SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2312.10656 · 2023

VidToMe: Video Token Merging for Zero-Shot Video Editing

arXiv · PDF · Open in the Atlas

Code that ran

12 of 16 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, 2 violated it and 0 have no contract verdict; 8 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/2312.10656.json · badge: /api/ran/2312.10656.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
lixirui142/VidToMe canonical (classifier label)
classifier basis: census
12 ran · 4 no recorded run
FunctionStatusWhere it lives
bipartite_soft_matching_randframe Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lixirui142/VidToMe/vidtome/merge.py
code served (permissive licence) · get_code("66c921d24bbf4a88")
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

torch.manual_seed(0)

# Create a small metric tensor [B, N, C]
# B=2 (batch), N=13 (total tokens), C=4 (channels)
metric = torch.randn(2, 13, 4)

# Frame number
F = 3

# Ratio of src tokens to remove
ratio = 0.5

# Number of unmerged tokens from previous ToMe
unm_pre = 1

# Random number generator
generator = torch.Generator()
generator.manual_seed(42)

# Target stride
target_stride = 2

# Align batch flag
align_batch = False

# Merge mode
merge_mode = 'replace'

ARGS = {
    'metric': metric,
    'F': F,
    'ratio': ratio,
    'unm_pre': unm_pre,
    'generator': generator,
    'target_stride': target_stride,
    'align_batch': align_batch,
    'merge_mode': merge_mode,
}
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
do_nothing Ran
ran, honoured the contract we drafted
lixirui142/VidToMe/vidtome/merge.py
code served (permissive licence) · get_code("bd041cf95d856b5d")
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, 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

torch.manual_seed(0)

ARGS = {
    "x": torch.randn(2, 4),
    "mode": None,
}
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, 4) — shape (2, 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
init_generator Ran
ran; our contract draft was wrong, not the code
lixirui142/VidToMe/vidtome/utils.py
code served (permissive licence) · get_code("5b667bce9438efa5")
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 Generator. 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

torch.manual_seed(0)

# Create a CPU device
cpu_device = torch.device('cpu')

# Create a fallback generator for edge cases
fallback_generator = torch.Generator(device='cpu').set_state(torch.get_rng_state())

ARGS = {
    'device': cpu_device,
    'fallback': fallback_generator,
}
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 Generator — a Generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Generator 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
ip2p_prepreocess Ran
ran, violated the contract we drafted
lixirui142/VidToMe/utils/controlnet_utils.py
code served (permissive licence) · get_code("517ed2f5e48e4506")
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, 4, 5) 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 torch.nn as nn
import numpy as np

torch.manual_seed(0)

# The function ip2p_prepreocess takes an image and **kwargs
# The body just returns the image, so we need a valid image tensor

# Create a small image tensor (batch=2, channels=3, height=4, width=5)
image = torch.randn(2, 3, 4, 5)

# kwargs can be empty since the function doesn't use them
ARGS = {
    "image": image
}
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, 4, 5) — shape (2, 3, 4, 5) 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
isinstance_str Ran
ran, violated the contract we drafted
lixirui142/VidToMe/vidtome/utils.py
code served (permissive licence) · get_code("cd5d5067de871c97")
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 bool, 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 torch.nn as nn

torch.manual_seed(0)

# Create a simple object to test isinstance_str
class CustomClass:
    pass

class DerivedClass(CustomClass):
    pass

# Create an instance of DerivedClass
obj = DerivedClass()

# Test with a class name that exists in the MRO
cls_name_valid = "DerivedClass"

# Test with a class name that exists in parent MRO
cls_name_parent = "CustomClass"

# Test with a class name that doesn't exist
cls_name_invalid = "NonExistentClass"

# Also test with built-in types
tensor = torch.randn(2, 4)
cls_name_tensor = "Tensor"

# Construct ARGS for the function
ARGS = {
    "x": obj,
    "cls_name": cls_name_valid
}
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 bool, 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
join_frame Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lixirui142/VidToMe/vidtome/utils.py
code served (permissive licence) · get_code("12347b72e892db7f")
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 3, dtype family float, shape (2, 12, 5) 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
from einops import rearrange

torch.manual_seed(0)

