SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2412.07720 · 2024

ACDiT: Interpolating Autoregressive Conditional Modeling and Diffusion Transformer

arXiv · PDF · Open in the Atlas

Code that ran

5 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 1 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/2412.07720.json · badge: /api/ran/2412.07720.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
thunlp/acdit canonical (classifier label)
classifier basis: census
5 ran · 1 no recorded run
FunctionStatusWhere it lives
center_crop_arr Ran
ran; our contract draft was wrong, not the code
thunlp/acdit/transform_utils.py
code served (permissive licence) · get_code("1712a07966b542ee")
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 Image. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
from PIL import Image

# Create a small test image
width, height = 256, 256
pil_image = Image.new('RGB', (width, height), color='red')

# Define arguments
ARGS = {
    'pil_image': pil_image,
    'image_size': 64
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Image — a Image output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Image output the recorded ndim 3 and dtype family int 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
crop Ran
ran; our contract draft was wrong, not the code
thunlp/acdit/transform_utils.py
code served (permissive licence) · get_code("b6159b54fe35ce26")
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, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small 4D video clip tensor with shape (T, C, H, W)
# T=2 (temporal), C=3 (channels), H=8 (height), W=6 (width)
clip = torch.randn(2, 3, 8, 6)

# Define crop parameters
i = 1  # starting row index
j = 2  # starting column index
h = 4  # crop height
w = 3  # crop width

ARGS = {
    'clip': clip,
    'i': i,
    'j': j,
    'h': h,
    'w': w
}
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, 3) — shape (2, 3, 4, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_2d_sincos_inblock_pos_embed Ran
ran, honoured the contract we drafted
thunlp/acdit/models.py
code served (permissive licence) · get_code("27f8382e2678b8bd")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (17, 8) 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

ARGS = {
    "embed_dim": 8,
    "grid_size": 4,
    "cls_token": True,
    "extra_tokens": 1,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (17, 8) — shape (17, 8) 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
modulate Ran
ran, honoured the contract we drafted
thunlp/acdit/models.py
code served (permissive licence) · get_code("c1e9c657601bb510")
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)

# Create small tensors for the modulate function
x = torch.randn(2, 4)  # batch size 2, dimension 4
shift = torch.randn(2, 4)  # same shape as x
scale = torch.randn(2, 4)  # same shape as x

ARGS = {
    "x": x,
    "shift": shift,
    "scale": scale,
}
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
skip_causal_attn_mask_mod_gen Ran
ran; our contract draft was wrong, not the code
thunlp/acdit/models.py
code served (permissive licence) · get_code("41ffe8fdbc1932e9")
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 1, dtype family float, shape (6,) 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)

# Analyzing the function:
# - b: batch size (used in torch.where but appears to be contextual)
# - h: head dimension (used in torch.where but appears to be contextual)
# - q_idx: query index tensor, gets divided by block_size
# - kv_idx: key-value index tensor, gets divided by block_size
# - block_size: scalar divisor
# - len1: scalar used in comparisons

# The function uses torch.where with tensor operations on q_idx and kv_idx
# These should be tensors that support broadcasting and arithmetic operations

b = 2
h = 4
q_idx = torch.tensor([0, 1, 2, 3, 4, 5], dtype=torch.long)
kv_idx = torch.tensor([0, 1, 2, 3, 4, 5], dtype=torch.long)
block_size = 2
len1 = 2

ARGS = {
    "b": b,
    "h": h,
    "q_idx": q_idx,
    "kv_idx": kv_idx,
    "block_size": block_size,
    "len1": len1
}
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 1, dtype family float, shape (6,) — shape (6,) 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
resize Not yet run
unverified (harvested, no recorded run)
thunlp/acdit/transform_utils.py
code served (permissive licence) · get_code("f1f9c7dfbe0b3105")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 1 the same month, 3 later) and 3 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
center_crop_arr 37 other papers
19 dated earlier, 4 the same month, 14 later
2211.17091 (2022-11)
Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models
2212.09748 (2022-12)
Scalable Diffusion Models with Transformers
2311.09257 (2023-11)
UFOGen: You Forward Once Large Scale Text-to-Image Generation via Diffusion GANs
and 34 more not named here
crop 5 other papers
3 dated earlier, 2 later
2402.16050 (2024-02)
Efficient Temporal Extrapolation of Multimodal Large Language Models with Temporal Grounding Bridge
2403.13802 (2024-03)
ZigMa: A DiT-style Zigzag Mamba Diffusion Model
2411.12951 (2024-11)
On the Consistency of Video Large Language Models in Temporal Comprehension
and 2 more not named here
resize 4 other papers
3 dated earlier, 1 later
2402.16050 (2024-02)
Efficient Temporal Extrapolation of Multimodal Large Language Models with Temporal Grounding Bridge
2403.13802 (2024-03)
ZigMa: A DiT-style Zigzag Mamba Diffusion Model
2411.12951 (2024-11)
On the Consistency of Video Large Language Models in Temporal Comprehension
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

Autoregressive and diffusion models have achieved remarkable progress in language models and visual generation, respectively. We present ACDiT, a novel Autoregressive blockwise Conditional Diffusion Transformer, that innovatively combines autoregressive and diffusion paradigms for continuous visual information. By introducing a block-wise autoregressive unit, ACDiT offers a flexible interpolation between token-wise autoregression and full-sequence diffusion, bypassing the limitations of discrete tokenization. The generation of each block is formulated as a conditional diffusion process, conditioned on prior blocks. ACDiT is easy to implement, as simple as applying a specially designed Skip-Causal Attention Mask on the standard diffusion transformer during training. During inference, the process iterates between diffusion denoising and autoregressive decoding that can make full use of KV-Cache. We validate the effectiveness of ACDiT on image, video, and text generation and show that ACDiT performs best among all autoregressive baselines under similar model scales on visual generation tasks. We also demonstrate that, benefiting from autoregressive modeling, pretrained ACDiT can be transferred in visual understanding tasks despite being trained with the generative objective. The analysis of the trade-off between autoregressive and diffusion demonstrates the potential of ACDiT to be used in long-horizon visual generation tasks. We hope that ACDiT offers a novel perspective on visual autoregressive generation and sheds light on new avenues for unified models.

For agents

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

get_harvested_code_for_paper("2412.07720")
get_code_for_paper("2412.07720")
have("2412.07720")

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

curl https://syntology.ai/api/ran/2412.07720.json

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

[![Syntology run record](https://syntology.ai/api/ran/2412.07720.svg)](https://syntology.ai/paper/2412.07720)

Connect an agent — have() is free.