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

[1/N] Remove CacheConfig import in all model files (#1658)

Browse Source
This commit is contained in:
Byron Hsu
2024-10-14 09:06:34 -07:00
committed by GitHub
parent 02bc95796d
commit 56503d9bc9
30 changed files with 64 additions and 91 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
python/sglang/srt/models/minicpm3.py
View File

@@ -21,7 +21,6 @@ from typing import Any, Dict, Iterable, Optional, Tuple
import torch
from torch import nn
from transformers import PretrainedConfig
from vllm.config import CacheConfig
from vllm.distributed import get_tensor_model_parallel_world_size
from vllm.model_executor.layers.linear import (
ColumnParallelLinear,
@@ -108,7 +107,7 @@ class MiniCPM3Attention(nn.Module):
rope_theta: float = 10000,
rope_scaling: Optional[Dict[str, Any]] = None,
max_position_embeddings: int = 8192,
cache_config: Optional[CacheConfig] = None,
cache_config=None,
quant_config: Optional[QuantizationConfig] = None,
layer_id=None,
) -> None:
@@ -252,7 +251,7 @@ class MiniCPM3AttentionMLA(nn.Module):
rope_theta: float = 10000,
rope_scaling: Optional[Dict[str, Any]] = None,
max_position_embeddings: int = 8192,
cache_config: Optional[CacheConfig] = None,
cache_config=None,
quant_config: Optional[QuantizationConfig] = None,
layer_id=None,
) -> None:
@@ -409,7 +408,7 @@ class MiniCPM3DecoderLayer(nn.Module):
self,
config: PretrainedConfig,
layer_id: int,
cache_config: Optional[CacheConfig] = None,
cache_config=None,
quant_config: Optional[QuantizationConfig] = None,
) -> None:
super().__init__()
@@ -501,7 +500,7 @@ class MiniCPM3Model(nn.Module):
def __init__(
self,
config: PretrainedConfig,
cache_config: Optional[CacheConfig] = None,
cache_config=None,
quant_config: Optional[QuantizationConfig] = None,
) -> None:
super().__init__()
@@ -552,7 +551,7 @@ class MiniCPM3ForCausalLM(nn.Module):
def __init__(
self,
config: PretrainedConfig,
cache_config: Optional[CacheConfig] = None,
cache_config=None,
quant_config: Optional[QuantizationConfig] = None,
) -> None:
super().__init__()