Skip to content

[Bugfix] Fix chunked prefill for GGUF #14666

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
merged 2 commits into from
Mar 13, 2025
Merged

[Bugfix] Fix chunked prefill for GGUF #14666

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6f58c56
fix chunked prefill for GGUF
SzymonOzog Mar 12, 2025
252fa99
rerun ci
SzymonOzog Mar 12, 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: 7 additions & 0 deletions vllm/model_executor/layers/quantization/gguf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -98,6 +98,13 @@ def get_quant_method(self, layer: torch.nn.Module,

def _fuse_mul_mat(x: torch.Tensor, qweight: torch.Tensor,
qweight_type: int) -> torch.Tensor:
# HACK: when doing chunked prefill we don't generate output tokens
# so input to logits generator is empty which causes invalid parameter
if x.shape[0] == 0:
return torch.empty(x.shape[0],
qweight.shape[0],
dtype=x.dtype,
device=x.device)
# there is no need to call any kernel for fp16/bf16
if qweight_type in UNQUANTIZED_TYPES:
return x @ qweight.T
Expand Down