SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.21705 · 2024

Tora: Trajectory-oriented Diffusion Transformer for Video Generation

arXiv · PDF · Open in the Atlas

Code that ran

9 of 12 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (4 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 1 violated it and 6 have no contract verdict; 2 failures of our instrument; 3 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2407.21705.json · badge: /api/ran/2407.21705.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
alibaba/Tora canonical (classifier label)
classifier basis: classifier_2025
9 ran · 3 no recorded run
FunctionStatusWhere it lives
Downsample Ran
ran on a synthesized input · constructed an object; no result was computed
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("6a1cf7ce76caca59")
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, Downsample: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from torch import nn

torch.manual_seed(0)

ARGS = {
    "channels": 16,
    "use_conv": True,
    "dims": 2,
    "out_channels": 32,
    "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 Downsample — 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 Downsample 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
FloatGroupNorm Ran
ran on a synthesized input
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("bd3af06052e1a872")
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 Tensor, ndim 4, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from torch import nn

class FloatGroupNorm(nn.GroupNorm):

    def forward(self, x):
        return super().forward(x.to(self.bias.dtype)).type(x.dtype)


INIT_ARGS = {
    'num_groups': 2,
    'num_channels': 4,
}

ARGS = {
    'x': torch.randn(2, 4, 3, 3),
}

if __name__ == '__main__':
    model = FloatGroupNorm(**INIT_ARGS)
    output = model(**ARGS)
    print(output.shape)
    print(output.dtype)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
MGF Ran
ran on a synthesized input
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("c6b9d6647323596a")
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 Tensor, ndim 4, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from torch import nn
from einops import rearrange, repeat

class FloatGroupNorm(nn.GroupNorm):
    def forward(self, x):
        return super().forward(x.to(self.bias.dtype)).type(x.dtype)

def zero_module(module):
    """
    Zero out the parameters of a module and return it.
    """
    for p in module.parameters():
        p.detach().zero_()
    return module

class MGF(nn.Module):
    def __init__(self, flow_in_channel=128, out_channels=1152):
        super().__init__()
        self.out_channels = out_channels
        self.flow_gamma_spatial = nn.Conv2d(flow_in_channel, self.out_channels // 4, 3, padding=1)
        self.flow_gamma_temporal = zero_module(nn.Conv1d(self.out_channels // 4, self.out_channels, kernel_size=3, stride=1, padding=1, padding_mode='replicate'))
        self.flow_beta_spatial = nn.Conv2d(flow_in_channel, self.out_channels // 4, 3, padding=1)
        self.flow_beta_temporal = zero_module(nn.Conv1d(self.out_channels // 4, self.out_channels, kernel_size=3, stride=1, padding=1, padding_mode='replicate'))
        self.flow_cond_norm = FloatGroupNorm(32, self.out_channels)

    def forward(self, h, flow, T):
        if flow is not None:
            gamma_flow = self.flow_gamma_spatial(flow)
            beta_flow = self.flow_beta_spatial(flow)
            _, _, hh, wh = beta_flow.shape
            gamma_flow = rearrange(gamma_flow, '(b f) c h w -> (b h w) c f', f=T)
            beta_flow = rearrange(beta_flow, '(b f) c h w -> (b h w) c f', f=T)
            gamma_flow = self.flow_gamma_temporal(gamma_flow)
            beta_flow = self.flow_beta_temporal(beta_flow)
            gamma_flow = rearrange(gamma_flow, '(b h w) c f -> (b f) c h w', h=hh, w=wh)
            beta_flow = rearrange(beta_flow, '(b h w) c f -> (b f) c h w', h=hh, w=wh)
            h = h + self.flow_cond_norm(h) * gamma_flow + beta_flow
        return h

