vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

GGUF: backend support, fixed-width I/O, misc fixes

Browse source
This commit is contained in:
Johannes Gäßler 2024-12-03 21:43:57 +01:00
parent cc98896db8
commit b88727009d
5 changed files with 334 additions and 271 deletions
Download patch file Download diff file Expand all files Collapse all files

3
examples/llava/clip.cpp
View file

@ -2566,7 +2566,8 @@ bool clip_model_quantize(const char * fname_inp, const char * fname_out, const i
total_size_org += orig_size;
total_size_new += new_size;
gguf_set_tensor_type(ctx_out, name.c_str(), new_type);
gguf_set_tensor_data(ctx_out, name.c_str(), new_data, new_size);
GGML_ASSERT(gguf_get_tensor_size(ctx_out, gguf_find_tensor(ctx_out, name.c_str())) == new_size);
gguf_set_tensor_data(ctx_out, name.c_str(), new_data);
fout.write((const char *)new_data, new_size);
size_t pad = GGML_PAD(new_size, gguf_get_alignment(ctx_out)) - new_size;
for (size_t j = 0; j < pad; ++j) {