[Bugfix] DiffusionGemma: fix multimodal support (#57589)

Signed-off-by: Matt Mastracci <matthew@mastracci.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
This commit is contained in:
Matt Mastracci
2026-09-21 00:05:24 -04:00
committed by GitHub
co-authored by Claude Fable 5.1 Lucas Wilkinson mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
parent 9a70c233cd
commit ff0c885494
@@ -197,6 +197,11 @@ class DiffusionGemmaForConditionalGeneration(
text_config.attention_k_eq_v = True
# ---- Vision tower ----
# Gemma4's image path, borrowed below, reads this flag.
lora_config = vllm_config.lora_config
self._enable_mm_lora = bool(
lora_config is not None and lora_config.enable_tower_connector_lora
)
vision_config = getattr(config, "vision_config", None)
self.embed_vision: Gemma4MultimodalEmbedder | None
if vision_config is not None: