A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 9, 00:00 UTC
Fix Gemma3 GGUF for Transformers v5 (#33683)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Harry Mellor committed
Feb 3, 2026 at 12:36 UTC
be8168ff889aa8981d4e8a158fc1b4d0a4deb18b
1 file changed
+1
-1
vllm/transformers_utils/gguf_utils.py
+1
-1
@@ -250,7 +250,7 @@ def maybe_patch_hf_config_from_gguf(
250
text_config = hf_config.get_text_config()
251
is_gemma3 = hf_config.model_type in ("gemma3", "gemma3_text")
252
if vision_config is not None and is_gemma3:
253
- new_hf_config = Gemma3Config.from_text_vision_configs(
253
+ new_hf_config = Gemma3Config(
254
text_config=text_config,
255
vision_config=vision_config,
256
architectures=["Gemma3ForConditionalGeneration"],