SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1908.04781 · 2019

Predicting 3D Human Dynamics from Video

arXiv · PDF · Open in the Atlas

Code that ran

7 of 17 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, 0 violated it and 7 have no contract verdict; 0 failures of our instrument; 10 no recorded run), measured 2026-11-21; 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/1908.04781.json · badge: /api/ran/1908.04781.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
jasonyzhang/phd canonical (classifier label)
classifier basis: census
7 ran · 10 no recorded run
FunctionStatusWhere it lives
get_dims Ran
ran on a synthesized input
jasonyzhang/phd/src/renderer.py
code served (permissive licence) · get_code("695ee32917b465d7")
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 int, ndim 0, dtype family int, 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 torch
import numpy as np

torch.manual_seed(0)

# The function get_dims takes a single parameter 'x'
# It checks if x is a torch.Tensor and calls .dim(), otherwise calls .ndim
# We need to provide both test cases to show it works with both

# Test with torch.Tensor
tensor_x = torch.randn(2, 4)

# Test with numpy array
numpy_x = np.random.randn(2, 4)

# We'll use the torch.Tensor as the primary argument
ARGS = {
    "x": tensor_x
}
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 int, ndim 0, dtype family int, 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
get_labels_poseflow Ran
ran on a synthesized input
jasonyzhang/phd/src/extract_tracks.py
code served (permissive licence) · get_code("3a63d35b434d3196")
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 list, ndim 3, 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 json
import tempfile
import os
import numpy as np

# Create a temporary JSON file with pose data in the format expected by the function
temp_dir = tempfile.mkdtemp()
json_path = os.path.join(temp_dir, "poses.json")

# Create sample pose data matching AlphaPose/PoseTrack format
# Each frame has people with keypoints and tracking indices
pose_data = {
    "image_0.jpg": [
        {
            "idx": 0,
            "keypoints": [10.0, 20.0, 0.9] * 17  # 17 keypoints (x, y, visibility)
        },
        {
            "idx": 1,
            "keypoints": [30.0, 40.0, 0.8] * 17
        }
    ],
    "image_1.jpg": [
        {
            "idx": 0,
            "keypoints": [11.0, 21.0, 0.9] * 17
        },
        {
            "idx": 1,
            "keypoints": [31.0, 41.0, 0.8] * 17
        }
    ],
    "image_2.jpg": [
        {
            "idx": 0,
            "keypoints": [12.0, 22.0, 0.9] * 17
        },
        {
            "idx": 1,
            "keypoints": [32.0, 42.0, 0.8] * 17
        }
    ],
    "image_3.jpg": [
        {
            "idx": 0,
            "keypoints": [13.0, 23.0, 0.9] * 17
        }
    ],
    "image_4.jpg": [
        {
            "idx": 0,
            "keypoints": [14.0, 24.0, 0.9] * 17
        }
    ]
}

with open(json_path, 'w') as f:
    json.dump(pose_data, f)

ARGS = {
    "json_path": json_path,
    "num_frames": 5,
    "min_kp_count": 3
}
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 list, ndim 3, dtype family float — a list 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
process_videos Ran
ran on a synthesized input
jasonyzhang/phd/src/evaluation/run_video.py
code served (permissive licence) · get_code("2cde9da4227a4934")
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. 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 os
import numpy as np

IMG_SIZE = 224

class Config:
    def __init__(self):
        self.fps = 30
        self.dataset = 'test_dataset'
        self.vid_path = '/path/to/video.mp4'
        self.start_frame = 0
        self.skip_rate = None
        self.vid_id = 'video_001.mp4'
        self.out_dir = '/tmp/output'
        self.batch_size = 2

# Create a config instance
config = Config()

# Create a list of dummy images (T frames of RGB images)
T = 16
images = [np.random.randint(0, 256, (IMG_SIZE, IMG_SIZE, 3), dtype=np.uint8) for _ in range(T + 5)]

# Create suffix string
suffix = 'test'

ARGS = {
    'config': config,
    'images': images,
    'T': T,
    'suffix': suffix
}
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 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
tf_pad Ran
ran on a synthesized input
jasonyzhang/phd/src/ops.py
code served (permissive licence) · get_code("05191822268f47e6")
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 EagerTensor, ndim 2, dtype family float, shape (5, 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 numpy as np
import tensorflow as tf

# Create a simple tensor to pad
tensor = tf.constant([[1.0, 2.0, 3.0, 4.0],
                       [5.0, 6.0, 7.0, 8.0]], dtype=tf.float32)

# Define paddings as list of list of non-negative ints
# Format: [[top, bottom], [left, right]] for 2D tensor
paddings = [[1, 2], [1, 1]]

