Pith. sign in

REVIEW 1 cited by

TR-BERT: Dynamic Token Reduction for Accelerating BERT Inference

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2105.11618 v1 pith:DKZEWOV6 submitted 2021-05-25 cs.CL

TR-BERT: Dynamic Token Reduction for Accelerating BERT Inference

classification cs.CL
keywords tr-berttokeninferenceplmsreductionbertdynamiclayer
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Existing pre-trained language models (PLMs) are often computationally expensive in inference, making them impractical in various resource-limited real-world applications. To address this issue, we propose a dynamic token reduction approach to accelerate PLMs' inference, named TR-BERT, which could flexibly adapt the layer number of each token in inference to avoid redundant calculation. Specially, TR-BERT formulates the token reduction process as a multi-step token selection problem and automatically learns the selection strategy via reinforcement learning. The experimental results on several downstream NLP tasks show that TR-BERT is able to speed up BERT by 2-5 times to satisfy various performance demands. Moreover, TR-BERT can also achieve better performance with less computation in a suite of long-text tasks since its token-level layer number adaption greatly accelerates the self-attention operation in PLMs. The source code and experiment details of this paper can be obtained from https://github.com/thunlp/TR-BERT.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. SemToken: Semantic-Aware Tokenization for Efficient Long-Context Language Modeling

    cs.CL 2025-08 reject novelty 4.0

    A semantic-aware tokenizer that merges similar and low-entropy text spans cuts long-context token counts by up to 59% and inference latency by roughly 2x, with no reported quality loss.