INIT_ARGS = {
    'flow_in_channel': 8,
    'out_channels': 288,
}

ARGS = {
    'h': torch.randn(4, 288, 4, 3),
    'flow': torch.randn(4, 8, 4, 3),
    'T': 2,
}

if __name__ == '__main__':
    model = MGF(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Expected shape: {ARGS['h'].shape}")
    assert output.shape == ARGS['h'].shape, f"Shape mismatch: {output.shape} vs {ARGS['h'].shape}"
    print("Test passed!")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 4, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Patchify3D Ran
ran on a synthesized input · constructed an object; no result was computed
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("62ae80ac1c524e6d")
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, Patchify3D: 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
from einops import rearrange, repeat

torch.manual_seed(0)

# Patchify3D only has one parameter: patch_size
# It can be an int or tuple of 3 ints
# The forward method expects input of shape [B C T H W]
# where T, H, W are divisible by the patch dimensions

# Create a test input tensor
# Shape: [B=2, C=4, T=6, H=6, W=6]
# This is divisible by patch_size=(2,2,2)
test_input = torch.randn(2, 4, 6, 6, 6)

ARGS = {
    "patch_size": 2
}
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 Patchify3D — 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 Patchify3D 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
ResnetBlock Ran
ran on a synthesized input · constructed an object; no result was computed
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("9dddf9902d99b405")
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, ResnetBlock: 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)

def conv_nd(dims, *args, **kwargs):
    """
    Create a 1D, 2D, or 3D convolution module.
    """
    if dims == 1:
        return nn.Conv1d(*args, **kwargs)
    elif dims == 2:
        return nn.Conv2d(*args, **kwargs)
    elif dims == 3:
        return nn.Conv3d(*args, **kwargs)
    raise ValueError(f'unsupported dimensions: {dims}')

def avg_pool_nd(dims, *args, **kwargs):
    """
    Create a 1D, 2D, or 3D average pooling module.
    """
    if dims == 1:
        return nn.AvgPool1d(*args, **kwargs)
    elif dims == 2:
        return nn.AvgPool2d(*args, **kwargs)
    elif dims == 3:
        return nn.AvgPool3d(*args, **kwargs)
    raise ValueError(f'unsupported dimensions: {dims}')

class Downsample(nn.Module):
    """
    A downsampling layer with an optional convolution.
    :param channels: channels in the inputs and outputs.
    :param use_conv: a bool determining if a convolution is applied.
    :param dims: determines if the signal is 1D, 2D, or 3D. If 3D, then
                 downsampling occurs in the inner-two dimensions.
    """

    def __init__(self, channels, use_conv, dims=2, out_channels=None, padding=1):
        super().__init__()
        self.channels = channels
        self.out_channels = out_channels or channels
        self.use_conv = use_conv
        self.dims = dims
        stride = 2 if dims != 3 else (1, 2, 2)
        if use_conv:
            self.op = conv_nd(dims, self.channels, self.out_channels, 3, stride=stride, padding=padding)
        else:
            assert self.channels == self.out_channels
            self.op = avg_pool_nd(dims, kernel_size=stride, stride=stride)

    def forward(self, x):
        assert x.shape[1] == self.channels
        return self.op(x)

class ResnetBlock(nn.Module):

    def __init__(self, in_c, out_c, down, ksize=3, sk=False, use_conv=True):
        super().__init__()
        ps = ksize // 2
        if in_c != out_c or sk == False:
            self.in_conv = nn.Conv2d(in_c, out_c, ksize, 1, ps)
        else:
            self.in_conv = None
        self.block1 = nn.Conv2d(out_c, out_c, 3, 1, 1)
        self.act = nn.ReLU()
        self.block2 = nn.Conv2d(out_c, out_c, ksize, 1, ps)
        self.bn1 = nn.BatchNorm2d(out_c)
        self.bn2 = nn.BatchNorm2d(out_c)
        if sk == False:
            self.skep = nn.Conv2d(out_c, out_c, ksize, 1, ps)
        else:
            self.skep = None
        self.down = down
        if self.down == True:
            self.down_opt = Downsample(in_c, use_conv=use_conv)

    def forward(self, x):
        if self.down == True:
            x = self.down_opt(x)
        if self.in_conv is not None:
            x = self.in_conv(x)
        h = self.bn1(x)
        h = self.act(h)
        h = self.block1(h)
        h = self.bn2(h)
        h = self.act(h)
        h = self.block2(h)
        if self.skep is not None:
            return h + self.skep(x)
        else:
            return h + x

