Skip to content

feat: Muse Glimmer support #639

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
giladgd merged 10 commits into from
Aug 11, 2026
Merged

feat: Muse Glimmer support #639

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
20fbc3a
fix: improve thought segment syntax extraction, improve function call…
giladgd Aug 10, 2026
f6074dd
fix: support more quant labels
giladgd Aug 10, 2026
96fcfb0
test: segment setting extraction
giladgd Aug 10, 2026
bcf46c5
feat(`inspect ggpu` command): print model parameters count
giladgd Aug 10, 2026
e83cdf2
chore: update modules
giladgd Aug 10, 2026
fb6725b
fix: improve chat wrapper equivalency check
giladgd Aug 10, 2026
bb6e630
fix: topP config when using temperature
giladgd Aug 10, 2026
3390a8c
test: fix tests
giladgd Aug 10, 2026
e9bb1f5
fix: bugs
giladgd Aug 10, 2026
76c6c8b
feat: Must Glimmer support
giladgd Aug 11, 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: 1 addition & 1 deletion llama/addon/AddonContext.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -235,7 +235,7 @@ class AddonContextSampleTokenWorker : public Napi::AsyncWorker {
llama_token_data_array cur_p;
sampler->sample(ctx->ctx, batchLogitIndex, cur_p, returnProbabilities || returnConfidence);

if (!(cur_p.selected >= 0 && cur_p.selected < (int32_t)cur_p.size)) {
if (cur_p.size == 0 || !(cur_p.selected >= 0 && cur_p.selected < (int32_t)cur_p.size)) {
no_output = true;
return;
}
Expand Down
10 changes: 7 additions & 3 deletions llama/addon/AddonSampler.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -173,6 +173,10 @@ void AddonSampler::acceptToken(llama_token token) {
void AddonSampler::sample(struct llama_context* llamaContext, int32_t batchLogitIndex, llama_token_data_array& curP, bool forceGrammar) {
setTokenCandidates(llamaContext, batchLogitIndex, curP);

if (curP.size == 0) {
return;
}

if (forceGrammar && grammarEvaluationState != nullptr && grammarEvaluationState->sampler != nullptr) {
llama_sampler_apply(grammarEvaluationState->sampler, &curP);
llama_sampler_apply(chain, &curP);
Expand Down Expand Up @@ -287,7 +291,7 @@ Napi::Value AddonSampler::ApplyConfig(const Napi::CallbackInfo& info) {
temperatureSampler = nullptr;
}

if (temperatureSampler_temperature <= 0) {
if (temperatureSampler_temperature <= 0.0f) {
greedySampler = llama_sampler_init_greedy();
} else {
temperatureSampler = llama_sampler_init_temp(temperatureSampler_temperature);
Expand Down Expand Up @@ -321,7 +325,7 @@ Napi::Value AddonSampler::ApplyConfig(const Napi::CallbackInfo& info) {
minPSampler = nullptr;
}

if (minPSampler_minP != 0) {
if (minPSampler_minP > 0.0f) {
minPSampler = llama_sampler_init_min_p(minPSampler_minP, min_keep);
}
}
Expand Down Expand Up @@ -366,7 +370,7 @@ Napi::Value AddonSampler::ApplyConfig(const Napi::CallbackInfo& info) {
topPSampler = nullptr;
}

if (topPSampler_topP >= 1) {
if (topPSampler_topP >= 0.0f && topPSampler_topP <= 1.0f) {
topPSampler = llama_sampler_init_top_p(topPSampler_topP, min_keep);
}
}
Expand Down
2 changes: 1 addition & 1 deletion llama/addon/AddonSampler.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ class AddonSampler : public Napi::ObjectWrap<AddonSampler> {
int topKSampler_topK = 0;

llama_sampler * topPSampler = nullptr;
float topPSampler_topP = 0.0f; // Top p sampling >=1.0 = disabled
float topPSampler_topP = 1.0f; // Top p sampling >=1.0 = disabled

llama_sampler * xtcSampler = nullptr;
float xtcSampler_probability = 0;
Expand Down
113 changes: 75 additions & 38 deletions package-lock.json
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

10 changes: 5 additions & 5 deletions package.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -144,7 +144,7 @@
"@nolebase/vitepress-plugin-og-image": "^2.18.2",
"@resvg/resvg-js": "^2.6.2",
"@semantic-release/exec": "^7.1.0",
"@semantic-release/github": "^12.0.6",
"@semantic-release/github": "^12.0.9",
"@semantic-release/npm": "^13.1.5",
"@shikijs/vitepress-twoslash": "^3.22.0",
"@stylistic/eslint-plugin": "^5.8.0",
Expand All @@ -170,7 +170,7 @@
"husky": "^9.1.7",
"rehype": "^13.0.2",
"rimraf": "^6.1.3",
"semantic-release": "^25.0.3",
"semantic-release": "^25.0.8",
"sharp": "^0.34.5",
"tslib": "^2.8.1",
"typedoc": "^0.28.17",
Expand All @@ -186,7 +186,7 @@
"zx": "^8.8.5"
},
"dependencies": {
"@huggingface/jinja": "^0.5.6",
"@huggingface/jinja": "^0.5.9",
"async-retry": "^1.3.3",
"bytes": "^3.1.2",
"chalk": "^5.6.2",
Expand All @@ -202,12 +202,12 @@
"lifecycle-utils": "^4.0.1",
"log-symbols": "^7.0.1",
"nanoid": "^5.1.6",
"node-addon-api": "^8.6.0",
"node-addon-api": "^8.9.1",
"ora": "^9.3.0",
"pretty-ms": "^9.3.0",
"proper-lockfile": "^4.1.2",
"semver": "^7.7.1",
"simple-git": "^3.33.0",
"simple-git": "^3.36.0",
"slice-ansi": "^8.0.0",
"stdout-update": "^4.0.1",
"strip-ansi": "^7.2.0",
Expand Down
Loading
Loading