SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2304.02198 · ICLR · 2023

EigenFold: Generative Protein Structure Prediction with Diffusion Models

Tommi Jaakkola, Gabriele Corso, Bowen Jing, Bonnie Berger, Ezra Erives, Peter Pao-Huang

arXiv · PDF · Open in the Atlas

Code that ran

6 of 8 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 6 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/2304.02198.json · badge: /api/ran/2304.02198.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
bjing2016/EigenFold canonical (classifier label)
classifier basis not recorded on the link
6 ran · 2 no recorded run
FunctionStatusWhere it lives
get_args_suffix Ran
ran on a synthesized input
bjing2016/EigenFold/make_embeddings.py
code served (permissive licence) · get_code("a7494421173fbc41")
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
import numpy as np

class Args:
    def __init__(self):
        self.learning_rate = 0.001
        self.batch_size = 32
        self.num_epochs = 10
        self.model_name = "test_model"
        self.seed = 42

arg_keys = ["learning_rate", "batch_size", "num_epochs", "model_name", "seed"]
args = Args()

ARGS = {
    "arg_keys": arg_keys,
    "args": args
}
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 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
get_dense_edges Ran
ran on a synthesized input
bjing2016/EigenFold/utils/dataset.py
code served (permissive licence) · get_code("7c5f8c9622efee41")
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 int, shape (2, 12) 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)

def get_dense_edges(n):
    atom_ids = np.arange(n)
    src, dst = (np.repeat(atom_ids, n), np.tile(atom_ids, n))
    mask = src != dst
    src, dst = (src[mask], dst[mask])
    edge_idx = np.stack([src, dst])
    return torch.tensor(edge_idx)

ARGS = {
    "n": 4
}
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 int, shape (2, 12) — shape (2, 12) 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_logger Ran
ran on a synthesized input
bjing2016/EigenFold/utils/logging.py
code served (permissive licence) · get_code("0df91f35231085f8")
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 Logger. 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 logging
import socket
import os

def get_logger(name, level='info'):
    logger = logging.Logger(name)
    level = {'crititical': 50, 'error': 40, 'warning': 30, 'info': 20, 'debug': 10}[level]
    logger.setLevel(level)
    ch = logging.StreamHandler()
    ch.setLevel(logging.DEBUG)
    formatter = logging.Formatter(f'%(asctime)s [{socket.gethostname()}:%(process)d] [%(levelname)s] %(message)s')
    ch.setFormatter(formatter)
    logger.addHandler(ch)
    return logger

ARGS = {
    'name': 'test_logger',
    'level': 'info'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Logger — a Logger output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Logger 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_score_fn Ran
ran on a synthesized input
bjing2016/EigenFold/utils/inference.py
code served (permissive licence) · get_code("bb9ab87b96e532ca")
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 function. 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 torch.nn as nn
import copy

torch.manual_seed(0)

# Create a minimal SDE class
class MinimalSDE:
    def __init__(self):
        self.N = 2
    
    def score_norm(self, t, k, adj=False):
        return torch.tensor(1.0)

# Create a minimal data object
class MinimalData:
    def __init__(self):
        self.sde = MinimalSDE()
        self.resi = type('obj', (object,), {
            'pos': torch.randn(2, 4),
            'num_nodes': 2,
            'node_t': torch.zeros(2)
        })()
        self.sidechain = type('obj', (object,), {
            'pos': torch.randn(3, 4)
        })()
        self.resi_sde = type('obj', (object,), {'N': 2})()
        self.score_norm = torch.tensor(1.0)
    
    def to(self, device):
        return self
    
    def __getitem__(self, key):
        if key == 'resi':
            return self.resi
        elif key == 'sidechain':
            return self.sidechain
        return None
    
    def __setitem__(self, key, value):
        if key == 'resi':
            self.resi = value
        elif key == 'sidechain':
            self.sidechain = value
        elif key == 'score_norm':
            self.score_norm = value

# Create a minimal model with enn method
class MinimalModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 4)
    
    def enn(self, data):
        # Return a dummy score tensor
        return torch.randn(5, 4)

