SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.15610 · 2022

LightHuBERT: Lightweight and Configurable Speech Representation Learning with Once-for-All Hidden-Unit BERT

arXiv · PDF · Open in the Atlas

Code that ran

5 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 4 have no contract verdict; 1 failure of our instrument; 3 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/2203.15610.json · badge: /api/ran/2203.15610.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
mechanicalsea/lighthubert canonical (classifier label)
classifier basis not recorded on the link
5 ran · 3 no recorded run
FunctionStatusWhere it lives
index_put Ran
ran on a synthesized input
mechanicalsea/lighthubert/lighthubert/functional/fairseq_utils.py
code served (permissive licence) · get_code("28bed6ccb2f1b89a")
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

torch.manual_seed(0)

# For the non-XLA case (CPU tensors)
tensor = torch.randn(2, 4)
indices = torch.tensor([[True, False, True, False], [False, True, False, True]])
value = torch.tensor(99.0)

ARGS = {
    "tensor": tensor,
    "indices": indices,
    "value": value,
}
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
is_xla_tensor Ran
ran on a synthesized input
mechanicalsea/lighthubert/lighthubert/functional/fairseq_utils.py
code served (permissive licence) · get_code("29c1031dcedd09de")
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 bool, 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

torch.manual_seed(0)

# The function checks if a tensor is a torch tensor and has device.type == 'xla'
# We need to pass a regular CPU tensor (since we can't use XLA in this sandbox)
# The function will return False for CPU tensors, which is expected behavior

tensor_cpu = torch.randn(2, 4)

ARGS = {
    "tensor": tensor_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 bool, 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
mask_padding Ran
ran on a synthesized input
mechanicalsea/lighthubert/lighthubert/functional/sliding_attn.py
code served (permissive licence) · get_code("d48916e1aac4a4de")
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, 5) 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)

# Create a small 1-d long tensor representing sequence lengths for a batch
# batch_size = 2, with different sequence lengths
seq_len = torch.LongTensor([3, 5])

# Optionally set max_len, or leave as None to let the function compute it
max_len = None

ARGS = {
    "seq_len": seq_len,
    "max_len": max_len,
}
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, 5) — shape (2, 5) 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
pad_as_attn_swz Ran
ran on a synthesized input
mechanicalsea/lighthubert/lighthubert/functional/sliding_attn.py
code served (permissive licence) · get_code("3bbfd682fbeb9a6d")
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
import torch.nn.functional as F

torch.manual_seed(0)

# Create small tensors for query, key, value
batch_size = 2
hidden_dim = 4
query_len = 5
key_len = 6

query = torch.randn(batch_size, query_len, hidden_dim)
key = torch.randn(batch_size, key_len, hidden_dim)
value = torch.randn(batch_size, key_len, hidden_dim)

# Sliding window size
swz = 3

# Create attention mask (3-d version)
attn_mask = torch.zeros(batch_size, query_len, key_len, dtype=torch.bool)

# Create key padding mask (2-d)
key_padding_mask = torch.zeros(batch_size, key_len, dtype=torch.bool)

ARGS = {
    "query": query,
    "key": key,
    "value": value,
    "swz": swz,
    "attn_mask": attn_mask,
    "key_padding_mask": key_padding_mask,
}
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
quant_noise Ran
ran; our contract draft was wrong, not the code
mechanicalsea/lighthubert/lighthubert/modules/fairseq_modules.py
code served (permissive licence) · get_code("50d074f5c3f3407d")
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 Linear. 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

torch.manual_seed(0)

# Create a simple Linear module for testing
linear_module = nn.Linear(in_features=8, out_features=4)

# Create a simple Embedding module for testing
embedding_module = nn.Embedding(num_embeddings=10, embedding_dim=8)

# Create a simple Conv2d module with 1x1 kernel
conv_1x1_module = nn.Conv2d(in_channels=8, out_channels=4, kernel_size=1)

# Create a simple Conv2d module with larger kernel
conv_3x3_module = nn.Conv2d(in_channels=8, out_channels=4, kernel_size=3, padding=1)

ARGS = {
    "module": linear_module,
    "p": 0.5,
    "block_size": 2,
}
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 Linear — a Linear output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Linear 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_mask_indices Not yet run
unverified (harvested, no recorded run)
mechanicalsea/lighthubert/lighthubert/lighthubert.py
code served (permissive licence) · get_code("6eeb1f787769a05e")
How it ran

no invocation recorded.

merge_padding_attm_mask Not yet run
unverified (harvested, no recorded run)
mechanicalsea/lighthubert/lighthubert/functional/sliding_attn.py
code served (permissive licence) · get_code("fb33b4608f21a92a")
How it ran

no invocation recorded.

pad_to_multiple Not yet run
unverified (harvested, no recorded run)
mechanicalsea/lighthubert/lighthubert/functional/fairseq_utils.py
pointer only (licence: MIT); the stored copy get_code serves, from speech-lab-iitm/ccc-wav2vec-2.0, is pointer only (licence: MIT) · get_code("b57d92d8a18d12a3")
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 speech-lab-iitm/ccc-wav2vec-2.0, and MIT, as GitHub detects it on that repository, is in the permissive set, but a licence decision keeps the body a pointer: its licence file names an outside party, so it may cover vendored code rather than the repository owner's files (recorded as pointer:licence_names_third_party_corporation). 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. 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, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 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
quant_noise 4 other papers
4 later · 1 only as an assembly brick
2307.13424 (2023-07)
Holistic Exploration on Universal Decompositional Semantic Parsing: Architecture, Data Augmentation, and LLM Paradigm
2407.03575 (2024-07)
DGR-MIL: Exploring Diverse Global Representation in Multiple Instance Learning for Whole Slide Image Classification
2602.15537 (2026-02)
ZeroSyl: Simple Zero-Resource Syllable Tokenization for Spoken Language Modeling
and 1 more not named here
pad_to_multiple 2 other papers
2 later
2207.00555 (2022-07)
FitHuBERT: Going Thinner and Deeper for Knowledge Distillation of Speech Self-Supervised Learning
2210.02592 (2022-10)
CCC-wav2vec 2.0: Clustering aided Cross Contrastive Self-supervised learning of speech representations

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

Self-supervised speech representation learning has shown promising results in various speech processing tasks. However, the pre-trained models, e.g., HuBERT, are storage-intensive Transformers, limiting their scope of applications under low-resource settings. To this end, we propose LightHuBERT, a once-for-all Transformer compression framework, to find the desired architectures automatically by pruning structured parameters. More precisely, we create a Transformer-based supernet that is nested with thousands of weight-sharing subnets and design a two-stage distillation strategy to leverage the contextualized latent representations from HuBERT. Experiments on automatic speech recognition (ASR) and the SUPERB benchmark show the proposed LightHuBERT enables over $10^9$ architectures concerning the embedding dimension, attention dimension, head number, feed-forward network ratio, and network depth. LightHuBERT outperforms the original HuBERT on ASR and five SUPERB tasks with the HuBERT size, achieves comparable performance to the teacher model in most tasks with a reduction of 29% parameters, and obtains a $3.5\times$ compression ratio in three SUPERB tasks, e.g., automatic speaker verification, keyword spotting, and intent classification, with a slight accuracy loss. The code and pre-trained models are available at https://github.com/mechanicalsea/lighthubert.

For agents

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

get_harvested_code_for_paper("2203.15610")
get_code_for_paper("2203.15610")
have("2203.15610")

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

curl https://syntology.ai/api/ran/2203.15610.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.15610.svg)](https://syntology.ai/paper/2203.15610)

Connect an agent — have() is free.