SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2102.02973 · AAAI · 2021

Show, Attend and Distill: Knowledge Distillation via Attention-based Feature Matching

Byeongho Heo, Sungrae Park, Mingi Ji

arXiv · PDF · Open in the Atlas

Code that ran

4 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 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; 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/2102.02973.json · badge: /api/ran/2102.02973.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
clovaai/attention-feature-distillation marked official in Papers with Code
Papers with Code archive
4 ran · 2 no recorded run
FunctionStatusWhere it lives
LinearTransformStudent Ran
ran on a synthesized input · constructed an object; no result was computed
clovaai/attention-feature-distillation/distill/AFD.py
code served (permissive licence) · get_code("bd57df9965444248")
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 an instance of its own class, LinearTransformStudent: the run constructed the object and did not call it, so no result was computed. 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)

class Args:
    def __init__(self):
        # t_shape: (N, C, H, W) for each teacher layer
        self.t_shapes = [
            (2, 3, 4, 5),  # batch=2, channels=3, height=4, width=5
            (2, 3, 4, 5),  # second teacher layer
        ]
        
        # unique_t_shapes used by Sample objects
        self.unique_t_shapes = [
            (2, 3, 4, 5),  # same as t_shapes
        ]
        
        # s_shape: (batch_size, feature_dim, ...) for each student layer
        # From code: channel_mean has shape (bs, C) after mean(3).mean(2)
        # key_layer expects s_shape[1] as input dimension
        self.s_shapes = [
            (2, 6),  # batch=2, feature_dim=6
            (2, 7),  # batch=2, feature_dim=7
        ]
        
        self.qk_dim = 8
        self.s = len(self.s_shapes)
        self.t = len(self.t_shapes)

args = Args()

ARGS = {
    'args': args,
}
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 LinearTransformStudent — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a LinearTransformStudent 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
LinearTransformTeacher Ran
ran on a synthesized input · constructed an object; no result was computed
clovaai/attention-feature-distillation/distill/AFD.py
code served (permissive licence) · get_code("4f18195b66b3a6c3")
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 an instance of its own class, LinearTransformTeacher: the run constructed the object and did not call it, so no result was computed. 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)

class Args:
    def __init__(self):
        # qk_dim is used in nn_bn_relu initialization
        self.qk_dim = 5
        # t_shapes is a list of shapes for teacher feature maps
        # Each shape should be (batch, channels, height, width)
        # We need shapes that match how they're used in forward():
        # - channel_mean: mean over dims 2,3 -> shape (bs, channels)
        # - spatial_mean: pow(2).mean(1) -> shape (bs, channels*height*width)
        self.t_shapes = [
            (2, 4, 3, 3),  # batch=2, channels=4, height=3, width=3
            (2, 6, 2, 2),  # batch=2, channels=6, height=2, width=2
        ]

args = Args()

ARGS = {
    "args": args
}
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 LinearTransformTeacher — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a LinearTransformTeacher 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
Sample Ran
ran on a synthesized input · constructed an object; no result was computed
clovaai/attention-feature-distillation/distill/AFD.py
code served (permissive licence) · get_code("63ab06a9f4beab91")
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 an instance of its own class, Sample: the run constructed the object and did not call it, so no result was computed. 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)

class Sample(nn.Module):
    def __init__(self, t_shape):
        super(Sample, self).__init__()
        t_N, t_C, t_H, t_W = t_shape
        self.sample = nn.AdaptiveAvgPool2d((t_H, t_W))

    def forward(self, g_s, bs):
        g_s = torch.stack([self.sample(f_s.pow(2).mean(1, keepdim=True)).view(bs, -1) for f_s in g_s], dim=1)
        return g_s

# Create valid arguments for Sample.__init__
# t_shape should be (t_N, t_C, t_H, t_W) where these are integers
t_N = 2
t_C = 3
t_H = 4
t_W = 5

ARGS = {
    't_shape': (t_N, t_C, t_H, t_W)
}
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 Sample — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Sample 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
nn_bn_relu Ran
ran on a synthesized input
clovaai/attention-feature-distillation/distill/AFD.py
code served (permissive licence) · get_code("47a6f3fb9c35225c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, 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 torch
import torch.nn as nn

class nn_bn_relu(nn.Module):

    def __init__(self, nin, nout):
        super(nn_bn_relu, self).__init__()
        self.linear = nn.Linear(nin, nout)
        self.bn = nn.BatchNorm1d(nout)
        self.relu = nn.ReLU(False)

    def forward(self, x, relu=True):
        if relu:
            return self.relu(self.bn(self.linear(x)))
        return self.bn(self.linear(x))


INIT_ARGS = {
    'nin': 4,
    'nout': 3,
}

ARGS = {
    'x': torch.randn(2, 4),
    'relu': True,
}

model = nn_bn_relu(**INIT_ARGS)
output = model(**ARGS)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
AFD Not yet run
unverified (harvested, no recorded run)
clovaai/attention-feature-distillation/distill/AFD.py
code served (permissive licence) · get_code("3c1cc274644e0b0c")
How it ran

no invocation recorded.

Attention Not yet run
unverified (harvested, no recorded run)
clovaai/attention-feature-distillation/distill/AFD.py
code served (permissive licence) · get_code("44c8fee3462b48ea")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 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.

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

Knowledge distillation extracts general knowledge from a pretrained teacher network and provides guidance to a target student network. Most studies manually tie intermediate features of the teacher and student, and transfer knowledge through pre-defined links. However, manual selection often constructs ineffective links that limit the improvement from the distillation. There has been an attempt to address the problem, but it is still challenging to identify effective links under practical scenarios. In this paper, we introduce an effective and efficient feature distillation method utilizing all the feature levels of the teacher without manually selecting the links. Specifically, our method utilizes an attention-based meta-network that learns relative similarities between features, and applies identified similarities to control distillation intensities of all possible pairs. As a result, our method determines competent links more efficiently than the previous approach and provides better performance on model compression and transfer learning tasks. Further qualitative analyses and ablative studies describe how our method contributes to better distillation. The implementation code is available at open sourced 1 .

For agents

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

get_harvested_code_for_paper("2102.02973")
get_code_for_paper("2102.02973")
have("2102.02973")

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

curl https://syntology.ai/api/ran/2102.02973.json

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

[![Syntology run record](https://syntology.ai/api/ran/2102.02973.svg)](https://syntology.ai/paper/2102.02973)

Connect an agent — have() is free.