[Bugfix] Pass drafter quant_config to ParallelLMHead in Eagle3 (#37280)

Signed-off-by: Matthias Gehre <matthias.gehre@amd.com>
This commit is contained in:
Matthias Gehre
2026-03-25 11:42:58 +00:00
committed by GitHub
parent ba2910f73a
commit a889b7f584
3 changed files with 48 additions and 0 deletions
@@ -294,6 +294,7 @@ class Eagle3LlamaForCausalLM(LlamaForCausalLM):
self.lm_head = ParallelLMHead(
self.config.draft_vocab_size,
self.config.hidden_size,
quant_config=get_draft_quant_config(vllm_config),
prefix=maybe_prefix(prefix, "lm_head"),
)
self.logits_processor = LogitsProcessor(