Giancarlo DelfinandGitHub 81bc196913 [Model Runner V2] Cache draft logits in model's LM head dtype (#50910) ...

Signed-off-by: Giancarlo Delfin <[email protected]>

2026-08-05 19:12:46 -07:00

..

__init__.py

Test: added acceptance length tests (#32030)

2026-01-20 18:55:15 +00:00

test_acceptance_length.py

Enable DeepSeek V4 and GLM-5.1 on SM120 (#43477)

2026-06-22 11:54:14 -07:00

test_backup_token_async_spec.py

[Bugfix][Async] Fix async spec decoding with hybrid models (#38556)

2026-03-31 11:08:54 -04:00

test_dflash_causality.py

[Bugfix] DFlash fc sized wrong when num_target_layers != num_hidden_layers (#48524)

2026-07-21 14:42:52 -07:00

test_dflash_lookahead.py

Fix DFlash prefix cache corruption due to missing lookahead block (#42971)

2026-06-02 12:06:33 +00:00

test_dspark_topk.py

[Spec Decode] Add top-k DSpark Markov projection (#49969)

2026-08-04 12:52:53 +00:00

test_dynamic_sd_cug.py

[MRV2][SD] Make Dynamic SD comatible with Full Cuda Graphs (#45953)

2026-07-03 23:58:27 -04:00

test_dynamic_sd.py

Disable dynamic speculative decoding when DP is enabled (#45963)

2026-07-07 12:52:41 +00:00

test_eagle_draft_attn_metadata.py

[Bugfix][Spec-Decode] Populate draft seq_lens_cpu_upper_bound for spec-decode attention metadata (#44492)

2026-07-20 20:50:58 +00:00

test_eagle_step_kernel.py

[XPU] Fix spec-decode UTs under tests/v1/spec_decode (#38491)

2026-04-11 01:31:00 +00:00

test_eagle.py

[Bugfix][Spec Decode] Fix probabilistic sampling for parallel drafting (#45956)

2026-06-24 05:36:23 +00:00

test_extract_hidden_states.py

[V1][Spec Decode] Add Dynamic SD (#32374)

2026-06-14 00:14:27 -07:00

test_llm_base_proposer_sampling.py

[New model] Kimi K3 (#50000)

2026-07-30 03:49:58 -07:00

test_llm_base_proposer.py

[BugFix] eagle draft max position embeddings (#49343)

2026-07-29 01:24:47 -07:00

test_max_len.py

[CI] Explicitly tear down speculative decode runners (#49910)

2026-07-27 05:32:46 +00:00

test_mtp_structured_output.py

[Bugfix][Structured Output][Spec Decode] Constrain bitmask and trim grammar advance at the reasoning boundary (#44297)

2026-07-04 09:08:45 +00:00

test_mtp.py

[V1][Spec Decode] Add Dynamic SD (#32374)

2026-06-14 00:14:27 -07:00

test_ngram.py

[V1][Spec Decode] Add Dynamic SD (#32374)

2026-06-14 00:14:27 -07:00

test_rejection_sampler_utils.py

[Model Runner V2] Cache draft logits in model's LM head dtype (#50910)

2026-08-05 19:12:46 -07:00

test_speculators_correctness.py

[Spec Decode] Add Qwen3 architecture support for EAGLE3 (#43132)

2026-06-21 13:55:26 -07:00

test_speculators_eagle3.py

[Rocm][CI] Fix test_speculator_eagle3 by skipping the CompressedTensorw4a16 Model (#30001)

2025-12-04 07:52:28 +00:00

test_synthetic_rejection_sampler_utils.py

[Feat] Unified Synthetic Acceptance Rate for V1 and V2 (#40662)

2026-04-24 00:48:08 +00:00

test_vocab_mapping.py

[Feature] Universal speculative decoding for heterogeneous vocabularies (TLI) (#38174)

2026-07-02 01:34:20 -07:00