EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Update model_loader deps and qqq quantization deps (#2220) (#2318)

Browse Source 
Co-authored-by: HandH1998 <1335248067@qq.com>
This commit is contained in:
Yineng Zhang
2024-12-02 23:22:13 +08:00
committed by GitHub
parent 33deca81b5
commit 85e1a6f3aa
58 changed files with 2363 additions and 366 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
python/sglang/srt/models/llava.py
View File

@@ -29,7 +29,6 @@ from transformers import (
SiglipVisionModel,
)
from transformers.models.llava.modeling_llava import LlavaMultiModalProjector
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from sglang.srt.layers.quantization.base_config import QuantizationConfig
from sglang.srt.managers.schedule_batch import ImageInputs
@@ -39,6 +38,7 @@ from sglang.srt.mm_utils import (
unpad_image_shape,
)
from sglang.srt.model_executor.forward_batch_info import ForwardBatch
from sglang.srt.model_loader.weight_utils import default_weight_loader
from sglang.srt.models.llama import LlamaForCausalLM
from sglang.srt.models.mistral import MistralForCausalLM
from sglang.srt.models.qwen2 import Qwen2ForCausalLM
@@ -451,7 +451,6 @@ class LlavaLlamaForCausalLM(LlavaBaseForCausalLM):
self,
config: LlavaConfig,
quant_config: Optional[QuantizationConfig] = None,
cache_config=None,
) -> None:
super().__init__()
@@ -473,7 +472,6 @@ class LlavaQwenForCausalLM(LlavaBaseForCausalLM):
self,
config: LlavaConfig,
quant_config: Optional[QuantizationConfig] = None,
cache_config=None,
) -> None:
super().__init__()
@@ -506,7 +504,6 @@ class LlavaMistralForCausalLM(LlavaBaseForCausalLM):
self,
config: LlavaConfig,
quant_config: Optional[QuantizationConfig] = None,
cache_config=None,
) -> None:
super().__init__()