SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2512.02206 · NeurIPS · 2025

WhAM: Towards A Translative Model of Sperm Whale Vocalization

U Berkeley, Bryan Pardo, Shafi Goldwasser, Orr Paradise, Pranav Muralikrishnan, Liangyuan Chen, Hugo Flores García, Roee Diamant, David Gruber, Shane Gero, Project Ceti

arXiv · PDF · Open in the Atlas

Code that ran

9 of 9 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; 3 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 6 failures of our instrument; 0 no recorded run), measured 2026-11-22; 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/2512.02206.json · badge: /api/ran/2512.02206.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
Project-CETI/wham canonical (classifier label)
classifier basis: census
9 ran · 0 no recorded run
FunctionStatusWhere it lives
calculate_fad Ran
ran, honoured the contract we drafted
Project-CETI/wham/wham/generation/eval/calculate_custom_fad.py
code served (permissive licence) · get_code("0ef4987eb211029f")
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
import torch

# Set random seed for reproducibility
torch.manual_seed(0)

# Create small embedding statistics for two groups
# Using dimension 4 for embeddings
embedding_dim = 4

# First group statistics
mu1 = np.random.randn(embedding_dim).astype(np.float32)
sigma1 = np.random.randn(embedding_dim, embedding_dim).astype(np.float32)
sigma1 = sigma1 @ sigma1.T  # Make it positive semi-definite

# Second group statistics
mu2 = np.random.randn(embedding_dim).astype(np.float32)
sigma2 = np.random.randn(embedding_dim, embedding_dim).astype(np.float32)
sigma2 = sigma2 @ sigma2.T  # Make it positive semi-definite

ARGS = {
    "mu1": mu1,
    "sigma1": sigma1,
    "mu2": mu2,
    "sigma2": sigma2,
}
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 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
calculate_mse Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Project-CETI/wham/wham/generation/eval/mse_error_by_freq.py
code served (permissive licence) · get_code("1a0e2559febcd066")
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 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 numpy as np

# Create frequency arrays with distinct ranges
freq1 = np.array([10.0, 20.0, 30.0, 40.0, 50.0])
freq2 = np.array([15.0, 25.0, 35.0, 45.0, 55.0])

# Create corresponding FFT magnitude arrays
fft1 = np.array([1.0, 2.0, 3.0, 2.5, 1.5])
fft2 = np.array([1.2, 2.1, 2.8, 2.3, 1.6])

ARGS = {
    "freq1": freq1,
    "freq2": freq2,
    "fft1": fft1,
    "fft2": fft2,
}
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, ndim 2, dtype family float — 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
collate_fn Ran
ran; our contract draft was wrong, not the code
Project-CETI/wham/wham/embedding/train_downstream.py
code served (permissive licence) · get_code("e082a4e27242a338")
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 torch

torch.manual_seed(0)

# Construct a batch of (x, y) tuples
# x should be a tensor that can be padded (2D or higher)
# y should be an integer label

# Create samples with different sequence lengths to test sorting
sample1_x = torch.randn(5, 4)  # sequence length 5, feature dim 4
sample1_y = 0

sample2_x = torch.randn(3, 4)  # sequence length 3, feature dim 4
sample2_y = 1

sample3_x = torch.randn(7, 4)  # sequence length 7, feature dim 4
sample3_y = 1

# Build the batch as a list of tuples
batch = [
    (sample1_x, sample1_y),
    (sample2_x, sample2_y),
    (sample3_x, sample3_y),
]

ARGS = {
    "batch": batch,
}
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 — 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
compute_fft_from_data Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Project-CETI/wham/wham/generation/eval/mse_error_by_freq.py
code served (permissive licence) · get_code("c2ee0896228920bb")
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 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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create sample audio data
# The function expects a 1D or 2D array of audio samples
# Let's create 1D audio data with enough samples for multiple frames
# window_size=1500, hop_size=1000, so we need at least 1500 samples
# Let's create 5000 samples to get multiple frames
audio_data = np.random.randn(5000).astype(np.float32)

# Alternatively, test with 2D data (multichannel audio)
# The function takes the first channel if data.shape > 1
audio_data_2d = np.random.randn(5000, 2).astype(np.float32)

ARGS = {
    "data": audio_data,
    "samplerate": 44100,
    "window_size": 1500,
    "hop_size": 1000,
}
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, ndim 2, dtype family float — 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
compute_stats Ran
ran; our contract draft was wrong, not the code
Project-CETI/wham/wham/generation/eval/calculate_custom_fad.py
code served (permissive licence) · get_code("281ee52c47284fc3")
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 numpy as np

# Create a small embeddings array
# Shape: (2, 4) - 2 samples, 4 dimensions
embeddings = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0]
])

