SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2401.03497 · IJCAI · 2024

EAT: Self-Supervised Pre-Training with Efficient Audio Transformer

Xie Chen, Zhisheng Zheng, Ziyang Ma, Wenxi Chen, Yuzhe Liang

arXiv · PDF · Open in the Atlas

Code that ran

13 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; 3 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 9 failures of our instrument; 4 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/2401.03497.json · badge: /api/ran/2401.03497.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
cwx-worst-one/eat canonical (classifier label)
classifier basis: census
13 ran · 4 no recorded run
FunctionStatusWhere it lives
build_dictionary Ran
ran; our contract draft was wrong, not the code
cwx-worst-one/eat/inference/inference.py
code served (permissive licence) · get_code("6a3e3a9bf8ae9914")
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 dict. 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 csv
import tempfile
import os

# Create a temporary CSV file with the expected format
temp_dir = tempfile.mkdtemp()
label_path = os.path.join(temp_dir, 'labels.csv')

with open(label_path, 'w', newline='') as f:
    writer = csv.writer(f)
    writer.writerow(['0', 'dummy_col1', 'cat'])
    writer.writerow(['1', 'dummy_col1', 'dog'])
    writer.writerow(['2', 'dummy_col1', 'bird'])

ARGS = {
    'label_path': label_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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
compute_block_mask_1d Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
cwx-worst-one/eat/utils/data_utils.py
code served (permissive licence) · get_code("66f7af815e734d4e")
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, 16) 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 math

torch.manual_seed(0)

ARGS = {
    "shape": (2, 16),
    "mask_prob": 0.3,
    "mask_length": 4,
    "mask_prob_adjust": 0.1,
    "inverse_mask": False,
    "require_same_masks": True,
    "expand_adjcent": False,
    "mask_dropout": 0.0,
    "non_overlapping": False,
}
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, 16) — shape (2, 16) 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_block_mask_2d Ran
ran on a synthesized input
cwx-worst-one/eat/utils/data_utils.py
code served (permissive licence) · get_code("067591e38407a2be")
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, 16) 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)
np.random.seed(0)

ARGS = {
    "shape": (2, 16),  # (batch_size=2, sequence_length=16)
    "mask_prob": 0.3,
    "mask_length": 3,
    "mask_prob_adjust": 0.1,
    "inverse_mask": False,
    "require_same_masks": True,
    "expand_adjcent": False,
    "mask_dropout": 0.0,
    "non_overlapping": False,
    "img_shape": (4, 4),  # 4x4 image grid matching L=16
    "flexible_mask": False,
}
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, 16) — shape (2, 16) 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_mask_indices Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
cwx-worst-one/eat/utils/data_utils.py
code served (permissive licence) · get_code("c4e0295ea525f219")
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 (2, 16) 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)
np.random.seed(0)

# Define arguments for compute_mask_indices
batch_size = 2
seq_length = 16

shape = (batch_size, seq_length)

# Optional padding mask: shape (batch_size, seq_length)
# False = not padded, True = padded
padding_mask = torch.tensor([
    [False, False, False, False, False, False, True, True, True, True, True, True, True, True, True, True],
    [False, False, False, False, False, False, False, False, True, True, True, True, True, True, True, True]
])

mask_prob = 0.15
mask_length = 4
mask_type = 'static'
mask_other = 0.0
min_masks = 1
no_overlap = False
min_space = 0
require_same_masks = True
mask_dropout = 0.0
add_masks = False
seed = 42
epoch = 0
indices = torch.tensor([0, 1])
idc_select_ver = 1
num_mask_ver = 2

ARGS = {
    'shape': shape,
    'padding_mask': padding_mask,
    'mask_prob': mask_prob,
    'mask_length': mask_length,
    'mask_type': mask_type,
    'mask_other': mask_other,
    'min_masks': min_masks,
    'no_overlap': no_overlap,
    'min_space': min_space,
    'require_same_masks': require_same_masks,
    'mask_dropout': mask_dropout,
    'add_masks': add_masks,
    'seed': seed,
    'epoch': epoch,
    'indices': indices,
    'idc_select_ver': idc_select_ver,
    'num_mask_ver': num_mask_ver,
}
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 2, dtype family float, shape (2, 16) — shape (2, 16) 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
extract_feature_tensor Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
cwx-worst-one/eat/feature_extract/feature_extract.py
code served (permissive licence) · get_code("22d21a785404efe9")
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, 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 torch
import torch.nn as nn

torch.manual_seed(0)

# Create a mock HuggingFace model
class MockHuggingFaceModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 8)
    
    def extract_features(self, x):
        return self.linear(x)

# Create a mock Fairseq model
class MockFairseqModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 8)
    
    def extract_features(self, x, padding_mask=None, mask=False, remove_extra_tokens=False):
        return {'x': self.linear(x)}