ARGS = {
    "in_c": 3,
    "out_c": 4,
    "down": False,
    "ksize": 3,
    "sk": False,
    "use_conv": True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ResnetBlock — 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 ResnetBlock 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
TrajExtractor Ran
ran on a synthesized input · constructed an object; no result was computed
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("7e6c1654297bad79")
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, TrajExtractor: 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
from einops import rearrange, repeat

torch.manual_seed(0)

ARGS = {
    "vae_downsize": (4, 8, 8),
    "patch_size": 2,
    "patch_size_t": 2,
    "channels": [320, 640, 1280, 1280],
    "nums_rb": 3,
    "cin": 2,
    "ksize": 3,
    "sk": False,
    "use_conv": True,
}
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 TrajExtractor — 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 TrajExtractor 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
ensure_tuple Ran
ran, violated the contract we drafted
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("6eb26eb0c6d9315b")
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 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 numpy as np

torch.manual_seed(0)

# Test case 1: xs is a tuple, ndim=3
xs_tuple = (1, 2, 3)
ndim_1 = 3

# Test case 2: xs is a list, ndim=2
xs_list = [5, 6]
ndim_2 = 2

# Test case 3: xs is a scalar, ndim=4
xs_scalar = 7
ndim_3 = 4

# Use the first test case for ARGS
ARGS = {
    "xs": (1, 2, 3),
    "ndim": 3
}
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 tuple, ndim 1, dtype family int — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_resize_crop_region_for_grid Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
alibaba/Tora/diffusers-version/tora/i2v_pipeline.py
code served (permissive licence) · get_code("734fd43f6aaef115")
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 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 numpy as np

# The function takes:
# - src: a tuple of (height, width) dimensions
# - tgt_width: target width (int)
# - tgt_height: target height (int)

ARGS = {
    "src": (480, 640),           # source dimensions as (height, width)
    "tgt_width": 224,            # target width
    "tgt_height": 224            # target height
}
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 tuple, ndim 2, dtype family int — 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
retrieve_timesteps Ran
ran; our contract draft was wrong, not the code
alibaba/Tora/diffusers-version/tora/i2v_pipeline.py
code served (permissive licence) · get_code("22b1f260da28f6a5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned tuple. 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 inspect
from typing import List, Optional, Union

torch.manual_seed(0)

class MockScheduler:
    """Mock scheduler that implements set_timesteps and has timesteps attribute."""
    
    def __init__(self, supports_timesteps=True, supports_sigmas=False):
        self.supports_timesteps = supports_timesteps
        self.supports_sigmas = supports_sigmas
        self.timesteps = torch.tensor([999, 500, 250, 0], dtype=torch.long)
        self._device = None
    
    def set_timesteps(self, num_inference_steps: Optional[int] = None, 
                      device: Optional[Union[str, torch.device]] = None,
                      timesteps: Optional[List[int]] = None,
                      sigmas: Optional[List[float]] = None,
                      **kwargs):
        """Mock set_timesteps method."""
        self._device = device
        
        if timesteps is not None:
            self.timesteps = torch.tensor(timesteps, dtype=torch.long)
        elif sigmas is not None:
            # Create timesteps from sigmas
            self.timesteps = torch.tensor([int(s * 1000) for s in sigmas], dtype=torch.long)
        elif num_inference_steps is not None:
            # Create default timesteps
            self.timesteps = torch.linspace(999, 0, num_inference_steps, dtype=torch.long)
        
        if device is not None:
            self.timesteps = self.timesteps.to(device)

# Create a basic scheduler instance
scheduler = MockScheduler(supports_timesteps=True, supports_sigmas=False)

ARGS = {
    "scheduler": scheduler,
    "num_inference_steps": 4,
    "device": "cpu",
    "timesteps": None,
    "sigmas": None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
CogVideoXBlock Not yet run
unverified (harvested, no recorded run)
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("696621f779b45cc6")
How it ran

no invocation recorded.