# Define mode as either 'ZERO' or 'EDGE'
mode = 'ZERO'

ARGS = {
    'tensor': tensor,
    'paddings': paddings,
    'mode': mode
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 2, dtype family float, shape (5, 6) — shape (5, 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
to_float_tensor Ran
ran on a synthesized input
jasonyzhang/phd/src/renderer.py
code served (permissive licence) · get_code("e5f55a88efd3909d")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# The function takes a single parameter 'x' which can be a numpy array
# It converts numpy arrays to torch FloatTensors
# We'll provide a small numpy array as the argument

x_numpy = np.array([[1.0, 2.0, 3.0, 4.0],
                    [5.0, 6.0, 7.0, 8.0]], dtype=np.float32)

ARGS = {
    'x': x_numpy
}
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 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
to_variable Ran
ran on a synthesized input
jasonyzhang/phd/src/renderer.py
code served (permissive licence) · get_code("2e0b2d1dc17197f6")
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
from torch.autograd import Variable

torch.manual_seed(0)

# The function to_variable takes parameter 'x' and converts it to a Variable if it isn't already one
# We need to test with something that is NOT already a Variable

# Create a small tensor
x = torch.randn(2, 4)

ARGS = {
    "x": x
}
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
undo_chumpy Ran
ran on a synthesized input
jasonyzhang/phd/src/tf_smpl/batch_smpl.py
code served (permissive licence) · get_code("e24744b154d5aa09")
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 int, shape (2, 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

class ChumbyStandIn:
    """Mock object that has a .r attribute for the non-ndarray case"""
    def __init__(self, data):
        self.r = data

# Test case 1: numpy array input
x_numpy = np.array([[1, 2, 3], [4, 5, 6]])

# Test case 2: object with .r attribute
x_chumpy = ChumbyStandIn(np.array([[7, 8, 9], [10, 11, 12]]))

# Create ARGS dict - we'll use the numpy array case as the primary test
ARGS = {
    "x": x_numpy
}
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 int, shape (2, 3) — shape (2, 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
batch_orth_proj_idrot Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/tf_smpl/projection.py
code served (permissive licence) · get_code("74c9b49c72a62f9d")
How it ran

no invocation recorded.

batch_orth_proj_optcam Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/tf_smpl/projection.py
code served (permissive licence) · get_code("c7079337574a09d2")
How it ran

no invocation recorded.

batch_rodrigues Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/tf_smpl/batch_lbs.py
code served (permissive licence) · get_code("785c66706574bea1")
How it ran

no invocation recorded.

batch_rot2aa Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/tf_smpl/batch_lbs.py
code served (permissive licence) · get_code("5b19861e6645f0be")
How it ran

no invocation recorded.

batch_skew Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/tf_smpl/batch_lbs.py
code served (permissive licence) · get_code("fca9ceed95411f50")
How it ran

no invocation recorded.

compute_tracks Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/extract_tracks.py
code served (permissive licence) · get_code("b29fcb58b1cffa5f")
How it ran

no invocation recorded.

get_output_path_name Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/evaluation/run_video.py
code served (permissive licence) · get_code("59c1352c7c8f0c7b")
How it ran

no invocation recorded.

procrustes2d_vis Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/tf_smpl/projection.py
code served (permissive licence) · get_code("8a9b610b726cca33")
How it ran

no invocation recorded.

resize_img Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/util/common.py
code served (permissive licence) · get_code("ab7d47670a4b49ba")
How it ran

no invocation recorded.

run_poseflow Not yet run
unverified (harvested, no recorded run)
jasonyzhang/phd/src/extract_tracks.py
code served (permissive licence) · get_code("0b220a0d1abc1008")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 1 later) and 15 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
undo_chumpy 2 other papers
1 the same month, 1 later
1908.07201 (2019-08)
Three-D Safari: Learning to Estimate Zebra Pose, Shape, and Texture from Images "In the Wild"
2007.11110 (2020-07)
Who Left the Dogs Out? 3D Animal Reconstruction with Expectation Maximization in the Loop
resize_img 1 other paper
1 dated earlier
1905.06817 (2019-05)
Learning to Regress 3D Face Shape and Expression from an Image without 3D Supervision

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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Given a video of a person in action, we can easily guess the 3D future motion of the person. In this work, we present perhaps the first approach for predicting a future 3D mesh model sequence of a person from past video input. We do this for periodic motions such as walking and also actions like bowling and squatting seen in sports or workout videos. While there has been a surge of future prediction problems in computer vision, most approaches predict 3D future from 3D past or 2D future from 2D past inputs. In this work, we focus on the problem of predicting 3D future motion from past image sequences, which has a plethora of practical applications in autonomous systems that must operate safely around people from visual inputs. Inspired by the success of autoregressive models in language modeling tasks, we learn an intermediate latent space on which we predict the future. This effectively facilitates autoregressive predictions when the input differs from the output domain. Our approach can be trained on video sequences obtained in-the-wild without 3D ground truth labels. The project website with videos can be found at https://jasonyzhang.com/phd.

For agents

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

get_harvested_code_for_paper("1908.04781")
get_code_for_paper("1908.04781")
have("1908.04781")

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

curl https://syntology.ai/api/ran/1908.04781.json

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

[![Syntology run record](https://syntology.ai/api/ran/1908.04781.svg)](https://syntology.ai/paper/1908.04781)

Connect an agent — have() is free.