Skip to content

Closed by author #9376

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

Closed
yyswhsccc wants to merge 1 commit into from
+3 −1
Closed

Closed by author #9376

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
4 changes: 3 additions & 1 deletion swift/rlhf_trainers/rollout_mixin.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,7 @@
from swift.template import Template
from swift.tuners import Swift
from swift.utils import (get_current_device, get_logger, is_deepspeed_enabled, is_vllm_available, remove_response,
to_device)
synchronize, to_device)
from .arguments import RolloutTrainerArgumentsMixin
from .rlhf_mixin import RLHFTrainerMixin
from .utils import (VLLM_LORA_INT_ID, VLLM_LORA_NAME, VLLM_LORA_PATH, FlattenedTensorBucket, TensorLoRARequest,
Expand Down Expand Up @@ -570,6 +570,8 @@ def _load_state_dict_to_vllm(self, state_dict):
# Patch MoE weight_loader if needed
patch_vllm_moe_model_weight_loader(llm_model)
llm_model.load_weights(state_dict.items())
# Keep ZeRO-3 gathered tensors valid until vLLM finishes any queued device copies.
synchronize()
del state_dict

def _fix_param_name_to_vllm(self, name: str, extra_prefixes: Optional[List[str]] = None) -> str:
Expand Down