[Model Runner v2] Fix block table IMA issue (#40648)
Signed-off-by: yewentao256 <zhyanwentao@126.com> Signed-off-by: Nick Hill <nickhill123@gmail.com> Co-authored-by: Nick Hill <nickhill123@gmail.com>
This commit is contained in:
@@ -558,6 +558,9 @@ class GPUModelRunner(LoRAModelRunnerMixin):
|
||||
del hidden_states, sample_hidden_states
|
||||
gc.collect()
|
||||
|
||||
def post_kv_cache_wake_up(self) -> None:
|
||||
self.block_tables.init_block_table_layout_tensors()
|
||||
|
||||
def reset_mm_cache(self) -> None:
|
||||
if self.encoder_cache is not None:
|
||||
self.encoder_cache.reset_mm_cache()
|
||||
|
||||
Reference in New Issue
Block a user