Skip to content

serve: one slot for Qwen3.5/3.8 on HRX when --parallel is not given #244

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
bong-water-water-bong merged 1 commit into from
Sep 30, 2026
Merged

serve: one slot for Qwen3.5/3.8 on HRX when --parallel is not given #244

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
7 changes: 7 additions & 0 deletions app/serve.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -717,6 +717,13 @@ Launch launch_for(const Options& o, const std::string& device, int child_port) {
const char* keep_split = std::getenv("ONEBIT_HRX_DECODE_SPLIT");
if (device == "hrx" && keep_split && std::string(keep_split) == "0")
env.push_back("GGML_HRX_DISABLE_DISPATCH=decode_split");
// Qwen3.5 / Qwen3.8 (gated delta net) decode one sequence per batch on HRX0. Without
// --parallel, llama-server's own default opens several slots, and a second concurrent
// request then fails (HTTP 500) at the multi-sequence softplus / GATED_DELTA_NET: one slot.
if (device == "hrx" && o.parallel == 0) {
const std::string arch = gguf_architecture(o.model);
if (arch == "qwen35" || arch == "qwen35moe" || arch == "qwen3next") argv.insert(argv.end(), {"-np", "1"});
}
} else if (device == "ds4") {
// DwarfStar: its own GGUF layouts only; it opens its port after the model has loaded
argv = {o.ds4.empty() ? default_ds4() : o.ds4, "-m", o.model,
Expand Down
Loading