[SpecDecode] Add LiLiCorr drafter (#57934)

Signed-off-by: Andrii Skliar <askliar@nvidia.com>
Signed-off-by: Andrii Skliar <andreyws96@gmail.com>
Co-authored-by: Andrii Skliar <askliar@nvidia.com>
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Benjamin Chislett <chislett.ben@gmail.com>
This commit is contained in:
Andrii Skliar
2026-09-25 10:08:28 -04:00
committed by GitHub
co-authored by Andrii Skliar Codex Benjamin Chislett
parent 25b0add7b8
commit 73a78e6f1f
21 changed files with 1944 additions and 110 deletions
+1
View File
@@ -633,6 +633,7 @@ _SPECULATIVE_DECODING_MODELS = {
"EagleLlama4ForCausalLM": ("llama4_eagle", "EagleLlama4ForCausalLM"),
"EagleMiniCPMForCausalLM": ("minicpm_eagle", "EagleMiniCPMForCausalLM"),
"DFlashDraftModel": ("qwen3_dflash", "DFlashQwen3ForCausalLM"),
"LiLiCorrDraftModel": ("lilicorr", "LiLiCorrForCausalLM"),
"DFlash2DraftModel": ("qwen3_dflash2", "DFlash2Qwen3ForCausalLM"),
# Muse Glimmer's DFlash draft head, reusing the generic qwen3_dflash
# implementation. EAGLEConfig rewrites a dflash draft's architecture to