Add pydocstyle to the ruff rules (#52136)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com> Signed-off-by: Harry Mellor <hej.mellor@gmail.com> Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Opus 5
parent
804e5377aa
commit
8c1557a79c
@@ -445,8 +445,7 @@ class GteNewModelConfig(VerifyAndUpdateConfig):
|
||||
class HybridAttentionMambaModelConfig(VerifyAndUpdateConfig):
|
||||
@classmethod
|
||||
def verify_and_update_config(cls, vllm_config: "VllmConfig") -> None:
|
||||
"""
|
||||
Perform early validation and setup for hybrid attention/mamba models.
|
||||
"""Perform early validation and setup for hybrid attention/mamba models.
|
||||
|
||||
Block size alignment with mamba page sizes is handled later by
|
||||
Platform.update_block_size_for_backend(), which runs after model
|
||||
@@ -454,6 +453,7 @@ class HybridAttentionMambaModelConfig(VerifyAndUpdateConfig):
|
||||
|
||||
Args:
|
||||
vllm_config: vLLM Config
|
||||
|
||||
"""
|
||||
# Enable FULL_AND_PIECEWISE by default
|
||||
MambaModelConfig.verify_and_update_config(vllm_config)
|
||||
@@ -609,12 +609,12 @@ class LlamaNemotronVLConfig(VerifyAndUpdateConfig):
|
||||
class MambaModelConfig(VerifyAndUpdateConfig):
|
||||
@classmethod
|
||||
def verify_and_update_config(cls, vllm_config: "VllmConfig") -> None:
|
||||
"""
|
||||
Enable FULL_AND_PIECEWISE cuda graph mode by default (required
|
||||
"""Enable FULL_AND_PIECEWISE cuda graph mode by default (required
|
||||
to get good performance for mamba layers in V1).
|
||||
|
||||
Args:
|
||||
vllm_config: vLLM Config
|
||||
|
||||
"""
|
||||
model_config = vllm_config.model_config
|
||||
cache_config = vllm_config.cache_config
|
||||
|
||||
Reference in New Issue
Block a user