SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2108.08020 · ICCV · 2021

Speech Drives Templates: Co-Speech Gesture Synthesis with Learned Templates

Shenghua Gao, Wen Liu, Shenhan Qian, Zhi Tu, Yihao Zhi

arXiv · PDF · Open in the Atlas

Code that ran

5 of 7 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; 0 honoured the contract we drafted, 1 violated it and 4 have no contract verdict; 0 failures of our instrument; 2 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/2108.08020.json · badge: /api/ran/2108.08020.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
shenhanqian/speechdrivestemplates canonical (classifier label)
classifier basis: census
5 ran · 2 no recorded run
FunctionStatusWhere it lives
calculate_frechet_distance Ran
ran, violated the contract we drafted
shenhanqian/speechdrivestemplates/core/utils/fgd.py
code served (permissive licence) · get_code("255ce544d2cf21df")
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 float32, 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 numpy as np
from scipy import linalg

# Set random seed for reproducibility
np.random.seed(42)

# Create small test data for Frechet distance calculation
# mu1 and mu2 are mean vectors (1D arrays)
dim = 4
mu1 = np.random.randn(dim).astype(np.float32)
mu2 = np.random.randn(dim).astype(np.float32)

# sigma1 and sigma2 are covariance matrices (2D arrays)
# Create positive semi-definite matrices by multiplying a random matrix by its transpose
A1 = np.random.randn(dim, dim).astype(np.float32)
sigma1 = (A1 @ A1.T).astype(np.float32)

A2 = np.random.randn(dim, dim).astype(np.float32)
sigma2 = (A2 @ A2.T).astype(np.float32)

# eps parameter (default already provided in function signature, but can override)
eps = 1e-06

ARGS = {
    'mu1': mu1,
    'sigma1': sigma1,
    'mu2': mu2,
    'sigma2': sigma2,
    'eps': eps
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float32, 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
compute_fgd Ran
ran on a synthesized input
shenhanqian/speechdrivestemplates/core/utils/fgd.py
code served (permissive licence) · get_code("787e0232fa9dfaf7")
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 (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 numpy as np
import torch

torch.manual_seed(0)

# Create two small feature matrices
# a: shape (2, 4) - 2 samples, 4 features
# b: shape (2, 4) - 2 samples, 4 features
a = np.random.randn(2, 4).astype(np.float32)
b = np.random.randn(2, 4).astype(np.float32)

ARGS = {
    "a": a,
    "b": b,
}
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 1, dtype family float, shape (1,) — shape (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
crop_pad_audio Ran
ran on a synthesized input
shenhanqian/speechdrivestemplates/core/utils/audio_processing.py
code served (permissive licence) · get_code("13e0fea7ebc7451f")
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 1, dtype family float, shape (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
import torch

# Analyze the function:
# - wav: must be array-like with len() and slicing support, appears to be audio samples
# - audio_length: integer specifying target length

# Create a small audio sample (1D numpy array)
wav_sample = np.array([0.1, 0.2, 0.3, 0.4, 0.5], dtype=np.float32)

# Create target audio length
audio_length = 8

ARGS = {
    "wav": wav_sample,
    "audio_length": audio_length
}
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 ndarray, ndim 1, dtype family float, shape (8,) — shape (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
draw_body_parts Ran
ran on a synthesized input
shenhanqian/speechdrivestemplates/core/utils/keypoint_visualization.py
code served (permissive licence) · get_code("a4c2cc5ad1ce27dc")
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 3, dtype family int, shape (480, 640, 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 numpy as np
import cv2

# Create a small test image
img = np.zeros((480, 640, 3), dtype=np.uint8)

# Create landmarks array with 135 keypoints (23 pose + 70 face + 21 left hand + 21 right hand)
# Each keypoint has [x, y, confidence] format
num_keypoints = 135
landmarks = np.random.rand(num_keypoints, 3) * 100 + 50
landmarks = landmarks.astype(np.float32)

ARGS = {
    'img': img,
    'landmarks': landmarks,
    'size': 2,
    'with_confidence_score': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (480, 640, 3) — shape (480, 640, 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
parse_audio_length Ran
ran on a synthesized input
shenhanqian/speechdrivestemplates/core/utils/audio_processing.py
code served (permissive licence) · get_code("cacc271304dc7f5b")
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

torch.manual_seed(0)

# parse_audio_length takes three numeric parameters:
# - audio_length: length of audio in samples (int or float)
# - sr: sample rate in Hz (int or float)
# - fps: frames per second (int or float)

ARGS = {
    "audio_length": 16000,
    "sr": 16000,
    "fps": 30
}
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 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
draw_hand_landmarks Not yet run
unverified (harvested, no recorded run)
shenhanqian/speechdrivestemplates/core/utils/keypoint_visualization.py
code served (permissive licence) · get_code("2a16e5bc90ddcb84")
How it ran

no invocation recorded.

draw_landmarks Not yet run
unverified (harvested, no recorded run)
shenhanqian/speechdrivestemplates/core/utils/keypoint_visualization.py
code served (permissive licence) · get_code("c95f0f8301dbcac0")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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 6 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
calculate_frechet_distance 3 other papers
1 dated earlier, 2 later
2103.15087 (2021-03)
Learning a Sketch Tensor Space for Image Inpainting of Man-made Scenes
2203.00867 (2022-03)
Incremental Transformer Structure Enhanced Image Inpainting with Masking Positional Encoding
2210.05950 (2022-10)
ZITS++: Image Inpainting by Improving the Incremental Transformer on Structural Priors

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

Co-speech gesture generation is to synthesize a gesture sequence that not only looks real but also matches with the input speech audio. Our method generates the movements of a complete upper body, including arms, hands, and the head. Although recent data-driven methods achieve great success, challenges still exist, such as limited variety, poor fidelity, and lack of objective metrics. Motivated by the fact that the speech cannot fully determine the gesture, we design a method that learns a set of gesture template vectors to model the latent conditions, which relieve the ambiguity. For our method, the template vector determines the general appearance of a generated gesture sequence, while the speech audio drives subtle movements of the body, both indispensable for synthesizing a realistic gesture sequence. Due to the intractability of an objective metric for gesturespeech synchronization, we adopt the lip-sync error as a proxy metric to tune and evaluate the synchronization ability of our model. Extensive experiments show the superiority of our method in both objective and subjective evaluations on fidelity and synchronization. 1 * Equal contribution. † Corresponding author.

For agents

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

get_harvested_code_for_paper("2108.08020")
get_code_for_paper("2108.08020")
have("2108.08020")

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

curl https://syntology.ai/api/ran/2108.08020.json

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

[![Syntology run record](https://syntology.ai/api/ran/2108.08020.svg)](https://syntology.ai/paper/2108.08020)

Connect an agent — have() is free.