ToraTransformer3DModel Not yet run
unverified (harvested, no recorded run)
alibaba/Tora/diffusers-version/tora/transformer_3d.py
code served (permissive licence) · get_code("0d292a2ea249dbd7")
How it ran

no invocation recorded.

retrieve_latents Not yet run
unverified (harvested, no recorded run)
alibaba/Tora/diffusers-version/tora/i2v_pipeline.py
code served (permissive licence) · get_code("8262566e1ed7239a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 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 8 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
retrieve_timesteps 46 other papers
4 dated earlier, 40 later, 2 undated
2312.04461 (2023-12)
PhotoMaker: Customizing Realistic Human Photos via Stacked ID Embedding
2402.10491 (2024-02)
Make a Cheap Scaling: A Self-Cascade Diffusion Model for Higher-Resolution Adaptation
2405.18991 (2024-05)
EasyAnimate: High-Performance Video Generation Framework with Hybrid Windows Attention and Reward Backpropagation
and 43 more not named here
retrieve_latents 19 other papers
5 dated earlier, 1 the same month, 12 later, 1 undated
2401.07519 (2024-01)
InstantID: Zero-shot Identity-Preserving Generation in Seconds
2403.14773 (2024-03)
StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text
2404.06666 (2024-04)
SafeGen: Mitigating Sexually Explicit Content Generation in Text-to-Image Models
and 16 more not named here
Downsample
brick of this paper's assembly: helper
12 other papers
9 dated earlier, 3 later · 10 only as an assembly brick
2301.05225 (2023-01)
Domain Expansion of Image Generators
2301.13721 (2023-01)
DisDiff: Unsupervised Disentanglement of Diffusion Probabilistic Models
2302.04304 (2023-02)
Q-Diffusion: Quantizing Diffusion Models
and 9 more not named here
get_resize_crop_region_for_grid 8 other papers
1 dated earlier, 7 later
2405.18991 (2024-05)
EasyAnimate: High-Performance Video Generation Framework with Hybrid Windows Attention and Reward Backpropagation
2410.07659 (2024-10)
MotionAura: Generating High-Quality and Motion Consistent Videos using Discrete Diffusion
2501.01790 (2025-01)
Ingredients: Blending Custom Photos with Video Diffusion Transformers
and 5 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

Recent advancements in Diffusion Transformer (DiT) have demonstrated remarkable proficiency in producing high-quality video content. Nonetheless, the potential of transformer-based diffusion models for effectively generating videos with controllable motion remains an area of limited exploration. This paper introduces Tora, the first trajectory-oriented DiT framework that concurrently integrates textual, visual, and trajectory conditions, thereby enabling scalable video generation with effective motion guidance. Specifically, Tora consists of a Trajectory Extractor (TE), a Spatial-Temporal DiT, and a Motion-guidance Fuser (MGF). The TE encodes arbitrary trajectories into hierarchical spacetime motion patches with a 3D motion compression network. The MGF integrates the motion patches into the DiT blocks to generate consistent videos that accurately follow designated trajectories. Our design aligns seamlessly with DiT's scalability, allowing precise control of video content's dynamics with diverse durations, aspect ratios, and resolutions. Extensive experiments demonstrate that Tora excels in achieving high motion fidelity compared to the foundational DiT model, while also accurately simulating the complex movements of the physical world. Code is made available at https://github.com/alibaba/Tora .

For agents

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

get_harvested_code_for_paper("2407.21705")
get_code_for_paper("2407.21705")
have("2407.21705")

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

curl https://syntology.ai/api/ran/2407.21705.json

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

[![Syntology run record](https://syntology.ai/api/ran/2407.21705.svg)](https://syntology.ai/paper/2407.21705)

Connect an agent — have() is free.