# Create test data
x_huggingface = torch.randn(2, 4)
x_fairseq = torch.randn(2, 4)

# Create model instances
model_huggingface = MockHuggingFaceModel()
model_fairseq = MockFairseqModel()

# Test with HuggingFace framework
ARGS = {
    'model': model_huggingface,
    'x': x_huggingface,
    'framework': 'huggingface'
}
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, 8) — shape (2, 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
get_2d_sincos_pos_embed Ran
ran, honoured the contract we drafted
cwx-worst-one/eat/models/mae.py
pointer only (licence: MIT); the stored copy get_code serves, from lluosi/RDB-CL, is pointer only (licence: NONE) · get_code("77e8a3ac46f3afec")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from lluosi/RDB-CL, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_2d_sincos_pos_embed_flexible Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
cwx-worst-one/eat/models/mae.py
code served (permissive licence) · get_code("68fc090d0a502993")
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 (12, 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

# Construct arguments for get_2d_sincos_pos_embed_flexible
embed_dim = 8  # must be even
grid_size = (4, 3)  # (height, width) - distinct dimensions
cls_token = False

ARGS = {
    "embed_dim": embed_dim,
    "grid_size": grid_size,
    "cls_token": cls_token,
}
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 (12, 8) — shape (12, 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
get_annealed_rate Ran
ran, honoured the contract we drafted
cwx-worst-one/eat/models/base.py
code served (permissive licence) · get_code("5d5be3c7916a67cb")
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 float, 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 torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - start: numeric value (used in arithmetic)
# - end: numeric value (used in arithmetic)
# - curr_step: numeric value (used in comparison and division)
# - total_steps: numeric value (used in division)
# All parameters are used in simple arithmetic operations

ARGS = {
    "start": 0.1,
    "end": 1.0,
    "curr_step": 25,
    "total_steps": 100
}
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 float, 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
get_layer_id_for_vit Ran
ran, honoured the contract we drafted
cwx-worst-one/eat/models/EAT_audio_classification.py
code served (permissive licence) · get_code("3f903f847b8f7f28")
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

torch.manual_seed(0)

# The function get_layer_id_for_vit takes two parameters:
# 1. name: a string representing a parameter name
# 2. num_layers: an integer representing the number of layers

# Let's construct valid test cases
name = "blocks.5.attn.qkv.weight"
num_layers = 12

ARGS = {
    "name": name,
    "num_layers": num_layers,
}
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 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
mixup_target Ran
ran; our contract draft was wrong, not the code
cwx-worst-one/eat/utils/mixup.py
code served (permissive licence) · get_code("6b945011d5d4e179")
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 (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)

# Construct arguments for mixup_target
target = torch.tensor([0, 1, 2, 3], dtype=torch.long)
num_classes = 4
lam = 0.8
smoothing = 0.1

ARGS = {
    "target": target,
    "num_classes": num_classes,
    "lam": lam,
    "smoothing": smoothing
}
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 1, dtype family float, shape (4,) — shape (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
one_hot Ran
ran; our contract draft was wrong, not the code
cwx-worst-one/eat/utils/mixup.py
pointer only (licence: MIT); the stored copy get_code serves, from uhlmanngroup/soft-target-infonce, is pointer only (licence: NONE) · get_code("6caffa789f1bdabe")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from uhlmanngroup/soft-target-infonce, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

rand_bbox Ran
ran; our contract draft was wrong, not the code
cwx-worst-one/eat/utils/mixup.py
pointer only (licence: MIT); the stored copy get_code serves, from terrypei/efficientvmamba, is pointer only (licence: NONE) · get_code("3913c620f6926f42")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from terrypei/efficientvmamba, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

random_masking Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
cwx-worst-one/eat/models/base.py
code served (permissive licence) · get_code("ee9ac852b4eb7975")
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 MaskInfo. 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 namedtuple

torch.manual_seed(0)

MaskSeed = namedtuple('MaskSeed', ['seed', 'update', 'ids'])

# Create input tensor: (batch_size=2, sequence_length=8, embedding_dim=4)
x = torch.randn(2, 8, 4)

# Create mask_ratio parameter
mask_ratio = 0.5

# Create mask_seed parameter with required fields
mask_seed_ids = torch.tensor([1, 2, 3])
mask_seed = MaskSeed(seed=42, update=1, ids=mask_seed_ids)

ARGS = {
    'x': x,
    'mask_ratio': mask_ratio,
    'mask_seed': mask_seed
}
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 MaskInfo — a MaskInfo 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
gather_unmasked Not yet run
unverified (harvested, no recorded run)
cwx-worst-one/eat/models/base.py
code served (permissive licence) · get_code("3a95f65cb24c84e6")
How it ran

no invocation recorded.

load_model Not yet run
unverified (harvested, no recorded run)
cwx-worst-one/eat/feature_extract/feature_extract.py
code served (permissive licence) · get_code("a663b7a41d863edb")
How it ran

no invocation recorded.

load_model Not yet run
unverified (harvested, no recorded run)
cwx-worst-one/eat/inference/inference.py
code served (permissive licence) · get_code("31694976f9ae8e78")
How it ran

no invocation recorded.

