xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-06-06 09:15:42 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Fix packed_factor missing attribute error (#23902)

Browse Source 
Signed-off-by: Kyuyeun Kim <kyuyeunk@google.com>
This commit is contained in:
Kyuyeun Kim 2025-09-02 10:56:31 -07:00 committed by GitHub
parent f399182e8c
commit 9480ae24e3
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 8 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
vllm/model_executor/layers/linear.py
View File

@ -722,8 +722,8 @@ class MergedColumnParallelLinear(ColumnParallelLinear):
# If quantized, we need to adjust the offset and size to account # If quantized, we need to adjust the offset and size to account
# for the packing. # for the packing.
if packed_dim == output_dim: if packed_dim == output_dim:
shard_size = shard_size // param.pack_factor shard_size = shard_size // param.packed_factor
shard_offset = shard_offset // param.pack_factor shard_offset = shard_offset // param.packed_factor
# Special case for Marlin. # Special case for Marlin.
shard_size, shard_offset = adjust_marlin_shard( shard_size, shard_offset = adjust_marlin_shard(
param, shard_size, shard_offset) param, shard_size, shard_offset)
@ -756,8 +756,8 @@ class MergedColumnParallelLinear(ColumnParallelLinear):
# for the packing. # for the packing.
packed_dim = getattr(param, "packed_dim", None) packed_dim = getattr(param, "packed_dim", None)
if packed_dim == output_dim: if packed_dim == output_dim:
shard_size = shard_size // param.pack_factor shard_size = shard_size // param.packed_factor
shard_offset = shard_offset // param.pack_factor shard_offset = shard_offset // param.packed_factor
# Special case for Marlin. # Special case for Marlin.
shard_size, shard_offset = adjust_marlin_shard( shard_size, shard_offset = adjust_marlin_shard(
param, shard_size, shard_offset) param, shard_size, shard_offset)
@ -1107,8 +1107,8 @@ class QKVParallelLinear(ColumnParallelLinear):
# If quantized, we need to adjust the offset and size to account # If quantized, we need to adjust the offset and size to account
# for the packing. # for the packing.
if packed_dim == output_dim: if packed_dim == output_dim:
shard_size = shard_size // param.pack_factor shard_size = shard_size // param.packed_factor
shard_offset = shard_offset // param.pack_factor shard_offset = shard_offset // param.packed_factor
# Special case for Marlin. # Special case for Marlin.
shard_size, shard_offset = adjust_marlin_shard( shard_size, shard_offset = adjust_marlin_shard(
@ -1155,8 +1155,8 @@ class QKVParallelLinear(ColumnParallelLinear):
# for the packing. # for the packing.
packed_dim = getattr(param, "packed_dim", None) packed_dim = getattr(param, "packed_dim", None)
if packed_dim == output_dim: if packed_dim == output_dim:
shard_size = shard_size // param.pack_factor shard_size = shard_size // param.packed_factor
shard_offset = shard_offset // param.pack_factor shard_offset = shard_offset // param.packed_factor
# Special case for Marlin. # Special case for Marlin.
shard_size, shard_offset = adjust_marlin_shard( shard_size, shard_offset = adjust_marlin_shard(