ARGS = {
    "embeddings": embeddings
}
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
load_embeddings Ran
ran, honoured the contract we drafted
Project-CETI/wham/wham/generation/eval/calculate_custom_fad.py
code served (permissive licence) · get_code("4346ebecc7c60973")
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 float, shape (6, 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 os
import numpy as np
import tempfile
import shutil

# Create a temporary directory with .npy files
temp_dir = tempfile.mkdtemp()

# Create sample embedding numpy arrays and save them
embed1 = np.random.randn(2, 4).astype(np.float32)
embed2 = np.random.randn(3, 4).astype(np.float32)
embed3 = np.random.randn(1, 4).astype(np.float32)

np.save(os.path.join(temp_dir, 'embed1.npy'), embed1)
np.save(os.path.join(temp_dir, 'embed2.npy'), embed2)
np.save(os.path.join(temp_dir, 'embed3.npy'), embed3)

ARGS = {
    'directory': temp_dir
}
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 float, shape (6, 4) — shape (6, 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
mkdir Ran
ran; our contract draft was wrong, not the code
Project-CETI/wham/vampnet/vampnet/beats.py
code served (permissive licence) · get_code("843cfb3ab9041539")
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 PosixPath. 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 tempfile
from pathlib import Path

# The function takes a single parameter 'path' which should be a Union[Path, str]
# The function creates a directory at that path using Path.mkdir()
# We'll provide a temporary directory path as a string

temp_dir = tempfile.gettempdir()
test_path = str(Path(temp_dir) / "test_mkdir_dir" / "nested" / "path")

ARGS = {
    "path": test_path
}
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 PosixPath — a PosixPath output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PosixPath 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
pathify Ran
ran; our contract draft was wrong, not the code
Project-CETI/wham/wham/utils.py
code served (permissive licence) · get_code("6f3d79c3b152f518")
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

def pathify(string):
    bad_chars = [' ', '.', '/', ':']
    return ''.join([c if c not in bad_chars else '_' for c in string])

# The function takes a single string parameter and replaces bad characters
# with underscores. We need to provide a string argument.

ARGS = {
    'string': 'Hello World. This/is:a test'
}
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
process_files Ran
ran, honoured the contract we drafted
Project-CETI/wham/wham/generation/eval/mse_error_by_freq.py
code served (permissive licence) · get_code("6a7a796665a4a0ee")
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 os
import numpy as np
import scipy.io.wavfile as wav
import tempfile
import shutil

# Create temporary directories with test audio files
original_dir = tempfile.mkdtemp()
gen_dir = tempfile.mkdtemp()

try:
    # Create a simple test audio file
    samplerate = 44100
    duration = 0.5  # seconds
    num_samples = int(samplerate * duration)
    
    # Generate simple sine wave for original file
    t = np.linspace(0, duration, num_samples)
    original_audio = np.sin(2 * np.pi * 440 * t).astype(np.float32)
    
    # Generate similar sine wave for generated file
    gen_audio = np.sin(2 * np.pi * 440 * t).astype(np.float32) * 0.95
    
    # Write test files
    filename = "test_audio.wav"
    original_file = os.path.join(original_dir, filename)
    gen_file = os.path.join(gen_dir, filename.replace('.wav', 'gen.wav'))
    
    wav.write(original_file, samplerate, (original_audio * 32767).astype(np.int16))
    wav.write(gen_file, samplerate, (gen_audio * 32767).astype(np.int16))
    
    # Define arguments for process_files
    ARGS = {
        "original_dir": original_dir,
        "gen_dir": gen_dir
    }

finally:
    # Cleanup will happen after execution
    pass
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 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

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 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, 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.

FunctionAlso present inEarliest-datedOthers named
mkdir 2 other papers
2 dated earlier
2307.04686 (2023-07)
VampNet: Music Generation via Masked Acoustic Token Modeling
2310.12404 (2023-10)
Loop Copilot: Conducting AI Ensembles for Music Generation and Iterative Editing

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

Sperm whales communicate in short sequences of clicks known as codas. We present WhAM (Whale Acoustics Model), the first transformer-based model capable of generating synthetic sperm whale codas from any audio prompt. WhAM is built by finetuning VampNet, a masked acoustic token model pretrained on musical audio, using 10k coda recordings collected over the past two decades. Through iterative masked token prediction, WhAM generates high-fidelity synthetic codas that preserve key acoustic features of the source recordings. We evaluate WhAM's synthetic codas using Fréchet Audio Distance and through perceptual studies with expert marine biologists. On downstream classification tasks including rhythm, social unit, and vowel classification, WhAM's learned representations achieve strong performance, despite being trained for generation rather than classification.

For agents

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

get_harvested_code_for_paper("2512.02206")
get_code_for_paper("2512.02206")
have("2512.02206")

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

curl https://syntology.ai/api/ran/2512.02206.json

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

[![Syntology run record](https://syntology.ai/api/ran/2512.02206.svg)](https://syntology.ai/paper/2512.02206)

Connect an agent — have() is free.