EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[Lint]Style: Convert vllm-ascend/ to ruff format(Batch #7) (#6023)

Browse Source 
### What this PR does / why we need it?
**Scope of Changes**:
| File Path |
| :--- |
|` vllm_ascend/quantization/compressed_tensors/compressed_tensors.py`|
|` vllm_ascend/quantization/quant_config.py`|
|` vllm_ascend/quantization/utils.py`|
|` vllm_ascend/quantization/w4a16.py`|
|` vllm_ascend/quantization/w4a4_flatquant_dynamic.py`|
|` vllm_ascend/quantization/w4a8_dynamic.py`|
|` vllm_ascend/quantization/w8a16.py`|
|` vllm_ascend/quantization/w8a8.py`|
|` vllm_ascend/quantization/w8a8_dynamic.py`|
|` vllm_ascend/quantization/w8a8_pdmix.py`|
|` vllm_ascend/quantization/w8a8mxfp8.py`|
|` vllm_ascend/sample/rejection_sampler.py`|
|` vllm_ascend/sample/sampler.py`|
|` vllm_ascend/worker/block_table.py`|

### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: v0.13.0
- vLLM main:
2c24bc6996

Signed-off-by: MrZ20 <2609716663@qq.com>
This commit is contained in:
SILONG ZENG
2026-02-06 14:56:53 +08:00
committed by GitHub
parent d0bc16859c
commit 99aedaff63
20 changed files with 997 additions and 1307 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
vllm_ascend/quantization/methods/w8a16.py
View File

@@ -15,7 +15,7 @@
# limitations under the License.
#
from typing import Any, Dict, Optional
from typing import Any
import torch
import torch_npu
@@ -29,7 +29,7 @@ from .registry import register_scheme
@register_scheme("W8A16", "linear")
class AscendW8A16LinearMethod(AscendLinearScheme):
"""Linear method for Ascend W8A16.
This scheme uses 8-bit quantized weights with 16-bit activations.
"""
@@ -41,39 +41,34 @@ class AscendW8A16LinearMethod(AscendLinearScheme):
input_size: int,
output_size: int,
params_dtype: torch.dtype = torch.bfloat16,
) -> Dict[str, Any]:
params_dict = {
"weight": torch.empty(output_size, input_size, dtype=torch.int8)
}
) -> dict[str, Any]:
params_dict = {"weight": torch.empty(output_size, input_size, dtype=torch.int8)}
return params_dict
def get_perchannel_param(
self,
output_size: int,
params_dtype: torch.dtype,
) -> Dict[str, Any]:
) -> dict[str, Any]:
params_dict = {}
params_dict["weight_scale"] = torch.empty(output_size,
1,
dtype=params_dtype)
params_dict["weight_offset"] = torch.empty(output_size,
1,
dtype=params_dtype)
params_dict["weight_scale"] = torch.empty(output_size, 1, dtype=params_dtype)
params_dict["weight_offset"] = torch.empty(output_size, 1, dtype=params_dtype)
return params_dict
def apply(
self,
layer: torch.nn.Module,
x: torch.Tensor,
bias: Optional[torch.Tensor] = None,
tp_rank: Optional[int] = 0,
bias: torch.Tensor | None = None,
tp_rank: int | None = 0,
) -> torch.Tensor:
output = torch_npu.npu_weight_quant_batchmatmul(
x=x,
weight=layer.weight,
antiquant_scale=layer.weight_scale,
antiquant_offset=layer.weight_offset,
bias=bias)
bias=bias,
)
return output
def process_weights_after_loading(self, layer):