# Build the data object
data = MinimalData()

# Build the model
model = MinimalModel()

# Define arguments
ARGS = {
    'args': None,  # args not used in the function body
    'model': model,
    'data': data,
    'key': 'resi',
    'device': 'cpu'
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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_timestep_embedding Ran
ran on a synthesized input
bjing2016/EigenFold/model/utils.py
code served (permissive licence) · get_code("beaa7b5b126512d3")
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 function. 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 torch.nn as nn
import torch.nn.functional as F
import numpy as np
import math

torch.manual_seed(0)

def sinusoidal_embedding(timesteps, embedding_dim, max_positions=10000):
    """ from https://github.com/hojonathanho/diffusion/blob/master/diffusion_tf/nn.py   """
    assert len(timesteps.shape) == 1
    half_dim = embedding_dim // 2
    emb = math.log(max_positions) / (half_dim - 1)
    emb = torch.exp(torch.arange(half_dim, dtype=torch.float32, device=timesteps.device) * -emb)
    emb = timesteps[:, None] * emb[None, :]
    emb = torch.cat([torch.sin(emb), torch.cos(emb)], dim=1)
    if embedding_dim % 2 == 1:
        emb = F.pad(emb, (0, 1), mode='constant')
    assert emb.shape == (timesteps.shape[0], embedding_dim)
    return emb

class GaussianFourierProjection(nn.Module):
    """Gaussian Fourier embeddings for noise levels.
    from https://github.com/yang-song/score_sde_pytorch/blob/1618ddea340f3e4a2ed7852a0694a809775cf8d0/models/layerspp.py#L32
    """

    def __init__(self, embedding_size=256, scale=1.0):
        super().__init__()
        self.W = nn.Parameter(torch.randn(embedding_size // 2) * scale, requires_grad=False)

    def forward(self, x):
        x_proj = x[:, None] * self.W[None, :] * 2 * np.pi
        emb = torch.cat([torch.sin(x_proj), torch.cos(x_proj)], dim=-1)
        return emb

def get_timestep_embedding(embedding_type, embedding_dim, embedding_scale=10000):
    if embedding_type == 'sinusoidal':
        emb_func = lambda x: sinusoidal_embedding(x, embedding_dim, max_positions=embedding_scale)
    elif embedding_type == 'fourier':
        emb_func = GaussianFourierProjection(embedding_size=embedding_dim, scale=embedding_scale)
    else:
        raise NotImplemented
    return emb_func

ARGS = {
    'embedding_type': 'sinusoidal',
    'embedding_dim': 8,
    'embedding_scale': 10000
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
sinusoidal_embedding Ran
ran on a synthesized input
bjing2016/EigenFold/model/utils.py
code served (permissive licence) · get_code("5a6ee1c70d1350c4")
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
import torch.nn.functional as F
import math

torch.manual_seed(0)

timesteps = torch.tensor([0, 1], dtype=torch.int64)
embedding_dim = 4
max_positions = 10000

ARGS = {
    "timesteps": timesteps,
    "embedding_dim": embedding_dim,
    "max_positions": max_positions
}
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
logp Not yet run
unverified (harvested, no recorded run)
bjing2016/EigenFold/diffusion/sampling.py
code served (permissive licence) · get_code("b6047193301bb1af")
How it ran

no invocation recorded.

reverse_sample Not yet run
unverified (harvested, no recorded run)
bjing2016/EigenFold/diffusion/sampling.py
code served (permissive licence) · get_code("e82cac0ee735eef1")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 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 8 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

Protein structure prediction has reached revolutionary levels of accuracy on single structures, yet distributional modeling paradigms are needed to capture the conformational ensembles and flexibility that underlie biological function. Towards this goal, we develop EIGENFOLD, a diffusion generative modeling framework for sampling a distribution of structures from a given protein sequence. We define a diffusion process that models the structure as a system of harmonic oscillators and which naturally induces a cascading-resolution generative process along the eigenmodes of the system. On recent CAMEO targets, EIGENFOLD achieves a median TMScore of 0.84, while providing a more comprehensive picture of model uncertainty via the ensemble of sampled structures relative to existing methods. We then assess EIGENFOLD's ability to model and predict conformational heterogeneity for fold-switching proteins and ligand-induced conformational change. Code is available at https://github.com/bjing2016/EigenFold.

For agents

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

get_harvested_code_for_paper("2304.02198")
get_code_for_paper("2304.02198")
have("2304.02198")

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

curl https://syntology.ai/api/ran/2304.02198.json

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

[![Syntology run record](https://syntology.ai/api/ran/2304.02198.svg)](https://syntology.ai/paper/2304.02198)

Connect an agent — have() is free.