[Feat][Spec Decode] DFlash (#36847)

Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
This commit is contained in:
Benjamin Chislett
2026-03-30 15:03:15 -04:00
committed by GitHub
parent ab1a6a43fa
commit 494636b29d
17 changed files with 1577 additions and 107 deletions

View File

@@ -546,6 +546,7 @@ _SPECULATIVE_DECODING_MODELS = {
"EagleLlamaForCausalLM": ("llama_eagle", "EagleLlamaForCausalLM"),
"EagleLlama4ForCausalLM": ("llama4_eagle", "EagleLlama4ForCausalLM"),
"EagleMiniCPMForCausalLM": ("minicpm_eagle", "EagleMiniCPMForCausalLM"),
"DFlashDraftModel": ("qwen3_dflash", "DFlashQwen3ForCausalLM"),
"Eagle3LlamaForCausalLM": ("llama_eagle3", "Eagle3LlamaForCausalLM"),
"LlamaForCausalLMEagle3": ("llama_eagle3", "Eagle3LlamaForCausalLM"),
"Eagle3Qwen2_5vlForCausalLM": ("llama_eagle3", "Eagle3LlamaForCausalLM"),