    'nested_dict': {
        'inner_list': ['a', 'b', 'c'],
        'deep_nested': {
            'another_list': [4, 5, 6]
        }
    },
    'mixed_list': [
        {'key': 'value'},
        [7, 8, 9],
        10
    ],
    'tensor_list': [
        torch.randn(2, 4),
        torch.randn(3, 4)
    ]
}

ARGS = {
    'd': test_dict
}
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
round_repeats Ran
ran on a synthesized input
Chianugoogidi/X3D-tf/utils.py
code served (permissive licence) · get_code("4ed34e164b9d3a9f")
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 int, ndim 0, dtype family int, shape () 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 math
import torch

def round_repeats(repeats, multiplier):
    """
  Round number of layers based on depth multiplier.
  Reference: https://github.com/facebookresearch/SlowFast/blob/master/slowfast/models/video_model_builder.py
  """
    multiplier = multiplier
    if not multiplier:
        return repeats
    return int(math.ceil(multiplier * repeats))

ARGS = {
    "repeats": 4,
    "multiplier": 1.5
}
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 int, ndim 0, dtype family int, shape () — shape () 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
round_width Ran
ran on a synthesized input
Chianugoogidi/X3D-tf/utils.py
code served (permissive licence) · get_code("b033c87e17ebe6b9")
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 int, ndim 0, dtype family int, shape () 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

ARGS = {
    "width": 64,
    "multiplier": 1.5,
    "min_depth": 8,
    "divisor": 8,
}
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 int, ndim 0, dtype family int, shape () — shape () 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
to_tf_example Ran
ran on a synthesized input
Chianugoogidi/X3D-tf/datasets/create_tfrecords.py
code served (permissive licence) · get_code("f040ea8252820867")
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 SequenceExample. 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 tensorflow as tf

# Create sample frames as numpy arrays
# The function expects a list of numpy arrays representing individual frames
# Each frame should be a valid image array (height, width, channels)
frames = [
    np.random.randint(0, 256, size=(32, 48, 3), dtype=np.uint8),
    np.random.randint(0, 256, size=(32, 48, 3), dtype=np.uint8),
]

# Convert to numpy array so it has a shape attribute (frames.shape[0])
frames = np.array(frames)

# Class ID is an integer representing the video label
class_id = 5

ARGS = {
    'frames': frames,
    'class_id': class_id,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SequenceExample — a SequenceExample output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SequenceExample 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
warmup_cosine_annealing_lr_v1 Ran
ran on a synthesized input
STONE-SAMA/x3d_mindspore/src/schedule/lr_schedule.py
code served (permissive licence) · get_code("1a80346f67ccb9cf")
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 (1000,) 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 math
import numpy as np

ARGS = {
    'lr': 0.1,
    'steps_per_epoch': 100,
    'warmup_epochs': 2,
    'max_epoch': 10,
    't_max': 8,
    'eta_min': 0.0
}
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 ndarray, ndim 1, dtype family float, shape (1000,) — shape (1000,) 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
warmup_step_lr Ran
ran on a synthesized input
STONE-SAMA/x3d_mindspore/src/schedule/lr_schedule.py
code served (permissive licence) · get_code("e90ba252e203e0e9")
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 (10000,) 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 collections import Counter

# Define ARGS dict with valid arguments for warmup_step_lr
ARGS = {
    "lr": 0.1,
    "lr_epochs": [30, 60],
    "steps_per_epoch": 100,
    "warmup_epochs": 5,
    "max_epoch": 100,
    "gamma": 0.1
}
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 (10000,) — shape (10000,) 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
check_number Not yet run
unverified (harvested, no recorded run)
STONE-SAMA/x3d_mindspore/src/utils/check_param.py
code served (permissive licence) · get_code("8ce6bc3f1fabc927")
How it ran

no invocation recorded.

check_number_range Not yet run
unverified (harvested, no recorded run)
STONE-SAMA/x3d_mindspore/src/utils/check_param.py
code served (permissive licence) · get_code("e5f77ac329bf7def")
How it ran

no invocation recorded.

conv3x3x3 Not yet run
unverified (harvested, no recorded run)
kkahatapitiya/X3D-Multigrid/x3d.py
code served (permissive licence) · get_code("624693b57e7dcd4a")
How it ran

no invocation recorded.

write_tfrecord Not yet run
unverified (harvested, no recorded run)
Chianugoogidi/X3D-tf/datasets/create_tfrecords.py
code served (permissive licence) · get_code("7be94293e72825f2")
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, 10 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 8 later) and 5 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
linear_warmup_learning_rate 7 other papers
4 dated earlier, 3 later
1506.01497 (2015-06)
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
1606.04797 (2016-06)
V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation
1606.06650 (2016-06)
3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation
and 4 more not named here
check_is_number 2 other papers
1 dated earlier, 1 later
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
2106.13230 (2021-06)
Video Swin Transformer
check_number 2 other papers
1 dated earlier, 1 later
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
2106.13230 (2021-06)
Video Swin Transformer
check_number_range 2 other papers
1 dated earlier, 1 later
1905.11946 (2019-05)
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
2106.13230 (2021-06)
Video Swin Transformer
get_inplanes 2 other papers
1 dated earlier, 1 later
1912.00998 (2019-12)
A Multigrid Method for Efficiently Training Video Models
2111.13675 (2021-11)
Weakly-guided Self-supervised Pretraining for Temporal Activity Detection
conv1x1x1 1 other paper
1 dated earlier
1912.00998 (2019-12)
A Multigrid Method for Efficiently Training Video Models
conv3x3x3 1 other paper
1 dated earlier
1912.00998 (2019-12)
A Multigrid Method for Efficiently Training Video Models
recur_list2tuple 1 other paper
1 later
2106.13230 (2021-06)
Video Swin Transformer
warmup_cosine_annealing_lr_v1 1 other paper
1 later
2106.13230 (2021-06)
Video Swin Transformer
warmup_step_lr 1 other paper
1 later
2106.13230 (2021-06)
Video Swin Transformer

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

This paper presents X3D, a family of efficient video networks that progressively expand a tiny 2D image classification architecture along multiple network axes, in space, time, width and depth. Inspired by feature selection methods in machine learning, a simple stepwise network expansion approach is employed that expands a single axis in each step, such that good accuracy to complexity trade-off is achieved. To expand X3D to a specific target complexity, we perform progressive forward expansion followed by backward contraction. X3D achieves state-of-the-art performance while requiring 4.8x and 5.5x fewer multiply-adds and parameters for similar accuracy as previous work. Our most surprising finding is that networks with high spatiotemporal resolution can perform well, while being extremely light in terms of network width and parameters. We report competitive accuracy at unprecedented efficiency on video classification and detection benchmarks. Code will be available at: https://github.com/facebookresearch/SlowFast

For agents

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

get_harvested_code_for_paper("2004.04730")
get_code_for_paper("2004.04730")
have("2004.04730")

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

curl https://syntology.ai/api/ran/2004.04730.json

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

[![Syntology run record](https://syntology.ai/api/ran/2004.04730.svg)](https://syntology.ai/paper/2004.04730)

Connect an agent — have() is free.