modify_relative_position_bias Not yet run
unverified (harvested, no recorded run)
cwx-worst-one/eat/models/mae.py
code served (permissive licence) · get_code("915281feeb90ee5a")
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, 9 are also present in other papers whose code we hold (9 with at least one paper dated earlier, 1 the same month, 9 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
get_2d_sincos_pos_embed 95 other papers
43 dated earlier, 2 the same month, 49 later, 1 same year, order not determinable · 33 only as an assembly brick
1807.10221 (2018-07)
Unified Perceptual Parsing for Scene Understanding
2105.12723 (2021-05)
Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
2111.06377 (2021-11)
Masked Autoencoders Are Scalable Vision Learners
and 92 more not named here
rand_bbox 14 other papers
12 dated earlier, 2 later
2109.03508 (2021-09)
RepNAS: Searching for Efficient Re-parameterizing Blocks
2111.06377 (2021-11)
Masked Autoencoders Are Scalable Vision Learners
2206.13559 (2022-06)
ST-Adapter: Parameter-Efficient Image-to-Video Transfer Learning
and 11 more not named here
get_2d_sincos_pos_embed_flexible 5 other papers
1 dated earlier, 4 later
2207.06405 (2022-07)
Masked Autoencoders that Listen
2403.07815 (2024-03)
Chronos: Learning the Language of Time Series
2405.00233 (2024-05)
SemantiCodec: An Ultra Low Bitrate Semantic Audio Codec for General Sound
and 2 more not named here
get_layer_id_for_vit 3 other papers
1 dated earlier, 2 later
2212.07525 (2022-12)
Efficient Self-supervised Learning with Contextualized Target Representations for Vision, Speech and Language
2410.13179 (2024-10)
EH-MAM: Easy-to-Hard Masked Acoustic Modeling for Self-Supervised Speech Representation Learning
2506.12222 (2025-06)
SSLAM: Enhancing Self-Supervised Models with Audio Mixtures for Polyphonic Soundscapes
modify_relative_position_bias 3 other papers
1 dated earlier, 2 later
2212.07525 (2022-12)
Efficient Self-supervised Learning with Contextualized Target Representations for Vision, Speech and Language
2410.13179 (2024-10)
EH-MAM: Easy-to-Hard Masked Acoustic Modeling for Self-Supervised Speech Representation Learning
2506.12222 (2025-06)
SSLAM: Enhancing Self-Supervised Models with Audio Mixtures for Polyphonic Soundscapes
gather_unmasked 2 other papers
1 dated earlier, 1 later
2312.15185 (2023-12)
emotion2vec: Self-Supervised Pre-Training for Speech Emotion Representation
2506.12222 (2025-06)
SSLAM: Enhancing Self-Supervised Models with Audio Mixtures for Polyphonic Soundscapes
get_annealed_rate 2 other papers
1 dated earlier, 1 later
2312.15185 (2023-12)
emotion2vec: Self-Supervised Pre-Training for Speech Emotion Representation
2506.12222 (2025-06)
SSLAM: Enhancing Self-Supervised Models with Audio Mixtures for Polyphonic Soundscapes
one_hot 2 other papers
1 dated earlier, 1 later
2310.04361 (2023-10)
Exploiting Activation Sparsity with Dense to Dynamic-k Mixture-of-Experts Conversion
2404.14076 (2024-04)
Towards noise contrastive estimation with soft targets for conditional models
random_masking 2 other papers
1 dated earlier, 1 later
2312.15185 (2023-12)
emotion2vec: Self-Supervised Pre-Training for Speech Emotion Representation
2506.12222 (2025-06)
SSLAM: Enhancing Self-Supervised Models with Audio Mixtures for Polyphonic Soundscapes

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

Audio self-supervised learning (SSL) pre-training, which aims to learn good representations from unlabeled audio, has made remarkable progress. However, the extensive computational demands during pre-training pose a significant barrier to the potential application and optimization of audio SSL models. In this paper, inspired by the success of data2vec 2.0 in image modality and Audio-MAE in audio modality, we introduce Efficient Audio Transformer (EAT) to further improve the effectiveness and efficiency in audio SSL. The proposed EAT adopts the bootstrap self-supervised training paradigm to the audio domain. A novel Utterance-Frame Objective (UFO) is designed to enhance the modeling capability of acoustic events. Furthermore, we reveal that the masking strategy is critical in audio SSL pre-training, and superior audio representations can be obtained with large inverse block masks. Experiment results demonstrate that EAT achieves state-of-the-art (SOTA) performance on a range of audio-related tasks, including AudioSet (AS-2M, AS-20K), ESC-50, and SPC-2, along with a significant pre-training speedup up to ∼15x compared to existing audio SSL models.

For agents

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

get_harvested_code_for_paper("2401.03497")
get_code_for_paper("2401.03497")
have("2401.03497")

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

curl https://syntology.ai/api/ran/2401.03497.json

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

[![Syntology run record](https://syntology.ai/api/ran/2401.03497.svg)](https://syntology.ai/paper/2401.03497)

Connect an agent — have() is free.