Skip to content

[Fix] Fix eplb bug and support fp8 load weight #5178

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Jiang-Jia-Jun merged 3 commits into from
Nov 24, 2025
Merged

[Fix] Fix eplb bug and support fp8 load weight #5178

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
386652f
fix eplb part2
xiaoxiaohehe001 Nov 23, 2025
3f652e2
fix eplb part2
xiaoxiaohehe001 Nov 23, 2025
e335f14
fix eplb part2
xiaoxiaohehe001 Nov 24, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 4 additions & 3 deletions fastdeploy/eplb/async_expert_loader.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -152,6 +152,9 @@ def load_tensor_from_shm_mem(tensor_infos, shm_ptr, logger=None):
# NumPy 不支持 bfloat16，因此先以 uint16 读取原始数据，再用 Paddle cast 为 bfloat16
tmp = np_array.view(np.uint16)
tensor = paddle.Tensor(tmp, dtype=paddle.bfloat16, place=paddle.CPUPlace(), zero_copy=True)
elif dtype == paddle.float8_e4m3fn:
tmp = np_array.view(np.uint8)
tensor = paddle.Tensor(tmp, dtype=paddle.float8_e4m3fn, place=paddle.CPUPlace(), zero_copy=True)
else:
raise TypeError(f"Unsupported dtype: {dtype}")

Expand Down Expand Up @@ -294,8 +297,6 @@ def load_safetensor_fp8_from_disk(self, need_to_reload: List[Tuple[int, int]]):
"""
up_gate_down = ["up_gate_proj", "down_proj"]
quant_weight_scale = ["quant_weight", "weight_scale"]
if self.moe_quant_type == "w4a8":
quant_weight_scale = ["quant_weight"]
ckpt_name = [
(f"ernie.layers.{layer_id}.mlp.experts.{expert_id}.{proj_name}.{quant_name}")
for layer_id, expert_id in need_to_reload
Expand All @@ -312,7 +313,7 @@ def load_safetensor_fp8_from_disk(self, need_to_reload: List[Tuple[int, int]]):
from safetensors import safe_open

for st_file in hf_weights_files:
with safe_open(st_file, framework="np", device="cpu") as f:
with safe_open(st_file, framework="paddle", device="cpu") as f:
for name in f.keys():
if name in ckpt_name:
weight = f.get_tensor(name)
Expand Down
12 changes: 7 additions & 5 deletions fastdeploy/eplb/experts_manager.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@ class RedundantExpertManager:
RedundantExpertManger
"""

def __init__(self, rank=0, ep_size=32, fd_config=None):
def __init__(self, rank=0, ep_size=64, fd_config=None):
self.logger = get_logger("eplb_expert_manager", "eplb_{0}.log".format(rank))

self.rank = rank
Expand Down Expand Up @@ -101,7 +101,7 @@ def __init__(self, rank=0, ep_size=32, fd_config=None):
self.http_timeout = 1
# 重置重排状态: 'done' -> 'free'
self.rearrange_end_ts = 0
self.rearrange_reset_interval = 300
self.rearrange_reset_interval = 30

self.tensor_infos = None

Expand Down Expand Up @@ -250,8 +250,8 @@ def caculate_expert_rank_table(self, is_init=False):
eplb_strategy = self.eplb_config.redundant_expert_eplb_strategy
if is_init:
num_groups = 1
num_nodes = 2
num_gpus = 2 * 8
num_nodes = 8
num_gpus = 8 * 8
eplb_strategy = ""
# eplb
rank_expert_list, logical_to_physical_map, expert_count = rebalance_experts(
Expand Down Expand Up @@ -420,7 +420,9 @@ def allreduce_load_weight_result(self):
if not exist_fail and all_success:
# prefill需要等待调度屏蔽
if (
self.fd_config.splitwise_role == "decode"
self.fd_config.scheduler_config.splitwise_role == "mixed"
or self.fd_config.scheduler_config.splitwise_role == "decode"
or self.fd_config.scheduler_config.splitwise_role == "prefill"
or not self.eplb_config.redundant_expert_enable_schedule_cordon
):
self.logger.info("redundant_expert: allreduce_load_weight_result success, notify infer.py")
Expand Down
4 changes: 4 additions & 0 deletions fastdeploy/model_executor/layers/moe/fused_moe_cutlass_backend.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1153,6 +1153,10 @@ def process_prequanted_weights(self, layer: nn.Layer, state_dict, is_rearrange:
# self.check(layer, up_gate_proj_weights, down_proj_weights)
up_gate_proj_weight_scale = []
down_proj_weight_scale = []

if isinstance(state_dict, list):
state_dict = dict(state_dict)

for expert_idx in logical_expert_ids:
up_gate_proj_weight_scale.append(
get_tensor(state_dict.pop(up_gate_proj_expert_weight_scale_key.format(expert_idx)))
Expand Down
4 changes: 4 additions & 0 deletions fastdeploy/model_executor/layers/moe/fused_moe_deepgemm_backend.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,6 +86,10 @@ def process_prequanted_weights(self, layer: nn.Layer, state_dict, is_rearrange:
# self.check(layer, up_gate_proj_weights, down_proj_weights)
up_gate_proj_weight_scale = []
down_proj_weight_scale = []

if isinstance(state_dict, list):
state_dict = dict(state_dict)

for expert_idx in logical_expert_ids:
up_gate_proj_expert_weight_scale_key_name = up_gate_proj_expert_weight_scale_key.format(expert_idx)
down_proj_expert_weight_scale_key_name = down_proj_expert_weight_scale_key.format(expert_idx)
Expand Down
8 changes: 8 additions & 0 deletions fastdeploy/worker/worker_process.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -260,13 +260,21 @@ def update_weights_from_tensor(self, mmap_infos):
"""
update_weights_from_tensor
"""
import time

while True:
if self.experts_manager.tensor_infos is None:
time.sleep(0.1)
else:
break
state_dicts = load_tensor_from_shm_mem(self.experts_manager.tensor_infos, mmap_infos[MODEL_MAIN_NAME], logger)
rank_expert_list, logical_to_physical_map, expert_count = self.experts_manager.get_ep_rank_to_expert_id_list()
self.worker.get_model().redundant_table_manger.update_expert_rank_table(
rank_expert_list, logical_to_physical_map, expert_count
)
# TO BE FIXED
self.worker.get_model().update_state_dict(state_dicts)
self.experts_manager.tensor_infos = None

def _broadcast_model_weights_signal(self, src: int, group) -> int:
model_weights_signal_tensor = paddle.full(shape=[1], fill_value=self.model_weights_signal[0], dtype="int32")
Expand Down
Loading