Skip to content

GitLab

Unverified Commit 3ed7b1e6 authored by Andreas Karatzas's avatar Andreas Karatzas Committed by GitHub
Browse files

[ROCm] Validate block_size for explicitly selected attention backends (#36846) 


Signed-off-by: default avatarAndreas Karatzas <akaratza@amd.com>
parent e8f9dbc3
Show whitespace changes
Inline Side-by-side
Showing with 0 additions and 2 deletions
vllm/platforms/rocm.py
View file @ 3ed7b1e6
...@@ -438,8 +438,6 @@ class RocmPlatform(Platform): ...@@ -438,8 +438,6 @@ class RocmPlatform(Platform):
device_capability = cls.get_device_capability() device_capability = cls.get_device_capability()
assert device_capability is not None assert device_capability is not None
attn_selector_config = attn_selector_config._replace(block_size=None)
# First try checking just the selected backend, if there is one. # First try checking just the selected backend, if there is one.
if selected_backend is not None: if selected_backend is not None:
try: try:
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment