freeliuzc e87ce4b8cd [Speculative Decoding] refactor MTP and optimize spec-decoding postprocess (#6973) ...

* support new mtp

* refactor(speculate_decoding and mtp): optimize mtp sturcture logic. Update spec-branch status-process

* fix cuda-graph for spec-decoding

* fix xpu mtp and fix some note

* fix unittest and optmize note

* fix model status update in eos-branch

2026-03-24 10:19:01 +08:00

..

draft_model_postprocess.cu

init (#6642)

2026-03-04 21:55:31 +08:00

draft_model_preprocess.cu

[Speculative Decoding] refactor MTP and optimize spec-decoding postprocess (#6973)

2026-03-24 10:19:01 +08:00

draft_model_set_value_by_flags.cu

init (#6642)

2026-03-04 21:55:31 +08:00

draft_model_update.cu

[Speculative Decoding] refactor MTP and optimize spec-decoding postprocess (#6973)

2026-03-24 10:19:01 +08:00

eagle_get_hidden_states.cu

init (#6642)

2026-03-04 21:55:31 +08:00

eagle_get_self_hidden_states.cu

init (#6642)

2026-03-04 21:55:31 +08:00

hydra_fetch_hidden_states.cu

init (#6642)

2026-03-04 21:55:31 +08:00

mtp_save_first_token.cc

[BugFix] replace ftok with custom_ftok in get_output/save_output ops (#6822)

2026-03-16 14:22:18 +08:00

mtp_step_paddle.cu

cuda13.0, implement changes to CCCL (#6751)

2026-03-10 16:47:02 +08:00

ngram_match_mixed.cu

init (#6642)

2026-03-04 21:55:31 +08:00