# Understanding the function:
# - join_frame takes x and fsize as parameters
# - x is rearranged from shape '(B F) N C' to 'B (F N) C'
# - fsize is the number of frames (F parameter in rearrange)
# 
# For a valid input:
# - x should be a tensor of shape (B*F, N, C) where:
#   - B is batch size
#   - F is number of frames (fsize)
#   - N is number of tokens
#   - C is channel/feature dimension
#
# Let's use distinct dimensions:
# - B = 2 (batch)
# - F = 3 (frames, passed as fsize)
# - N = 4 (tokens)
# - C = 5 (channels)

B = 2
F = 3
N = 4
C = 5

# x should have shape (B*F, N, C) = (6, 4, 5)
x = torch.randn(B * F, N, C)

ARGS = {
    "x": x,
    "fsize": F
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 12, 5) — shape (2, 12, 5) 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
load_config Ran
ran; our contract draft was wrong, not the code
lixirui142/VidToMe/utils/config_utils.py
code served (permissive licence) · get_code("81e5978bbc052c2b")
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 DictConfig. 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 argparse
from omegaconf import OmegaConf, DictConfig
import tempfile
import os

# Create a temporary directory and config files for testing
temp_dir = tempfile.mkdtemp()

# Create base config file
base_config_path = os.path.join(temp_dir, 'base_config.yaml')
with open(base_config_path, 'w') as f:
    f.write("""
generation:
  prompt: "base prompt"
model:
  hidden_size: 256
""")

# Create main config file that references base config
main_config_path = os.path.join(temp_dir, 'tea-pour.yaml')
with open(main_config_path, 'w') as f:
    f.write(f"""
base_config: {base_config_path}
generation:
  prompt: "edit prompt"
model:
  hidden_size: 512
  layers: 4
""")

# Mock sys.argv to pass the config argument
import sys
original_argv = sys.argv
sys.argv = ['test_script.py', '--config', main_config_path]

# Define ARGS - the function load_config takes print_config as parameter
ARGS = {
    'print_config': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
omegaconf — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DictConfig — a DictConfig output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DictConfig 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
load_image Ran
ran, honoured the contract we drafted
lixirui142/VidToMe/utils/utils.py
code served (permissive licence) · get_code("08704c5df640348e")
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 (1, 3, 3, 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 tempfile
from PIL import Image
import os

torch.manual_seed(0)

# Create a temporary image file
temp_dir = tempfile.gettempdir()
image_path = os.path.join(temp_dir, "test_image.png")

# Create a small test image (4x3 RGB)
test_image = Image.new('RGB', (4, 3), color=(255, 0, 0))
test_image.save(image_path)

ARGS = {
    'image_path': image_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, torch, torchvision — 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 (1, 3, 3, 4) — shape (1, 3, 3, 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
make_diffusers_tome_block Ran
ran; our contract draft was wrong, not the code
lixirui142/VidToMe/vidtome/patch.py
code served (permissive licence) · get_code("e4a3b96bb406cafa")
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 type. 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 math
from typing import Type, Dict, Any, Tuple, Callable

torch.manual_seed(0)

# Create minimal merge module with required functions
class MockMergeModule:
    @staticmethod
    def do_nothing(x):
        return x
    
    @staticmethod
    def bipartite_soft_matching_randframe(tokens, curF, merge_ratio, unm, generator, target_stride, align_batch):
        # Returns (merge_fn, unmerge_fn, dict)
        def merge_fn(x):
            return x
        def unmerge_fn(x):
            return x
        return merge_fn, unmerge_fn, {'unm_num': 0}
    
