biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Docs] Fix warnings in mkdocs build (continued) (#24740)

Browse Source 
Signed-off-by: Zerohertz <ohg3417@gmail.com>
This commit is contained in:
Hyogeun Oh (오효근)
2025-09-12 22:43:15 +09:00
committed by GitHub
parent bcb06d7baf
commit 41f17bf290
10 changed files with 121 additions and 176 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
vllm/model_executor/layers/quantization/torchao.py
View File

@@ -144,9 +144,9 @@ def torchao_quantize_param_data(param: torch.Tensor,
"""Quantize a Tensor with torchao quantization specified by torchao_config
Args:
`param`: weight parameter of the linear module
`torchao_config`: type of quantization and their arguments we want to
use to quantize the Tensor
param: weight parameter of the linear module
torchao_config: type of quantization and their arguments we want to
use to quantize the Tensor
"""
from torchao.core.config import AOBaseConfig
from torchao.quantization import quantize_
@@ -172,8 +172,8 @@ class TorchAOLinearMethod(LinearMethodBase):
"""Linear method for torchao.
Args:
torchao_config: The torchao quantization config, a string
that encodes the type of quantization and all relevant arguments.
quant_config: The torchao quantization config, a string that encodes
the type of quantization and all relevant arguments.
"""
def __init__(self, quant_config: TorchAOConfig):

2
vllm/model_executor/layers/quantization/utils/int8_utils.py
View File

@@ -423,7 +423,7 @@ def w8a8_block_int8_matmul(
Bs: The per-block quantization scale for `B`.
block_size: The block size for per-block quantization. It should be
2-dim, e.g., [128, 128].
output_dytpe: The dtype of the returned tensor.
output_dtype: The dtype of the returned tensor.
Returns:
torch.Tensor: The result of matmul.