Skip to content

qwen3 vl skip layer id for pp #19135

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
ShangmingCai merged 2 commits into from
Apr 3, 2026
Merged

qwen3 vl skip layer id for pp #19135

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f873a80
qwen3 vl skip layer id for pp
narutolhy Feb 22, 2026
27311a4
fix
narutolhy Feb 22, 2026
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
12 changes: 12 additions & 0 deletions python/sglang/srt/models/qwen3_vl_moe.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,7 @@
from sglang.srt.eplb.expert_location import ModelConfigForExpertLocation
from sglang.srt.layers.moe.fused_moe_triton.layer import FusedMoE
from sglang.srt.layers.quantization.base_config import QuantizationConfig
from sglang.srt.layers.utils import get_layer_id
from sglang.srt.model_executor.forward_batch_info import ForwardBatch, PPProxyTensors
from sglang.srt.model_loader.weight_utils import default_weight_loader
from sglang.srt.models.qwen3_moe import Qwen3MoeDecoderLayer, Qwen3MoeModel
Expand Down Expand Up @@ -232,6 +233,17 @@ def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
params_dict = self._cached_params_dict
for name, loaded_weight in weights:
name = name.replace(r"model.language_model.", r"model.")
layer_id = get_layer_id(name)
if (
"visual" not in name
and layer_id is not None
and hasattr(self.model, "start_layer")
and (
layer_id < self.model.start_layer
or layer_id >= self.model.end_layer
)
):
continue

for param_name, weight_name, shard_id in stacked_params_mapping:
if "experts.gate_up_proj" in name or "experts.down_proj" in name:
Expand Down
Loading