tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-29 20:45:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

model : gemma3n text-only (#14400)

Browse Source 
* gemma3n

* add llm_graph_input_one
This commit is contained in:
Xuan-Son Nguyen
2025-06-26 19:34:02 +02:00
committed by GitHub
parent a01047b041
commit 8846aace49
13 changed files with 960 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
src/llama-model.h
View File

@ -95,6 +95,8 @@ enum llm_type {
LLM_TYPE_17B_128E, // llama4 Maverick
LLM_TYPE_30B_A3B,
LLM_TYPE_235B_A22B,
LLM_TYPE_E2B,
LLM_TYPE_E4B,
};
std::string llama_rope_scaling_type_name(llama_rope_scaling_type rope_scaling_type);
@ -316,6 +318,19 @@ struct llama_layer {
struct ggml_tensor * ffn_up_scale = nullptr;
struct ggml_tensor * ffn_down_scale = nullptr;
// altup & laurel
struct ggml_tensor * per_layer_inp_gate = nullptr;
struct ggml_tensor * per_layer_proj = nullptr;
struct ggml_tensor * per_layer_post_norm = nullptr;
struct ggml_tensor * altup_correct_coef = nullptr;
struct ggml_tensor * altup_correct_scale = nullptr;
struct ggml_tensor * altup_predict_coef = nullptr;
struct ggml_tensor * altup_router = nullptr;
struct ggml_tensor * altup_router_norm = nullptr;
struct ggml_tensor * laurel_l = nullptr;
struct ggml_tensor * laurel_r = nullptr;
struct ggml_tensor * laurel_post_norm = nullptr;
struct llama_layer_posnet posnet;
struct llama_layer_convnext convnext;
@ -354,6 +369,13 @@ struct llama_model {
struct ggml_tensor * conv1d = nullptr;
struct ggml_tensor * conv1d_b = nullptr;
// gemma3n altup
struct ggml_tensor * tok_embd_per_layer = nullptr;
struct ggml_tensor * altup_proj = nullptr;
struct ggml_tensor * altup_unembd_proj = nullptr;
struct ggml_tensor * per_layer_model_proj = nullptr;
struct ggml_tensor * per_layer_proj_norm = nullptr;
std::vector<llama_layer> layers;
llama_model_params params;