squall/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

[Bugfix] Fix PerTensorScaleParameter weight loading for fused models (#7376)

Browse Source
This commit is contained in:
Dipika Sikka 2024-08-09 17:23:46 -04:00 committed by GitHub
parent 933790c209
commit 5c6c54d67a
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 14 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
vllm/model_executor/layers/linear.py
View File

@ -14,7 +14,8 @@ from vllm.logger import init_logger
from vllm.model_executor.layers.quantization.base_config import ( from vllm.model_executor.layers.quantization.base_config import (
QuantizationConfig, QuantizeMethodBase) QuantizationConfig, QuantizeMethodBase)
from vllm.model_executor.parameter import (BasevLLMParameter, from vllm.model_executor.parameter import (BasevLLMParameter,
PackedvLLMParameter) PackedvLLMParameter,
PerTensorScaleParameter)
from vllm.model_executor.utils import set_weight_attrs from vllm.model_executor.utils import set_weight_attrs
logger = init_logger(__name__) logger = init_logger(__name__)
@ -573,11 +574,13 @@ class MergedColumnParallelLinear(ColumnParallelLinear):
param: BasevLLMParameter, param: BasevLLMParameter,
loaded_weight: torch.Tensor, loaded_weight: torch.Tensor,
loaded_shard_id: Optional[int] = None): loaded_shard_id: Optional[int] = None):
param_data = param.data
if loaded_shard_id is None: if loaded_shard_id is None:
if param.output_dim is None: if isinstance(param, PerTensorScaleParameter):
assert param_data.shape == loaded_weight.shape param.load_merged_column_weight(loaded_weight=loaded_weight,
param_data.copy_(loaded_weight) shard_id=0)
return
elif type(param) is BasevLLMParameter:
param.load_merged_column_weight(loaded_weight=loaded_weight)
return return
self._load_fused_module_from_checkpoint(param, loaded_weight) self._load_fused_module_from_checkpoint(param, loaded_weight)
return return
@ -720,11 +723,13 @@ class QKVParallelLinear(ColumnParallelLinear):
param: BasevLLMParameter, param: BasevLLMParameter,
loaded_weight: torch.Tensor, loaded_weight: torch.Tensor,
loaded_shard_id: Optional[str] = None): loaded_shard_id: Optional[str] = None):
param_data = param.data
if loaded_shard_id is None: # special case for certain models if loaded_shard_id is None: # special case for certain models
if param.output_dim is None: if isinstance(param, PerTensorScaleParameter):
assert param_data.shape == loaded_weight.shape param.load_merged_column_weight(loaded_weight=loaded_weight,
param_data.copy_(loaded_weight) shard_id=0)
return
elif type(param) is BasevLLMParameter:
param.load_merged_column_weight(loaded_weight=loaded_weight)
return return
self._load_fused_module_from_checkpoint(param, loaded_weight) self._load_fused_module_from_checkpoint(param, loaded_weight)
return return