Skip to content

Update clip.cpp #9482

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
wants to merge 7 commits into from
Closed

Update clip.cpp #9482

Changes from 5 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
aa9e721
Update clip.cpp
Tejaakshaykumar Sep 14, 2024
28d1c45
Update examples/llava/clip.cpp
Tejaakshaykumar Sep 17, 2024
cba0340
Refactored error handling for hyperparameter validation in clip.cpp
Tejaakshaykumar Sep 17, 2024
e08b907
Update clip.cpp
Tejaakshaykumar Sep 18, 2024
9bea433
Update clip.cpp by deleting optional logs
Tejaakshaykumar Sep 23, 2024
36d9bbc
Updated examples/llava/clip.cpp
Tejaakshaykumar Sep 24, 2024
46d20e1
Updated clip.cpp
Tejaakshaykumar Sep 24, 2024
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
26 changes: 17 additions & 9 deletions examples/llava/clip.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1270,15 +1270,23 @@ struct clip_ctx * clip_model_load(const char * fname, const int verbosity = 1) {
// load vision model
auto & vision_model = new_clip->vision_model;
auto & hparams = vision_model.hparams;
hparams.hidden_size = get_u32(ctx, format(KEY_N_EMBD, "vision"));
hparams.n_head = get_u32(ctx, format(KEY_N_HEAD, "vision"));
hparams.n_intermediate = get_u32(ctx, format(KEY_N_FF, "vision"));
hparams.n_layer = get_u32(ctx, format(KEY_N_BLOCK, "vision"));
hparams.image_size = get_u32(ctx, KEY_IMAGE_SIZE);
hparams.patch_size = get_u32(ctx, KEY_PATCH_SIZE);
hparams.projection_dim = get_u32(ctx, format(KEY_PROJ_DIM, "vision"));
hparams.eps = get_f32(ctx, format(KEY_LAYER_NORM_EPS, "vision"));

try{
hparams.hidden_size = get_u32(ctx, format(KEY_N_EMBD, "vision"));
hparams.n_head = get_u32(ctx, format(KEY_N_HEAD, "vision"));
hparams.n_intermediate = get_u32(ctx, format(KEY_N_FF, "vision"));
hparams.n_layer = get_u32(ctx, format(KEY_N_BLOCK, "vision"));
hparams.image_size = get_u32(ctx, KEY_IMAGE_SIZE);
hparams.patch_size = get_u32(ctx, KEY_PATCH_SIZE);
hparams.projection_dim = get_u32(ctx, format(KEY_PROJ_DIM, "vision"));
hparams.eps = get_f32(ctx, format(KEY_LAYER_NORM_EPS, "vision"));
if (hparams.hidden_size <= 0 || hparams.n_head <= 0 || hparams.n_layer <= 0 || hparams.n_intermediate <= 0 || hparams.image_size <= 0 || hparams.patch_size <= 0 || hparams.projection_dim <= 0 || hparams.eps <= 0) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am not sure that these checks are necessary.

LOG_TEE("Error: Invalid hyperparameter values\n");
return false;
}
} catch (const std::exception& e) {
LOG_TEE("Error while loading hyperparameters: %s\n", e.what());
     }

try {
int idx = get_key_idx(ctx, KEY_IMAGE_GRID_PINPOINTS);
int n = gguf_get_arr_n(ctx, idx);
Expand Down