Skip to content

Qwen4exp correctness fixes #27879

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
tarruda wants to merge 10 commits into from
Closed

Qwen4exp correctness fixes #27879

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
c07204d
qwen4exp: fix sparse-attention block selection
tarruda Aug 28, 2026
fce561d
qwen4exp: support independent PLE embedding widths
tarruda Aug 28, 2026
14b71e5
qwen4exp: validate model metadata
tarruda Aug 28, 2026
6ef2436
qwen4exp: update indexer cache after sequence copies
tarruda Aug 28, 2026
edb6dec
qwen4exp: enable recurrent state rollback
tarruda Aug 28, 2026
8427625
qwen4exp: disable tensor split
tarruda Aug 28, 2026
a7fc7e4
qwen4exp: fix GDN QK normalization
tarruda Aug 28, 2026
dfc7932
qwen4exp: avoid CUDA RMSNorm grid limit
tarruda Aug 29, 2026
01ac8a2
Merge branch 'master' into qwen4exp-correctness-fixes
tarruda Aug 30, 2026
50d6b3d
qwen4exp: share QSA block tables across queries
tarruda Aug 29, 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
2 changes: 2 additions & 0 deletions src/llama-arch.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1100,6 +1100,7 @@ bool llm_arch_supports_rs_rollback(const llm_arch & arch) {
switch (arch) {
case LLM_ARCH_QWEN35:
case LLM_ARCH_QWEN35MOE:
case LLM_ARCH_QWEN4EXP:
case LLM_ARCH_DEEPSEEK4:
case LLM_ARCH_NEMOTRON_H:
case LLM_ARCH_NEMOTRON_H_MOE:
Expand Down Expand Up @@ -1141,6 +1142,7 @@ bool llm_arch_supports_sm_tensor(const llm_arch & arch) {
case LLM_ARCH_KIMI_LINEAR:
case LLM_ARCH_BAILINGMOE3:
case LLM_ARCH_KIMI_K3:
case LLM_ARCH_QWEN4EXP:
case LLM_ARCH_QWEN3TTS:
return false;
default:
Expand Down
Loading