char-fork/llama.cpp

git clone https://git.t4t.associates/char-fork/llama.cpp

Pedro Cuencallama: Restore quantization of mmprojs (#26818)86c298fb8

char/aliceai-t5-moe
brancheschar/aliceai-t5-moeall refs →
636 B18 linesraw
1#include "models.h"
2
3// Stub to allow llama-quantize to open mmproj GGUFs
4
5[[noreturn]]
6void llama_model_clip::load_arch_hparams(llama_model_loader &) {
7    GGML_ABORT("CLIP is a quant-only stub; load_arch_hparams should not be called");
8}
9
10[[noreturn]]
11void llama_model_clip::load_arch_tensors(llama_model_loader &) {
12    GGML_ABORT("CLIP is a quant-only stub; load_arch_tensors should not be called");
13}
14
15[[noreturn]]
16std::unique_ptr<llm_graph_context> llama_model_clip::build_arch_graph(const llm_graph_params &) const {
17    GGML_ABORT("CLIP has no inference graph via llama_model dispatch; runtime lives in tools/mtmd/clip.cpp");
18}