SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2304.04746 · EMNLP · 2023

A Cheaper and Better Diffusion Language Model with Soft-Masked Noise

Aston Zhang, Mu Li, Alex Smola, Jiaao Chen, Diyi Yang

arXiv · PDF · Open in the Atlas

Code that ran

1 of 2 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 0 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/2304.04746.json · badge: /api/ran/2304.04746.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
amazon-science/masked-diffusion-lm marked official in Papers with Code
Papers with Code archive
1 ran · 1 no recorded run
FunctionStatusWhere it lives
GaussianDiffusion Ran
ran on a synthesized input · constructed an object; no result was computed
amazon-science/masked-diffusion-lm/improved-diffusion/improved_diffusion/gaussian_diffusion.py
code served (permissive licence) · get_code("a49081fe140ee2f0")
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, GaussianDiffusion: 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 numpy as np
import enum

torch.manual_seed(0)

class LossType(enum.Enum):
    MSE = enum.auto()
    RESCALED_MSE = enum.auto()
    KL = enum.auto()
    RESCALED_KL = enum.auto()
    E2E_KL = enum.auto()
    E2E_MSE = enum.auto()
    E2E_Simple_MSE = enum.auto()
    E2E_Simple_KL = enum.auto()

    def is_vb(self):
        return self == LossType.KL or self == LossType.RESCALED_KL

class ModelMeanType(enum.Enum):
    PREVIOUS_X = enum.auto()
    START_X = enum.auto()
    EPSILON = enum.auto()

class ModelVarType(enum.Enum):
    LEARNED = enum.auto()
    FIXED_SMALL = enum.auto()
    FIXED_LARGE = enum.auto()
    LEARNED_RANGE = enum.auto()

# Create betas array (1-D numpy array for each diffusion timestep)
# Typical diffusion schedules have timesteps from 1000 down to 1
num_timesteps = 1000
betas = np.linspace(0.0001, 0.02, num_timesteps)

# Create model_mean_type
model_mean_type = ModelMeanType.EPSILON

# Create model_var_type
model_var_type = ModelVarType.LEARNED

# Create loss_type
loss_type = LossType.MSE

# Create rescale_timesteps flag
rescale_timesteps = False

# Create model_arch (can be None or a string identifier)
model_arch = None

# Create training_mode
training_mode = 'emb'

ARGS = {
    'betas': betas,
    'model_mean_type': model_mean_type,
    'model_var_type': model_var_type,
    'loss_type': loss_type,
    'rescale_timesteps': rescale_timesteps,
    'model_arch': model_arch,
    'training_mode': training_mode,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type GaussianDiffusion — 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 GaussianDiffusion 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
LossType Not yet run
unverified (harvested, no recorded run)
amazon-science/masked-diffusion-lm/improved-diffusion/improved_diffusion/gaussian_diffusion.py
code served (permissive licence) · get_code("5ed7993d9c8b5aa3")
How it ran

no invocation recorded.

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 2 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.

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 that are based on iterative denoising have been recently proposed and leveraged in various generation tasks like image generation. Whereas, as a way inherently built for continuous data, existing diffusion models still have some limitations in modeling discrete data, e.g., languages. For example, the generally used Gaussian noise can not handle the discrete corruption well, and the objectives in continuous spaces fail to be stable for textual data in the diffusion process especially when the dimension is high. To alleviate these issues, we introduce a novel diffusion model for language modeling, Masked-Diffusion LM, with lower training cost and better performances, inspired by linguistic features in languages. Specifically, we design a linguistic-informed forward process which adds corruptions to the text through strategically soft-masking to better noise the textual data. Also, we directly predict the categorical distribution with cross-entropy loss function in every diffusion step to connect the continuous space and discrete space in a more efficient and straightforward way. Through experiments on 5 controlled generation tasks, we demonstrate that our Masked-Diffusion LM can achieve better generation quality than the state-of-the-art diffusion models with better efficiency. Code is available at https://github. com/SALT-NLP/Masked_Diffusioin_LM.

For agents

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

get_harvested_code_for_paper("2304.04746")
get_code_for_paper("2304.04746")
have("2304.04746")

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

curl https://syntology.ai/api/ran/2304.04746.json

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

[![Syntology run record](https://syntology.ai/api/ran/2304.04746.svg)](https://syntology.ai/paper/2304.04746)

Connect an agent — have() is free.