SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2312.09767 · 2023

DreamTalk: When Emotional Talking Head Generation Meets Diffusion Probabilistic Models

arXiv · PDF · Open in the Atlas

Code that ran

6 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; 1 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 3 failures of our instrument; 0 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/2312.09767.json · badge: /api/ran/2312.09767.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
ali-vilab/dreamtalk canonical (classifier label)
classifier basis: census
6 ran · 0 no recorded run
FunctionStatusWhere it lives
face3d_norm_to_raw Ran
ran, violated the contract we drafted
ali-vilab/dreamtalk/core/networks/diffusion_net.py
code served (permissive licence) · get_code("393ccc99a3b65cf6")
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, 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 numpy as np

torch.manual_seed(0)

# face3d_norm: (B, L, C_face3d) - normalized face 3D features
# exp_min: (C_face3d) - minimum values for denormalization
# exp_max: (C_face3d) - maximum values for denormalization

B = 2  # batch size
L = 4  # sequence length
C_face3d = 5  # feature dimension

# Create normalized face3d tensor in range [-1, 1]
face3d_norm = torch.randn(B, L, C_face3d)

# Create min and max tensors for denormalization
exp_min = torch.randn(C_face3d) * 10  # arbitrary min values
exp_max = exp_min + torch.rand(C_face3d) * 20  # max should be > min

ARGS = {
    "face3d_norm": face3d_norm,
    "exp_min": exp_min,
    "exp_max": exp_max,
}
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 3, dtype family float, shape (2, 4, 5) — shape (2, 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
face3d_raw_to_norm Ran
ran, violated the contract we drafted
ali-vilab/dreamtalk/core/networks/diffusion_net.py
code served (permissive licence) · get_code("7e7a2a7876ca493c")
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, 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)

# Parameters for face3d_raw_to_norm
B = 2  # batch size
L = 3  # sequence length
C_face3d = 4  # number of face3d channels

# Create face3d_raw tensor: (B, L, C_face3d)
face3d_raw = torch.randn(B, L, C_face3d)

# Create exp_min tensor: (C_face3d,)
exp_min = torch.tensor([-1.0, -0.5, -2.0, -1.5])

# Create exp_max tensor: (C_face3d,)
exp_max = torch.tensor([1.0, 0.5, 2.0, 1.5])

ARGS = {
    "face3d_raw": face3d_raw,
    "exp_min": exp_min,
    "exp_max": exp_max,
}
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, 4) — shape (2, 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
get_face3d_clip Ran
ran, honoured the contract we drafted
ali-vilab/dreamtalk/core/utils.py
code served (permissive licence) · get_code("58ed5133da1179a6")
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 (5, 64) 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 os
import numpy as np
import torch
import tempfile
from scipy.io import savemat

torch.manual_seed(0)
np.random.seed(0)

# Create a temporary directory for test files
temp_dir = tempfile.mkdtemp()

# Create a test .mat file
mat_data = {
    'coeff': np.random.randn(10, 200)  # 10 frames, 200 coefficients (includes expression at 80:144)
}
mat_file_path = os.path.join(temp_dir, 'test_video.mat')
savemat(mat_file_path, mat_data)

# Create a test .txt file
txt_data = np.random.randn(10, 64)  # 10 frames, 64 expression coefficients
txt_file_path = os.path.join(temp_dir, 'test_video.txt')
np.savetxt(txt_file_path, txt_data)

ARGS = {
    'video_name': 'test_video.mat',
    'video_root_dir': temp_dir,
    'num_frames': 5,
    'start_idx': 'middle',
    'dtype': torch.float32
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, 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 (5, 64) — shape (5, 64) 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_style_video_lists Ran
ran; our contract draft was wrong, not the code
ali-vilab/dreamtalk/core/utils.py
code served (permissive licence) · get_code("ff61b3c1dcddc54d")
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 defaultdict. 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 collections import defaultdict

def get_video_style(video_name, style_type):
    person_id, direction, emotion, level, *_ = video_name.split('_')
    if style_type == 'id_dir_emo_level':
        style = '_'.join([person_id, direction, emotion, level])
    elif style_type == 'emotion':
        style = emotion
    elif style_type == 'id':
        style = person_id
    else:
        raise ValueError('Unknown style type')
    return style

def get_style_video_lists(video_list, style_type):
    style2video_list = defaultdict(list)
    for video in video_list:
        style = get_video_style(video, style_type)
        style2video_list[style].append(video)
    return style2video_list

# Create sample video names that parse correctly
# Format: person_id_direction_emotion_level
video_list = [
    'person1_left_happy_high',
    'person1_right_sad_low',
    'person2_left_happy_medium',
    'person2_right_angry_high',
    'person1_left_sad_low',
    'person3_center_neutral_high'
]

style_type = 'emotion'

ARGS = {
    'video_list': video_list,
    'style_type': style_type
}
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 defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict 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
get_video_style Ran
ran; our contract draft was wrong, not the code
ali-vilab/dreamtalk/core/utils.py
code served (permissive licence) · get_code("eff3585e52fa9c67")
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 str. 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

ARGS = {
    'video_name': 'person_001_left_happy_high_extra_info',
    'style_type': 'id_dir_emo_level'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
mish Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ali-vilab/dreamtalk/core/networks/mish.py
code served (permissive licence) · get_code("e8af53cde1de7240")
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; 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 torch.nn.functional as F

torch.manual_seed(0)

# The mish function takes a single parameter 'input' which should be a tensor
# The function applies element-wise operations: input * tanh(softplus(input))
# So input should be a tensor of any shape

input_tensor = torch.randn(2, 4)

ARGS = {
    "input": input_tensor
}
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 — no output shape is stated: observe_signature.py could not observe it (import:OSError)
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

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, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 5 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
mish 5 other papers
4 dated earlier, 1 later
1603.06937 (2016-03)
Stacked Hourglass Networks for Human Pose Estimation
1610.02357 (2016-10)
Xception: Deep Learning with Depthwise Separable Convolutions
1908.08681 (2019-08)
Mish: A Self Regularized Non-Monotonic Activation Function
and 2 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

Emotional talking head generation has attracted growing attention. Previous methods, which are mainly GAN-based, still struggle to consistently produce satisfactory results across diverse emotions and cannot conveniently specify personalized emotions. In this work, we leverage powerful diffusion models to address the issue and propose DreamTalk, a framework that employs meticulous design to unlock the potential of diffusion models in generating emotional talking heads. Specifically, DreamTalk consists of three crucial components: a denoising network, a style-aware lip expert, and a style predictor. The diffusion-based denoising network can consistently synthesize high-quality audio-driven face motions across diverse emotions. To enhance lip-motion accuracy and emotional fullness, we introduce a style-aware lip expert that can guide lip-sync while preserving emotion intensity. To more conveniently specify personalized emotions, a diffusion-based style predictor is utilized to predict the personalized emotion directly from the audio, eliminating the need for extra emotion reference. By this means, DreamTalk can consistently generate vivid talking faces across diverse emotions and conveniently specify personalized emotions. Extensive experiments validate DreamTalk's effectiveness and superiority. The code is available at https://github.com/ali-vilab/dreamtalk.

For agents

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

get_harvested_code_for_paper("2312.09767")
get_code_for_paper("2312.09767")
have("2312.09767")

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

curl https://syntology.ai/api/ran/2312.09767.json

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

[![Syntology run record](https://syntology.ai/api/ran/2312.09767.svg)](https://syntology.ai/paper/2312.09767)

Connect an agent — have() is free.