feat: Use BucketVocabStore in PipelineBPE#2188
Open
SBrandeis wants to merge 19 commits into
Open
Conversation
SBrandeis
marked this pull request as ready for review
July 11, 2026 09:14
McPatate
force-pushed
the
feat/no-alloc-model-2
branch
from
July 15, 2026 17:44
528e154 to
87b2740
Compare
Base automatically changed from
feat/no-alloc-model-2
to
feat/train_encode_split
July 15, 2026 17:45
…zers into feat/bucket-vocab-bpe
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
TL;DR
Use BucketVocabStore instead of VocabbStore in the BPE model
PipelineTokenizer benchmark
7 / 8 models supported — PipelineTokenizer vs
tokenizersv0.23.1 (latest release) · ~10 kB inputs · single thread + 1/2/4/8/max-thread sweepd1b26a01a · 2026-07-17 15:24 UTC· Intel(R) Xeon(R) Platinum 8375C CPU @ 2.90GHz · 16 coresvs base branch (
2c4f6100b) — per-model geomean ×speedup of this PR's PipelineTokenizer against the base branch's; regressions in red.bert-base-uncased — normalizer-heavy WordPiece · ×5.54 vs v0.23.1 · ×0.97 vs base
Memory (RSS MB, load+encode): v0.23.1 2+0 (peak 6) · Pipeline 7+0 (peak 7)
deepseek-v4 — deepseek 3-regex split-heavy byte-level BPE · ×5.49 vs v0.23.1 · ×1.01 vs base
Memory (RSS MB, load+encode): v0.23.1 52+0 (peak 58) · Pipeline 75+0 (peak 74)
Pre-tokenize:
classify + fsmvs regex engines — ns/byte, lower better. The fsm is the scalar jump-table in both pipe columns; SIMD / scalar is the classify pass (regex pre-tokenizers have no SIMD fsm).×vs= engine ÷ our pipeline (SIMD / scalar classify);onig&pcre2(JIT) are C,fancyis pure-Rust fancy-regex,logosis a compile-time DFA lexer (approximate grammar; n/a for deepseek).gpt2 — gpt2 ByteLevel regex · ×9.86 vs v0.23.1 · ×1.01 vs base
Memory (RSS MB, load+encode): v0.23.1 17+2 (peak 19) · Pipeline 20+0 (peak 20)
Pre-tokenize:
classify + fsmvs regex engines — ns/byte, lower better. The fsm is the scalar jump-table in both pipe columns; SIMD / scalar is the classify pass (regex pre-tokenizers have no SIMD fsm).×vs= engine ÷ our pipeline (SIMD / scalar classify);onig&pcre2(JIT) are C,fancyis pure-Rust fancy-regex,logosis a compile-time DFA lexer (approximate grammar; n/a for deepseek).gpt-oss — o200k-regex byte-level BPE (gpt-oss) · ×8.01 vs v0.23.1 · ×0.99 vs base
Memory (RSS MB, load+encode): v0.23.1 2+4 (peak 6) · Pipeline 3+0 (peak 6)
Pre-tokenize:
classify + fsmvs regex engines — ns/byte, lower better. The fsm is the scalar jump-table in both pipe columns; SIMD / scalar is the classify pass (regex pre-tokenizers have no SIMD fsm).×vs= engine ÷ our pipeline (SIMD / scalar classify);onig&pcre2(JIT) are C,fancyis pure-Rust fancy-regex,logosis a compile-time DFA lexer (approximate grammar; n/a for deepseek).glm-5.2 — cl100k-variant regex byte-level BPE (glm-5.2) · ×13.35 vs v0.23.1 · ×0.98 vs base
Memory (RSS MB, load+encode): v0.23.1 2+4 (peak 6) · Pipeline 3+0 (peak 6)
Pre-tokenize:
classify + fsmvs regex engines — ns/byte, lower better. The fsm is the scalar jump-table in both pipe columns; SIMD / scalar is the classify pass (regex pre-tokenizers have no SIMD fsm).×vs= engine ÷ our pipeline (SIMD / scalar classify);onig&pcre2(JIT) are C,fancyis pure-Rust fancy-regex,logosis a compile-time DFA lexer (approximate grammar; n/a for deepseek).llama-2 — model-bounded BPE, no pre-tokenizer · ×4.06 vs v0.23.1 · ×0.99 vs base
Memory (RSS MB, load+encode): v0.23.1 15+0 (peak 15) · Pipeline 14+0 (peak 14)
llama-3 — cl100k-regex byte-level BPE (llama-3), single regex · ×9.28 vs v0.23.1 · ×1.01 vs base
Memory (RSS MB, load+encode): v0.23.1 129+0 (peak 189) · Pipeline 130+0 (peak 189)
Pre-tokenize:
classify + fsmvs regex engines — ns/byte, lower better. The fsm is the scalar jump-table in both pipe columns; SIMD / scalar is the classify pass (regex pre-tokenizers have no SIMD fsm).×vs= engine ÷ our pipeline (SIMD / scalar classify);onig&pcre2(JIT) are C,fancyis pure-Rust fancy-regex,logosis a compile-time DFA lexer (approximate grammar; n/a for deepseek).Not yet supported:
t5-base