    @staticmethod
    def bipartite_soft_matching_2s(tokens, src_len, merge_ratio, align_batch, unmerge_chunk=0):
        def merge_fn(x):
            return x
        def unmerge_fn(x):
            return x
        return merge_fn, unmerge_fn, {}

merge = MockMergeModule()

# Helper functions used in compute_merge
def init_generator(device, fallback=None):
    return torch.Generator(device=device)

def join_frame(x, fsize):
    return x

def join_warper(fsize):
    def wrapper(x):
        return x
    return wrapper

def split_warper(fsize):
    def wrapper(x):
        return x
    return wrapper

def func_warper(funcs):
    def wrapper(x):
        for f in funcs:
            x = f(x)
        return x
    return wrapper

# Create a minimal block class with all required attributes and methods
class MinimalBlock(nn.Module):
    def __init__(self):
        super().__init__()
        self.use_ada_layer_norm = False
        self.use_ada_layer_norm_zero = False
        self.only_cross_attention = False
        self.global_tokens = None
        self.generator = torch.Generator()
        
        # Minimal norm layers
        self.norm1 = nn.Identity()
        self.norm2 = nn.Identity()
        self.norm3 = nn.Identity()
        
        # Minimal attention layers
        self.attn1 = nn.Identity()
        self.attn2 = nn.Identity()
        
        # Minimal feed-forward
        self.ff = nn.Identity()
        
        self._tome_info = {
            'size': (4, 4),
            'args': {
                'batch_size': 2,
                'max_downsample': 8,
                'generator': None,
                'local_merge_ratio': 0.5,
                'target_stride': 4,
                'align_batch': False,
                'merge_global': False,
                'global_rand': 0.5,
                'global_merge_ratio': 0.5,
            }
        }

ARGS = {
    'block_class': MinimalBlock
}
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 type — a type output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a type 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
mps_gather_workaround Ran
ran; our contract draft was wrong, not the code
lixirui142/VidToMe/vidtome/merge.py
code served (permissive licence) · get_code("16eb26a585c085e1")
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 3, dtype family float, shape (2, 3, 1) 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)

# Create test tensors for the function
# input: tensor to gather from
# dim: dimension along which to gather
# index: indices to gather

# Case 1: input with shape[-1] == 1 (the workaround case)
input_tensor = torch.randn(2, 4, 1)
dim = 1
index = torch.randint(0, 4, (2, 3, 1))

ARGS = {
    "input": input_tensor,
    "dim": dim,
    "index": index
}
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 3, dtype family float, shape (2, 3, 1) — shape (2, 3, 1) 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
process_frames Ran
ran; our contract draft was wrong, not the code
lixirui142/VidToMe/utils/utils.py
code served (permissive licence) · get_code("a7a7bcc32acb5c29")
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 (1, 3, 192, 384) 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
import torch
import torchvision.transforms as T

torch.manual_seed(0)

# Create a small tensor for frames with shape (C, H, W) = (3, 128, 256)
# The function expects frames to be a tensor with at least 3 dimensions
frames = torch.randint(0, 256, (3, 128, 256), dtype=torch.uint8).float()

# Height and width parameters for resizing/cropping
h = 192
w = 384

ARGS = {
    "frames": frames,
    "h": h,
    "w": w,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, torchvision — 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 (1, 3, 192, 384) — shape (1, 3, 192, 384) 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
tile_preprocess Ran
ran; our contract draft was wrong, not the code
lixirui142/VidToMe/utils/controlnet_utils.py
code served (permissive licence) · get_code("944a3cb6dbaef645")
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, 4, 5) 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 torch.nn.functional as F

torch.manual_seed(0)

# Create a small image tensor with shape (batch=2, channels=3, height=4, width=5)
# This represents a batch of 2 images with 3 channels and spatial dimensions 4x5
image = torch.randn(2, 3, 4, 5)

