Papers › SpecHub: Provable Acceleration to Multi-Draft Speculative Decoding

SpecHub: Provable Acceleration to Multi-Draft Speculative Decoding

8 Nov 2024arXiv:2411.05289archive 2025-07-28

Ryan Sun, Tianyi Zhou, Xun Chen, Lichao Sun

Large Language Models (LLMs) have become essential in advancing natural language processing (NLP) tasks, but their sequential token generation limits inference speed. Multi-Draft Speculative Decoding (MDSD) offers a promising solution by using a smaller draft model to generate multiple token sequences, which the target LLM verifies in parallel. However, current heuristic approaches, such as Recursive Rejection Sampling (RRS), suffer from low acceptance rates in subsequent drafts, limiting the advantages of using multiple drafts. Meanwhile, Optimal Transport with Membership Cost (OTM) can theoretically improve acceptance rates, but its computational cost is too high for real-time use. We present SpecHub, a novel, efficient sampling-verification method for MDSD that improves acceptance rates with only linear computational overhead. By simplifying the OTM problem into a compact Linear Programming model, SpecHub significantly reduces computational complexity. It further accelerates sampling by leveraging a sparse joint distribution, focusing computation on high-probability token sequences. In extensive experiments, Spechub consistently generates 0.05-0.27 and 0.02-0.16 more tokens per step than RRS and RRS without replacement. We attach our code at https://github.com/MasterGodzilla/Speculative_decoding_OT.

PaperPDFCodeCode Syntology ranMethod as code

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

Syntology 21 code samples harvested from 1 repository linked to this paper: 9 ran (of which 7 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 2 where Syntology's instrument failed) · 12 unverified. Where Syntology's instrument failed: 2 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 21 samples from 1 repository, 9 ran (7 with no instrument failure; 2 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.

Levels written on 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

mastergodzilla/speculative_decoding_ot officialmentioned in paperpytorchSyntology: 21 samples, 9 ran (2 instrument-failed) · 12 unverified 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.

Code Syntology ran Syntology

21 samples harvested: 9 ran (of which 7 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 2 where Syntology's instrument failed) · 12 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

2ran; the contract check could not run
7ran
12unverified

Licence: 21 of the 21 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 mastergodzilla/speculative_decoding_ot. “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”. For 2 samples that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “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.

mastergodzilla/speculative_decoding_ot official repository 21 samples, 9 ran (2 instrument-failed) · 12 unverified
KV_Cache Sequoia/Tree/SpecTree.py ran no licence file found · pointer only · 1080bbf045764637 · report fixture: returned KV_Cache · level recorded 2026-09-22
LlamaAttention_FI Sequoia/Tree/SpecTree.py ran no licence file found · pointer only · 0f36d7967c772ba7 · report fixture: returned LlamaAttention_FI · level recorded 2026-09-22
LlamaAttention_TG Sequoia/Tree/SpecTree.py ran no licence file found · pointer only · a7975bacb53acbd6 · report fixture: returned LlamaAttention_TG · level recorded 2026-09-22
LlamaMLP_FI Sequoia/Tree/SpecTree.py ran fingerprinted no licence file found · pointer only · 4d241e71f8a7a763 · report fixture: returned LlamaMLP_FI · level recorded 2026-09-22
LlamaRMSNorm_FI Sequoia/Tree/SpecTree.py ran no licence file found · pointer only · 6240a83f9a6521cf · report fixture: returned LlamaRMSNorm_FI · level recorded 2026-09-22
LlamaRotaryEmbedding_FI Sequoia/Tree/SpecTree.py ran no licence file found · pointer only · 39802465997cc25d · report fixture: returned LlamaRotaryEmbedding_FI · level recorded 2026-09-22
Tree in 10 papers Sequoia/Tree/SpecTree.py ran no licence file found · pointer only · 6c730d4a44e0ced5 · report fixture: returned Tree · level recorded 2026-09-22
_make_causal_mask in 8 papers Sequoia/Tree/SpecTree.py ran; the contract check could not run (RAISES) no licence file found · pointer only · 3d5d12b34c6bc8f3 · report fixture: signature input_ids_shape:Size/1/int, dtype:dtype/0/?, device:device/0/? · returned Tensor · level recorded 2026-09-22
get_sampling_logits Sequoia/Tree/SpecTree.py ran; the contract check could not run (RAISES) no licence file found · pointer only · 1823e59e5f171174 · report fixture: signature logits:Tensor/2/float, top_p:float/0/float, T:float/0/float, replicate:bool/0/bool · returned Tensor · level recorded 2026-09-22
GraphInferenceEngine Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · b706495d45bf74dd · report
GraphInferenceEngineTG Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · b251248cec118216 · report
InferenceEngine Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · 85a9b61d47defe2d · report
InferenceEngineTG Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · c652c9e8d68d727a · report
LlamaDecoderLayer_FI Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · 774786e151f678c0 · report
LlamaDecoderLayer_TG Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · dc6a8ea4f5fcee7a · report
LlamaForCausalLM_FI Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · 8653a0e682ed6728 · report
LlamaForCausalLM_TG Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · e6b7a28738c8bd31 · report
LlamaModel_FI Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · 4580bc94dbaf3361 · report
LlamaModel_TG Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · 68e76f48680e026b · report
SpecTree Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · 718aa0b1a1a6203e · report
capture_graph Sequoia/Tree/SpecTree.py unverified no licence file found · pointer only · c07e4861bb7ab066 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for SpecHub (graph method: SpecHub). 21 bricks; individually on a fixture: 9 ran (of which 7 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 2 where Syntology's instrument failed) · 12 unverified. The assembly has not been run end-to-end.

  1. SpecTree entry · level 0 · unverified · 718aa0b1a1a6203e
  2. GraphInferenceEngine model · level 0 · unverified · b706495d45bf74dd
  3. GraphInferenceEngineTG model · level 0 · unverified · b251248cec118216
  4. Tree helper · level 1 · ran · 6c730d4a44e0ced5
  5. get_sampling_logits helper · level 1 · ran; the contract check could not run (RAISES) · 1823e59e5f171174
  6. InferenceEngine helper · level 0 · unverified · 85a9b61d47defe2d
  7. capture_graph helper · level 0 · unverified · c07e4861bb7ab066
  8. InferenceEngineTG helper · level 0 · unverified · c652c9e8d68d727a
  9. _make_causal_mask helper · level 1 · ran; the contract check could not run (RAISES) · 3d5d12b34c6bc8f3
  10. KV_Cache helper · level 1 · ran · 1080bbf045764637
  11. LlamaForCausalLM_FI helper · level 0 · unverified · 8653a0e682ed6728
  12. LlamaForCausalLM_TG helper · level 0 · unverified · e6b7a28738c8bd31
  13. LlamaModel_FI helper · level 0 · unverified · 4580bc94dbaf3361
  14. LlamaModel_TG helper · level 0 · unverified · 68e76f48680e026b
  15. LlamaDecoderLayer_FI helper · level 0 · unverified · 774786e151f678c0
  16. LlamaRMSNorm_FI helper · level 1 · ran · 6240a83f9a6521cf
  17. LlamaDecoderLayer_TG helper · level 0 · unverified · dc6a8ea4f5fcee7a
  18. LlamaAttention_FI helper · level 1 · ran · 0f36d7967c772ba7
  19. LlamaMLP_FI helper · level 1 · ran · 4d241e71f8a7a763
  20. LlamaAttention_TG helper · level 1 · ran · a7975bacb53acbd6
  21. LlamaRotaryEmbedding_FI helper · level 1 · ran · 39802465997cc25d

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2411.05289") returns these assemblies with each brick's role and wiring (how to connect).

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

OTM

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