Skip to content

Fix AWQ with enable MLA #2364

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
zhyncs merged 1 commit into from
Dec 5, 2024
Merged

Fix AWQ with enable MLA #2364

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
15 changes: 14 additions & 1 deletion python/sglang/srt/models/deepseek_v2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@
import torch
from torch import nn
from transformers import PretrainedConfig
from vllm import _custom_ops as ops
from vllm.distributed import (
get_tensor_model_parallel_rank,
get_tensor_model_parallel_world_size,
Expand Down Expand Up @@ -894,7 +895,19 @@ def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
if not global_server_args_dict["disable_mla"]:
for layer_id in range(self.config.num_hidden_layers):
self_attn = self.model.layers[layer_id].self_attn
w_kc, w_vc = self_attn.kv_b_proj.weight.unflatten(
if hasattr(self_attn.kv_b_proj, "qweight"):
# AWQ compatible
w = ops.awq_dequantize(
self_attn.kv_b_proj.qweight,
self_attn.kv_b_proj.scales,
self_attn.kv_b_proj.qzeros,
0,
0,
0,
).T
else:
w = self_attn.kv_b_proj.weight
w_kc, w_vc = w.unflatten(
0, (-1, self_attn.qk_nope_head_dim + self_attn.v_head_dim)
).split([self_attn.qk_nope_head_dim, self_attn.v_head_dim], dim=1)
self_attn.w_kc = w_kc.transpose(1, 2).contiguous().transpose(1, 2)
Expand Down
Loading