SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.10211 · ICCV · 2021

Space-Time Crop & Attend: Improving Cross-modal Video Representation Learning

A Research, Ishan Misra, João Henriques, Yuki Asano, Mandela Patrick, Andrea Facebook, Po-Yao Huang, Florian Metze

arXiv · PDF · Open in the Atlas

Code that ran

12 of 23 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (11 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 12 have no contract verdict; 0 failures of our instrument; 11 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/2103.10211.json · badge: /api/ran/2103.10211.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
facebookresearch/GDT marked official in Papers with Code
Papers with Code archive
12 ran · 11 no recorded run
FunctionStatusWhere it lives
Base Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("9d7e4b24b5660ac5")
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, Base: 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)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
        self.norm = nn.LayerNorm(3)
    
    def forward(self, x):
        return self.norm(self.linear(x))

ARGS = {}
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 Base — 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 Base 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
BasicBlock Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("249b5820589b96d3")
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, BasicBlock: 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
import torch.nn as nn

torch.manual_seed(0)

# Create a small downsample module for testing
downsample_module = nn.Sequential(
    nn.Conv2d(4, 8, kernel_size=1, stride=2, bias=False),
    nn.BatchNorm2d(8)
)

ARGS = {
    "inplanes": 4,
    "planes": 8,
    "stride": 2,
    "downsample": downsample_module,
    "groups": 1,
    "base_width": 64,
    "dilation": 1,
    "norm_layer": nn.BatchNorm2d
}
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 BasicBlock — 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 BasicBlock 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
Bottleneck Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("346198401060aa08")
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, Bottleneck: 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
import torch.nn as nn

torch.manual_seed(0)

# Build a minimal downsample module (Sequential with Conv2d and BatchNorm2d)
downsample = nn.Sequential(
    nn.Conv2d(64, 256, kernel_size=1, stride=1, bias=False),
    nn.BatchNorm2d(256)
)

# Define the ARGS dictionary with all Bottleneck parameters
ARGS = {
    'inplanes': 64,
    'planes': 64,
    'stride': 1,
    'downsample': downsample,
    'groups': 1,
    'base_width': 64,
    'dilation': 1,
    'norm_layer': nn.BatchNorm2d
}
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 Bottleneck — 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 Bottleneck 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
Conv2Plus1D Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("4ab7412ae6131e98")
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, Conv2Plus1D: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'in_planes': 3,
    'out_planes': 64,
    'midplanes': 32,
    'stride': 1,
    'padding': 1,
}
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 Conv2Plus1D — 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 Conv2Plus1D output the recorded ndim 1 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
FFN Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("ba5523a1a6748c82")
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, FFN: 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
import torch.nn as nn
import torch.nn.functional as F

torch.manual_seed(0)

ARGS = {
    "in_dim": 4,
    "dim_hidden": 8,
    "out_dim": 3,
    "dropout": 0.1,
}
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 FFN — 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 FFN 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
MLP Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("66c25c20b741db7d")
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, MLP: 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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "n_input": 16,
    "n_classes": 5,
    "n_hidden": 512,
    "p": 0.3,
}
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 MLP — 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 MLP 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
MultiHeadAttention Ran
ran on a synthesized input
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("025f425a305c99a1")
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 numpy as np
from torch import nn

INIT_ARGS = {
    "n_heads": 2,
    "dim": 4,
    "dp": 0.1,
    "qkv_mha": False
}

# Construct input tensor: batch_size=2, sequence_length=3, dim=4
input_tensor = torch.randn(2, 3, 4)

# Construct mask tensor: batch_size=2, sequence_length=3 (broadcastable to attention scores)
mask = torch.ones(2, 3)

# For self-attention, kv is None
ARGS = {
    "input": input_tensor,
    "mask": mask,
    "kv": None,
    "cache": None,
    "head_mask": None
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — 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)
R2Plus1dStem Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("dfb98feede0d7132")
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, R2Plus1dStem: 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
import torch.nn as nn

torch.manual_seed(0)

# R2Plus1dStem.__init__ takes no parameters beyond self
# The class is a nn.Sequential that chains together several layers:
# - Conv3d(3, 45, ...)
# - BatchNorm3d(45)
# - ReLU(inplace=True)
# - Conv3d(45, 64, ...)
# - BatchNorm3d(64)
# - ReLU(inplace=True)

# Since __init__ takes no arguments (only self), ARGS should be empty
ARGS = {}
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 R2Plus1dStem — 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 R2Plus1dStem output the recorded ndim 1 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
Transformer Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("a45dddd3d20ddb5b")
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, Transformer: 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
import torch.nn as nn

torch.manual_seed(0)

# Examine Transformer.__init__ parameters:
# - hidden_dim: int
# - heads: int (default 4)
# - dropout: float (default 0.0)
# - qkv_mha: bool (default False)

