[Model] Add VaultGemma via Transformers modeling backend (#49803)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2026-07-25 16:54:15 +00:00
committed by GitHub
parent 7fe6d3c76b
commit 26d725c334
6 changed files with 8 additions and 17 deletions
+3 -5
View File
@@ -102,10 +102,7 @@ from vllm.logger import init_logger, suppress_logging
from vllm.platforms import CpuArchEnum, current_platform
from vllm.plugins import load_general_plugins
from vllm.ray.lazy_utils import is_in_ray_actor, is_ray_initialized
from vllm.transformers_utils.config import (
is_interleaved,
maybe_override_with_speculators,
)
from vllm.transformers_utils.config import maybe_override_with_speculators
from vllm.transformers_utils.repo_utils import get_model_path
from vllm.transformers_utils.utils import is_cloud_storage
from vllm.utils.argparse_utils import (
@@ -1908,7 +1905,8 @@ class EngineArgs:
self._set_default_chunked_prefill_and_prefix_caching_args(model_config)
self._set_default_reasoning_config_args()
sliding_window: int | None = None
if not is_interleaved(model_config.hf_text_config):
layer_types = getattr(model_config.hf_text_config, "layer_types", None)
if layer_types is None or all(lt == "sliding_attention" for lt in layer_types):
# Only set CacheConfig.sliding_window if the model is all sliding
# window. Otherwise CacheConfig.sliding_window will override the
# global layers in interleaved sliding window models.