Skip to content

mtmd : build_vit GQA support #23782

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
ngxson merged 1 commit into from
May 28, 2026
Merged

mtmd : build_vit GQA support #23782

Show file tree
Hide file tree
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
1 change: 1 addition & 0 deletions tools/mtmd/clip-graph.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,6 +29,7 @@ struct clip_graph {
const int n_patches;
const int n_embd;
const int n_head;
const int n_head_kv;
const int d_head;
const int n_layer;
const int n_mmproj_embd;
Expand Down
10 changes: 7 additions & 3 deletions tools/mtmd/clip.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -246,6 +246,7 @@ clip_graph::clip_graph(clip_ctx * ctx, const clip_image_f32 & img) :
n_patches(n_patches_x * n_patches_y),
n_embd(hparams.n_embd),
n_head(hparams.n_head),
n_head_kv(hparams.n_head_kv),
d_head(n_embd / n_head),
n_layer(hparams.n_layer),
n_mmproj_embd(clip_n_mmproj_embd(ctx)),
Expand Down Expand Up @@ -401,9 +402,9 @@ ggml_tensor * clip_graph::build_vit(
}
}

Qcur = ggml_reshape_3d(ctx0, Qcur, d_head, n_head, n_pos);
Kcur = ggml_reshape_3d(ctx0, Kcur, d_head, n_head, n_pos);
Vcur = ggml_reshape_3d(ctx0, Vcur, d_head, n_head, n_pos);
Qcur = ggml_reshape_3d(ctx0, Qcur, d_head, n_head, n_pos);
Kcur = ggml_reshape_3d(ctx0, Kcur, d_head, n_head_kv, n_pos);
Vcur = ggml_reshape_3d(ctx0, Vcur, d_head, n_head_kv, n_pos);

if (norm_per_head) {
if (layer.q_norm) {
Expand Down Expand Up @@ -1120,6 +1121,9 @@ struct clip_model_loader {
get_u32(string_format(KEY_PROJ_DIM, prefix), hparams.projection_dim);
get_f32(string_format(KEY_LAYER_NORM_EPS, prefix), hparams.eps);

// n_head_kv is optional (for GQA), default to n_head
hparams.n_head_kv = hparams.n_head;

if (is_vision) {
get_u32(KEY_IMAGE_SIZE, hparams.image_size);
get_u32(KEY_PATCH_SIZE, hparams.patch_size);
Expand Down
Loading