sglang

cpp_ngram

Remove fields that are never used in spec/engine/disagg (#25434 )

2026-05-16 09:18:00 +08:00

triton_ops

Reland spec v2 tree drafting (eagle topk>1) with page_size==1 (#26866 ) (#26997 )

2026-06-03 15:40:05 -04:00

adaptive_runtime_state.py

[SPEC][5/N] feat: batchsize-aware support for adaptive speculative_num_steps (#24055 )

2026-06-05 15:43:02 -07:00

adaptive_spec_params.py

[SPEC][5/N] feat: batchsize-aware support for adaptive speculative_num_steps (#24055 )

2026-06-05 15:43:02 -07:00

base_spec_worker.py

[SPEC][5/N] feat: batchsize-aware support for adaptive speculative_num_steps (#24055 )

2026-06-05 15:43:02 -07:00

dflash_info.py

Fix the misnamed request finish-check method to reflect its mutating semantics (#25725 )

2026-05-19 09:22:28 +08:00

dflash_utils.py

Move module-level helpers out of scheduler.py (#25638 )

2026-05-18 18:45:38 +08:00

dflash_worker.py

[NPU] DFlash Speculative Decoding Support NPU (#23122 )

2026-05-30 15:13:59 +08:00

draft_utils.py

Support Flashinfer Cute-DSL MLA attention (#24737 )

2026-05-28 00:21:32 -07:00

eagle_disaggregation.py

Allow custom speculative algorithm to support disaggregation (#26195 )

2026-05-27 09:54:53 -07:00

eagle_draft_cuda_graph_runner.py

Replace skip_attn_backend_init with a batch-carried attention plan marker (+ staleness re-plan) (#27193 )

2026-06-04 17:13:18 -07:00

eagle_draft_extend_cuda_graph_runner.py

[spec_v2] Enable trtllm_mha draft-extend CUDA graph with v2 semantics (#25002 )

2026-06-04 17:50:12 -07:00

eagle_info_v2.py

[spec] Consolidate the per-decode KV alloc reserve into one helper (#27458 )

2026-06-06 14:58:50 -07:00

eagle_info.py

[SPEC] fix: import copy module for eagle sampling info clone (#27390 )

2026-06-06 12:12:46 +00:00

eagle_utils.py

[Perf][Spec Decoding] Skip cat/topk/sort/gather in draft_forward for topk=1 (#26424 )

2026-06-01 15:37:47 -07:00

eagle_worker_v2.py

Support topk > 1 tree drafting for mamba/hybrid-linear models on spec v2 (#27463 )

2026-06-07 17:04:09 -07:00

eagle_worker.py

[Bugfix] Restore overridden HF config fields and support index_skip_topk_offset for DSA topk sharing (#27114 )

2026-06-06 13:26:04 +08:00

external_corpus_manager.py

[Spec][Ngram] Return token counts in list_external_corpora API (#22471 )

2026-04-10 21:50:02 -07:00

frozen_kv_mtp_cuda_graph_runner.py

Replace skip_attn_backend_init with a batch-carried attention plan marker (+ staleness re-plan) (#27193 )

2026-06-04 17:13:18 -07:00

frozen_kv_mtp_info.py

[Spec V1] Split draft-extend phase from EagleDraftInput into new EagleDraftExtendInput (#24859 )

2026-05-10 01:07:45 -07:00

frozen_kv_mtp_utils.py

feat(model_runner): remove pool/backend refs from ForwardBatch via ForwardContext (#25983 )

2026-05-21 14:01:49 -07:00

frozen_kv_mtp_worker_v2.py

[Gemma 4] Adding MTP support (#24436 )

2026-05-07 14:08:41 -07:00

frozen_kv_mtp_worker.py

FrozenKVMTPVerifyInput: add _draft_preprocess_idle call for when all requests in the verify batch finish in the same iteration (#26859 )

2026-06-04 21:47:32 -07:00

multi_layer_eagle_draft_extend_cuda_graph_runner.py

Unify full→SWA index translation in init_forward_metadata; drop pool caches (#27091 )

2026-06-03 16:12:27 -07:00

multi_layer_eagle_utils.py

[codex] Centralize Triton utility kernels (#26000 )

2026-06-02 16:47:45 +08:00

multi_layer_eagle_worker_v2.py

[spec] Misc defensive guards for EAGLE draft KV indexing (#27486 )

2026-06-07 21:27:46 -07:00

multi_layer_eagle_worker.py

[spec] Misc defensive guards for EAGLE draft KV indexing (#27486 )

2026-06-07 21:27:46 -07:00

ngram_info.py

[Bug] ngram verify: keep batch.seq_lens_sum in sync after accept (#26753 )

2026-05-29 18:00:11 -07:00

ngram_worker.py

refactor: remove ModelWorkerBatch indirection (#25516 )

2026-05-17 18:05:39 -07:00

spec_info.py

[core] Make spec_v2 seq_lens_cpu optional via backend needs_cpu_seq_lens; Triton opts out (#26128 )

2026-05-29 13:00:32 -07:00

spec_registry.py

fix(spec): complete CustomSpecAlgo duck-typing interface and guard against drift (#27300 )

2026-06-04 15:52:46 -07:00

spec_utils.py

[spec] Misc defensive guards for EAGLE draft KV indexing (#27486 )

2026-06-07 21:27:46 -07:00

standalone_worker_v2.py

[Perf][Spec Decoding] Skip cat/topk/sort/gather in draft_forward for topk=1 (#26424 )

2026-06-01 15:37:47 -07:00

standalone_worker.py

[Spec] fold can_run_cuda_graph into EagleVerifyOutput; drop dead extend-after-decode check (#25566 )

2026-05-18 14:24:54 -07:00