# resample_rate parameter - using a scale factor for interpolation
resample_rate = 1.0

# kwargs for any additional arguments (empty dict since function doesn't require any)
kwargs = {}

ARGS = {
    'image': image,
    'resample_rate': resample_rate,
    'kwargs': kwargs
}
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 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 4, 5) 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
compute_merge Not yet run
unverified (harvested, no recorded run)
lixirui142/VidToMe/vidtome/patch.py
code served (permissive licence) · get_code("bbc5db4d6781ec3d")
How it ran

no invocation recorded.

make_tome_block Not yet run
unverified (harvested, no recorded run)
lixirui142/VidToMe/vidtome/patch.py
code served (permissive licence) · get_code("6d0482841fa4703b")
How it ran

no invocation recorded.

register_attention_control Not yet run
unverified (harvested, no recorded run)
lixirui142/VidToMe/utils/pnp_utils.py
code served (permissive licence) · get_code("1a0f6cae08d3d7af")
How it ran

no invocation recorded.

register_conv_control Not yet run
unverified (harvested, no recorded run)
lixirui142/VidToMe/utils/pnp_utils.py
code served (permissive licence) · get_code("4700b1ebfdc72fef")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 1 the same month, 4 later) and 12 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
isinstance_str 10 other papers
2 dated earlier, 1 the same month, 7 later
2303.17604 (2023-03)
Token Merging for Fast Stable Diffusion
2307.10373 (2023-07)
TokenFlow: Consistent Diffusion Features for Consistent Video Editing
2312.02212 (2023-12)
Portrait Diffusion: Training-free Face Stylization with Chain-of-Painting
and 7 more not named here
do_nothing 7 other papers
2 dated earlier, 5 later · 2 only as an assembly brick
2210.09461 (2022-10)
Token Merging: Your ViT But Faster
2303.17604 (2023-03)
Token Merging for Fast Stable Diffusion
2403.14622 (2024-03)
Language Repository for Long Video Understanding
and 4 more not named here
mps_gather_workaround 7 other papers
2 dated earlier, 5 later · 2 only as an assembly brick
2210.09461 (2022-10)
Token Merging: Your ViT But Faster
2303.17604 (2023-03)
Token Merging for Fast Stable Diffusion
2403.14622 (2024-03)
Language Repository for Long Video Understanding
and 4 more not named here
init_generator 4 other papers
1 dated earlier, 3 later
2303.17604 (2023-03)
Token Merging for Fast Stable Diffusion
2403.14622 (2024-03)
Language Repository for Long Video Understanding
2025.findings-acl.294 (2025 (year only))
Language Repository for Long Video Understanding
and 1 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

Diffusion models have made significant advances in generating high-quality images, but their application to video generation has remained challenging due to the complexity of temporal motion. Zero-shot video editing offers a solution by utilizing pre-trained image diffusion models to translate source videos into new ones. Nevertheless, existing methods struggle to maintain strict temporal consistency and efficient memory consumption. In this work, we propose a novel approach to enhance temporal consistency in generated videos by merging self-attention tokens across frames. By aligning and compressing temporally redundant tokens across frames, our method improves temporal coherence and reduces memory consumption in self-attention computations. The merging strategy matches and aligns tokens according to the temporal correspondence between frames, facilitating natural temporal consistency in generated video frames. To manage the complexity of video processing, we divide videos into chunks and develop intra-chunk local token merging and inter-chunk global token merging, ensuring both short-term video continuity and long-term content consistency. Our video editing approach seamlessly extends the advancements in image editing to video editing, rendering favorable results in temporal consistency over state-of-the-art methods.

For agents

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

get_harvested_code_for_paper("2312.10656")
get_code_for_paper("2312.10656")
have("2312.10656")

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

curl https://syntology.ai/api/ran/2312.10656.json

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

[![Syntology run record](https://syntology.ai/api/ran/2312.10656.svg)](https://syntology.ai/paper/2312.10656)

Connect an agent — have() is free.