Papers › Patient Knowledge Distillation for BERT Model Compression

Patient Knowledge Distillation for BERT Model Compression

25 Aug 2019IJCNLP 2019 11arXiv:1908.09355archive 2025-07-28

Siqi Sun, Yu Cheng, Zhe Gan, Jingjing Liu

Pre-trained language models such as BERT have proven to be highly effective for natural language processing (NLP) tasks. However, the high demand for computing resources in training such models hinders their application in practice. In order to alleviate this resource hunger in large-scale model training, we propose a Patient Knowledge Distillation approach to compress an original large model (teacher) into an equally-effective lightweight shallow network (student). Different from previous knowledge distillation methods, which only use the output from the last layer of the teacher network for distillation, our student model patiently learns from multiple intermediate layers of the teacher model for incremental knowledge extraction, following two strategies: (i) PKD-Last: learning from the last k layers; and (ii) PKD-Skip: learning from every k layers. These two patient distillation schemes enable the exploitation of rich information in the teacher's hidden layers, and encourage the student model to patiently learn from and imitate the teacher through a multi-layer distillation process. Empirically, this translates into improved results on multiple NLP tasks with significant gain in training efficiency, without sacrificing model accuracy.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

The archive's official repository (intersun/PKD-for-BERT-Model-Compression) was not harvested; the 27 samples below come from 2 community repositories.

Syntology 27 code samples harvested from 2 repositories linked to this paper: 19 ran (of which 0 constructed an object rather than computing a result; 13 with no instrument failure: 3 honoured, 1 violated, 9 with no contract checked; 6 where Syntology's instrument failed) · 8 unverified. Where Syntology's instrument failed: 5 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 27 samples from 2 repositories, 19 ran (13 with no instrument failure; 6 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 2 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 19 samples that ran have no recorded date.

intersun/PKD-for-BERT-Model-Compression officialmentioned in papermentioned on GitHubpytorchSyntology: not harvested report
Daniel-H-99/Patient-Knowledge-Distillation mentioned on GitHubpytorchMITSyntology: 17 samples, 13 ran (3 instrument-failed) · 4 unverified report
eunanomist/PKD_BERT mentioned on GitHubpytorchMITSyntology: 10 samples, 6 ran (3 instrument-failed) · 4 unverified report
MindSpore-scientific/code-11 mindsporeSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to SignSGD with Federated Voting (arXiv:2403.16372; that paper's own run record: 3 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
MindSpore-scientific/code-13 mindsporeSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to SVGDreamer: Text Guided SVG Generation with Diffusion Model (arXiv:2312.16476; that paper's own run record: 10 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 2 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

27 samples harvested: 19 ran (of which 0 constructed an object rather than computing a result; 13 with no instrument failure: 3 honoured, 1 violated, 9 with no contract checked; 6 where Syntology's instrument failed) · 8 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

3ran · honoured contract
1ran · violated contract
5ran · our draft was wrong
1ran; the contract check could not run
9ran
8unverified

Licence: 27 of the 27 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

Daniel-H-99/Patient-Knowledge-Distillation community (archive-listed) 17 samples, 13 ran (3 instrument-failed) · 4 unverified
swish in 258 papers
census: 14 implementations from 37 papers, 2 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
BERT/pytorch_pretrained_bert/modeling.py ran · our draft was wrong fingerprinted MIT recorded; this copy not marked cleared · pointer only · 0f786c407fb1ee4c · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded, date not recorded
acc_and_f1 in 39 papers src/nli_data_processing.py ran; the contract check could not run (RAISES) · metamorphic tier: deterministic MIT recorded; this copy not marked cleared · pointer only · cac113ca87b9d9f3 · report fixture: signature preds:ndarray/1/int, labels:ndarray/1/int · returned dict · level recorded 2026-09-22
boolean_string in 17 papers src/argument_parser.py ran · violated contract MIT recorded; this copy not marked cleared · pointer only · b1f9cebf0d177140 · report fixture: signature s:str/0/str · returned bool · level recorded 2026-09-22
convert_examples_to_features in 134 papers src/nli_data_processing.py ran MIT recorded; this copy not marked cleared · pointer only · 379ade5858acdbfa · report fixture: signature examples:list/1/SimpleExample, label_list:list/1/str, max_seq_length:int/0/int, tokenizer:SimpleTokenizer/0/?, output_mode:str/0/str · returned list · level recorded 2026-09-24
convert_examples_to_features in 134 papers src/race_data_processing.py ran MIT recorded; this copy not marked cleared · pointer only · b30ee2907e5cd3a6 · report fixture: signature examples:list/1/Example, tokenizer:MinimalTokenizer/0/?, max_seq_length:int/0/int, is_training:bool/0/bool · returned list · level recorded 2026-09-24
count_parameters in 278 papers src/utils.py ran MIT recorded; this copy not marked cleared · pointer only · 3ea89c558af6c4b4 · report fixture: signature model:callable/0/?, trainable_only:bool/0/bool, is_dict:bool/0/bool · returned int · level recorded 2026-09-24
distillation_loss in 10 papers src/KD_loss.py ran MIT recorded; this copy not marked cleared · pointer only · 5ff7ab3a0c0d6da0 · report fixture: signature y:Tensor/2/float, labels:Tensor/1/int, teacher_scores:Tensor/2/float, T:float/0/float, alpha:float/0/float, reduction_kd:str/0/str, reduction_nll:str/0/str · returned tuple · level recorded 2026-09-24
gelu in 308 papers
census: 43 implementations from 74 papers, 6 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
BERT/pytorch_pretrained_bert/modeling_gpt2.py ran · our draft was wrong fingerprinted MIT recorded; this copy not marked cleared · pointer only · 8d23fbe2b99b840b · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded, date not recorded
gelu in 308 papers
census: 43 implementations from 74 papers, 6 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
BERT/pytorch_pretrained_bert/modeling.py ran · honoured contract fingerprinted MIT recorded; this copy not marked cleared · pointer only · fdc64f4c72036ae4 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
is_folder_empty src/argument_parser.py ran MIT recorded; this copy not marked cleared · pointer only · 615f129ea793785e · report fixture: signature folder_name:str/0/str · returned bool · level recorded 2026-09-24
load_model in 479 papers src/utils.py ran MIT recorded; this copy not marked cleared · pointer only · 9e42bf8972501658 · report fixture: signature model:callable/0/?, checkpoint:NoneType/0/?, args:ArgsConfig/0/?, mode:str/0/str, train_mode:str/0/str, verbose:bool/0/bool, DEBUG:bool/0/bool · returned DummyModel · level recorded 2026-09-24
patience_loss src/KD_loss.py ran MIT recorded; this copy not marked cleared · pointer only · 16afb11eb02e8e88 · report fixture: signature teacher_patience:Tensor/3/float, student_patience:Tensor/3/float, normalized_patience:bool/0/bool · returned Tensor · level recorded 2026-09-24
read_mrc_examples src/race_data_processing.py ran MIT recorded; this copy not marked cleared · pointer only · 28d61d48898d30c4 · report fixture: signature input_file:str/0/str · returned list · level recorded 2026-09-24
fill_tensor src/utils.py unverified MIT recorded; this copy not marked cleared · pointer only · 01803af97cd5feef · report
load_tf_weights_in_bert in 135 papers BERT/pytorch_pretrained_bert/modeling.py unverified MIT recorded; this copy not marked cleared · pointer only · b1a7180868dd8d6a · report
load_tf_weights_in_gpt2 in 68 papers BERT/pytorch_pretrained_bert/modeling_gpt2.py unverified MIT recorded; this copy not marked cleared · pointer only · a62604f618a8cc82 · report
simple_accuracy in 66 papers src/nli_data_processing.py unverified MIT recorded; this copy not marked cleared · pointer only · 3c241ecfe3749a6d · report
eunanomist/PKD_BERT community (archive-listed) 10 samples, 6 ran (3 instrument-failed) · 4 unverified
build_tf_to_pytorch_map in 14 papers BERT/pytorch_pretrained_bert/modeling_transfo_xl.py ran · our draft was wrong MIT recorded; this copy not marked cleared · pointer only · 35ef111fc00c21b4 · report fixture: signature model:callable/0/?, config:Config/0/? · returned dict · level recorded 2026-09-22
sample_logits in 16 papers BERT/pytorch_pretrained_bert/modeling_transfo_xl_utilities.py ran · our draft was wrong MIT recorded; this copy not marked cleared · pointer only · 93228a3eb5c4d179 · report fixture: signature embedding:callable/0/?, bias:Tensor/1/float, labels:Tensor/2/int, inputs:Tensor/3/float, sampler:LogUniformSampler/0/? · returned Tensor · level recorded 2026-09-22
url_to_filename in 73 papers BERT/pytorch_pretrained_bert/file_utils.py ran · our draft was wrong fingerprinted MIT recorded; this copy not marked cleared · pointer only · 5eff22fa0a651276 · report fixture: signature url:str/0/str, etag:str/0/str · returned str · level recorded 2026-09-22
warmup_constant in 30 papers BERT/pytorch_pretrained_bert/optimization.py ran · honoured contract fingerprinted MIT recorded; this copy not marked cleared · pointer only · 59e4e730a9dc5dc6 · report fixture: signature x:float/0/float, warmup:float/0/float · returned float · level recorded 2026-09-26
warmup_cosine in 30 papers BERT/pytorch_pretrained_bert/optimization.py ran fingerprinted MIT recorded; this copy not marked cleared · pointer only · 007a8d2955fb555d · report fixture: signature x:float/0/float, warmup:float/0/float · returned float · level recorded 2026-09-25
warmup_linear in 45 papers BERT/pytorch_pretrained_bert/optimization.py ran · honoured contract fingerprinted MIT recorded; this copy not marked cleared · pointer only · e0960afdb0d64aa7 · report fixture: signature x:float/0/float, warmup:float/0/float · returned float · level recorded 2026-09-22
cached_path in 47 papers BERT/pytorch_pretrained_bert/file_utils.py unverified MIT recorded; this copy not marked cleared · pointer only · df9f10631fa5d322 · report
filename_to_url in 54 papers BERT/pytorch_pretrained_bert/file_utils.py unverified MIT recorded; this copy not marked cleared · pointer only · 6db16fe8f67e56b6 · report
load_tf_weights_in_openai_gpt in 13 papers BERT/pytorch_pretrained_bert/modeling_openai.py unverified MIT recorded; this copy not marked cleared · pointer only · 46f305675277e757 · report
load_tf_weights_in_transfo_xl in 12 papers BERT/pytorch_pretrained_bert/modeling_transfo_xl.py unverified MIT recorded; this copy not marked cleared · pointer only · 1ebc2bf2c3cddc4d · report

Tasks

Knowledge DistillationModel Compressionmodel

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamAttentionAttention DropoutBERTDense ConnectionsDropoutKnowledge DistillationLayer NormalizationLinear LayerLinear Warmup With Linear DecayMulti-Head AttentionResidual ConnectionSoftmaxWeight DecayWordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections