Refactor: move all quantization-related code to srt/layer/quantization (#7989)

This commit is contained in:
Cheng Wan
2025-07-17 00:47:07 -07:00
committed by GitHub
parent 02404a1e35
commit 49b8777460
22 changed files with 1095 additions and 1175 deletions
+1 -5
View File
@@ -19,15 +19,11 @@ import torch
import torch.nn.functional as F
from sglang.srt.eplb import expert_location_dispatch
from sglang.srt.eplb.expert_distribution import (
ExpertDistributionRecorder,
get_global_expert_distribution_recorder,
)
from sglang.srt.eplb.expert_distribution import get_global_expert_distribution_recorder
from sglang.srt.eplb.expert_location_dispatch import (
ExpertLocationDispatchInfo,
topk_ids_logical_to_physical,
)
from sglang.srt.managers.schedule_batch import global_server_args_dict
from sglang.srt.utils import (
cpu_has_amx_support,
get_bool_env_var,