chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

Fix kernel type annotations for fp8 quant and logging (#16994)

Browse Source
This commit is contained in:
Lianmin Zheng
2026-01-13 18:14:32 -08:00
committed by GitHub
parent afe285f7bd
commit a4825ed588
7 changed files with 797 additions and 684 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
python/sglang/srt/layers/quantization/base_config.py
View File

@@ -230,7 +230,9 @@ class QuantizationConfig(ABC):
"""
raise NotImplementedError()
def apply_sglang_mapper(self, hf_to_sglang_mapper: "WeightsMapper"): # noqa: B027
def apply_weight_name_mapper(
self, hf_to_sglang_mapper: "WeightsMapper"
): # noqa: B027
"""
Interface for models to update module names referenced in
quantization configs in order to reflect the sglang model structure

2
python/sglang/srt/layers/quantization/compressed_tensors/compressed_tensors.py
View File

@@ -123,7 +123,7 @@ class CompressedTensorsConfig(QuantizationConfig):
def get_scaled_act_names(self) -> List[str]:
return []
def apply_sglang_mapper(self, hf_to_sglang_mapper: "WeightsMapper"):
def apply_weight_name_mapper(self, hf_to_sglang_mapper: "WeightsMapper"):
self.target_scheme_map = hf_to_sglang_mapper.apply_dict(self.target_scheme_map)
self.ignore = hf_to_sglang_mapper.apply_list(self.ignore)
self.sparsity_scheme_map = hf_to_sglang_mapper.apply_dict(