# Examine Transformer.forward signature:
# - x: tensor of shape (bs, qlen, hidden_dim)
# - mask: tensor of shape (bs, qlen) or (bs, 1, qlen)

# Create minimal valid inputs
batch_size = 2
seq_len = 3
hidden_dim = 4
num_heads = 2

# Create input tensor
x = torch.randn(batch_size, seq_len, hidden_dim)

# Create mask tensor (2D: bs x qlen)
# The code expects mask values where mask == 0 means masked positions
# Use 1 for valid positions, 0 for masked positions
mask = torch.ones(batch_size, seq_len, dtype=torch.long)

ARGS = {
    "hidden_dim": hidden_dim,
    "heads": num_heads,
    "dropout": 0.0,
    "qkv_mha": False,
}
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 Transformer — 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 Transformer 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
TransformerPooling Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("784b7c6f51d0dc00")
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, TransformerPooling: 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
import torch.nn.functional as F
import math

torch.manual_seed(0)

ARGS = {
    'emb_dim': 8,
    'hidden_dim': 16,
    'num_layer': 2,
    'dp': 0.1,
    'num_head': 4,
    'all_state': False,
    'positional_emb': False,
    'num_speed': 1,
    'qkv_mha': False,
}
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 TransformerPooling — 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 TransformerPooling 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
TransposeSqueeze Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("bd97e7875c7f6d01")
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, TransposeSqueeze: 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)

# TransposeSqueeze.__init__ takes fdim and tdim as parameters
# TransposeSqueeze.forward takes x and does x.view(-1, self.fdim, self.tdim).transpose(-1, -2)
# This means x needs to have total elements divisible by (fdim * tdim)

# Let's use distinct dimension sizes:
# fdim = 4
# tdim = 3
# We need a batch of tensors where total elements = batch_size * 4 * 3 = batch_size * 12

fdim = 4
tdim = 3

# Create input tensor with shape that works with the view operation
# If we have batch_size=2, we need 2*4*3 = 24 elements total
# Input shape could be (2, 12) or (24,) or anything that flattens to 24 elements
x = torch.randn(2, 12)

ARGS = {
    "fdim": fdim,
    "tdim": tdim,
}
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 TransposeSqueeze — 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 TransposeSqueeze 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
VideoResNet Ran
ran on a synthesized input · constructed an object; no result was computed
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("1c92b909d0cdece1")
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, VideoResNet: 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
import torch.nn as nn

torch.manual_seed(0)

# Define the stem - a simple 3D conv stem that outputs 64 channels
class SimpleStem(nn.Module):
    def __init__(self):
        super(SimpleStem, self).__init__()
        self.conv = nn.Conv3d(3, 64, kernel_size=7, stride=2, padding=3, bias=False)
        self.bn = nn.BatchNorm3d(64)
        self.relu = nn.ReLU(inplace=True)
    
    def forward(self, x):
        x = self.conv(x)
        x = self.bn(x)
        x = self.relu(x)
        return x

# Define a simple conv_maker class that provides the interface needed
class ConvMaker:
    def __init__(self, temporal_stride=1):
        self.temporal_stride = temporal_stride
    
    def get_downsample_stride(self, spatial_stride):
        return (self.temporal_stride, spatial_stride, spatial_stride)

# Define a minimal block for VideoResNet that matches the interface
class VideoBottleneck(nn.Module):
    expansion = 4
    
    def __init__(self, inplanes, planes, conv_builder, stride=1, downsample=None):
        super(VideoBottleneck, self).__init__()
        self.conv1 = nn.Conv3d(inplanes, planes, kernel_size=1, stride=1, bias=False)
        self.bn1 = nn.BatchNorm3d(planes)
        self.conv2 = nn.Conv3d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=False)
        self.bn2 = nn.BatchNorm3d(planes)
        self.conv3 = nn.Conv3d(planes, planes * self.expansion, kernel_size=1, stride=1, bias=False)
        self.bn3 = nn.BatchNorm3d(planes * self.expansion)
        self.relu = nn.ReLU(inplace=True)
        self.downsample = downsample
        self.stride = stride
    
    def forward(self, x):
        identity = x
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)
        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)
        out = self.conv3(out)
        out = self.bn3(out)
        if self.downsample is not None:
            identity = self.downsample(x)
        out += identity
        out = self.relu(out)
        return out

ARGS = {
    'block': VideoBottleneck,
    'conv_makers': [ConvMaker(1), ConvMaker(1), ConvMaker(1), ConvMaker(1)],
    'layers': [3, 4, 6, 3],
    'stem': SimpleStem,
    'num_classes': 400,
    'zero_init_residual': False,
    'larger_last': False,
}
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 VideoResNet — 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 VideoResNet 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
AudioBaseNetwork Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("05373243575c5edd")
How it ran

no invocation recorded.

