ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

llama : add option to override model tensor buffers (#11397)

Browse source 
* llama : add option to override tensor buffers

* ggml : fix possible underflow in ggml_nbytes
This commit is contained in:
Diego Devesa 2025-04-02 14:52:01 +02:00 • committed by GitHub
parent a10b36c91a
commit e0e912f49b
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
12 changed files with 108 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

8
src/llama-model-loader.h
View file

@ -77,8 +77,9 @@ struct llama_model_loader {
llama_mmaps mappings;
std::map<std::string, struct llama_tensor_weight, weight_name_comparer> weights_map;
std::unordered_map<std::string, struct llama_model_kv_override> kv_overrides;
std::map<std::string, llama_tensor_weight, weight_name_comparer> weights_map;
std::unordered_map<std::string, llama_model_kv_override> kv_overrides;
const llama_model_tensor_buft_override * tensor_buft_overrides;
gguf_context_ptr meta;
std::vector<ggml_context_ptr> contexts;
@ -95,7 +96,8 @@ struct llama_model_loader {
std::vector<std::string> & splits, // optional, only need if the split does not follow naming scheme
bool use_mmap,
bool check_tensors,
const struct llama_model_kv_override * param_overrides_p);
const llama_model_kv_override * param_overrides_p,
const llama_model_tensor_buft_override * param_tensor_buft_overrides_p);
template<typename T>
typename std::enable_if<std::is_integral<T>::value, bool>::type