[Core] Simplify core kv-cache blocks initialization logic (#36521)
Signed-off-by: Nick Hill <nickhill123@gmail.com>
This commit is contained in:
@@ -104,10 +104,6 @@ class WorkerBase:
|
||||
"""
|
||||
raise NotImplementedError
|
||||
|
||||
def initialize_cache(self, num_gpu_blocks: int, num_cpu_blocks: int) -> None:
|
||||
"""Initialize the KV cache with the given size in blocks."""
|
||||
raise NotImplementedError
|
||||
|
||||
def reset_mm_cache(self) -> None:
|
||||
reset_fn = getattr(self.model_runner, "reset_mm_cache", None)
|
||||
if callable(reset_fn):
|
||||
|
||||
Reference in New Issue
Block a user