ResNet Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("6840fa8301c4edba")
How it ran

no invocation recorded.

Stica_TransformerFMCrop Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("85fbc7f944b570fc")
How it ran

no invocation recorded.

VideoBaseNetwork Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("cf7d2feaca7a7e0e")
How it ran

no invocation recorded.

_resnet Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("f8df10b38e4d9f0d")
How it ran

no invocation recorded.

_video_resnet Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("471f1e36d731e3a0")
How it ran

no invocation recorded.

get_audio_feature_extractor Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("cf0d617f6919c7db")
How it ran

no invocation recorded.

get_video_feature_extractor Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("dfcc255237250347")
How it ran

no invocation recorded.

r2plus1d_18 Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("9982977c79b6bef4")
How it ran

no invocation recorded.

r2plus1d_34 Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("dcfa82af3e6afc58")
How it ran

no invocation recorded.

resnet9 Not yet run
unverified (harvested, no recorded run)
facebookresearch/GDT/model.py
code served (permissive licence) · get_code("f78b8e83121feab2")
How it ran

no invocation recorded.

Also present in other papers' code

Of 23 samples lifted from this paper's linked repositories, 23 carry a function identity and 0 do not. Of the 23 distinct functions with an identity, 6 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 6 later) and 17 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
BasicBlock
brick of this paper's assembly: helper
16 other papers
2 dated earlier, 14 later · 13 only as an assembly brick
1512.03385 (2015-12)
Deep Residual Learning for Image Recognition
2003.11942 (2020-03)
Towards Backward-Compatible Representation Learning
2104.02691 (2021-04)
Localizing Visual Sounds the Hard Way
and 13 more not named here
Bottleneck
brick of this paper's assembly: helper
15 other papers
2 dated earlier, 13 later · 12 only as an assembly brick
2003.10152 (2020-03)
SOLOv2: Dynamic and Fast Instance Segmentation
2003.11942 (2020-03)
Towards Backward-Compatible Representation Learning
2104.02691 (2021-04)
Localizing Visual Sounds the Hard Way
and 12 more not named here
_resnet
brick of this paper's assembly: helper
8 other papers
8 later
2105.03245 (2021-05)
Adaptive Focus for Efficient Video Recognition
2107.02156 (2021-07)
Do Different Tracking Tasks Require Different Appearance Models?
2204.04662 (2022-04)
FOSTER: Feature Boosting and Compression for Class-Incremental Learning
and 5 more not named here
r2plus1d_18
brick of this paper's assembly: helper
3 other papers
1 dated earlier, 2 later
1711.11248 (2017-11)
A Closer Look at Spatiotemporal Convolutions for Action Recognition
136670659 (2022 (year only))
MovieCuts: A New Dataset and Benchmark for Cut Type Recognition
Ding_Motion-Aware_Contrastive_Video_Representation_Learning_via_Foreground-Background_Merging_CVPR_2022_paper (2022 (year only))
Motion-aware Contrastive Video Representation Learning via Foreground-background Merging
Conv2Plus1D
brick of this paper's assembly: helper
1 other paper
1 later
2207.12795 (2022-07)
Static and Dynamic Concepts for Self-supervised Video Representation Learning
R2Plus1dStem
brick of this paper's assembly: helper
1 other paper
1 later
2207.12795 (2022-07)
Static and Dynamic Concepts for Self-supervised Video Representation Learning

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

The quality of the image representations obtained from self-supervised learning depends strongly on the type of data augmentations used in the learning formulation. Recent papers have ported these methods from still images to videos and found that leveraging both audio and video signals yields strong gains; however, they did not find that spatial augmentations such as cropping, which are very important for still images, work as well for videos. In this paper, we improve these formulations in two ways unique to the spatio-temporal aspect of videos. First, for space, we show that spatial augmentations such as cropping do work well for videos too, but that previous implementations, due to the high processing and memory cost, could not do this at a scale sufficient for it to work well. To address this issue, we first introduce Feature Crop, a method to simulate such augmentations much more efficiently directly in feature space. Second, we show that as opposed to naïve average pooling, the use of transformer-based attention improves performance significantly, and is well suited for processing feature crops. Combining both of our discoveries into a new method, Space-Time Crop & Attend (STiCA) we achieve state-of-the-art performance across multiple video-representation learning benchmarks. In particular, we achieve new state-of-the-art accuracies of 67.0% on HMDB-51 and 93.1% on UCF-101 when pre-training on Kinetics-400. Code and pretrained models are available 1 . * Equal contribution.

For agents

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

get_harvested_code_for_paper("2103.10211")
get_code_for_paper("2103.10211")
have("2103.10211")

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

curl https://syntology.ai/api/ran/2103.10211.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.10211.svg)](https://syntology.ai/paper/2103.10211)

Connect an agent — have() is free.