CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-31 08:51:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : gemma3 : use output tensor if it exists in model weight (#12506)

Browse Source 
* llama : gemma3 : use output tensor if it exists in model weight

* also add to the llm_tensor_names
This commit is contained in:
Xuan-Son Nguyen
2025-03-22 23:28:19 +01:00
committed by GitHub
parent ba932dfb50
commit fbdfefe74e
3 changed files with 8 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
gguf-py/gguf/constants.py
View File

@@ -1113,6 +1113,7 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
], ],
MODEL_ARCH.GEMMA3: [ MODEL_ARCH.GEMMA3: [
MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.OUTPUT,
MODEL_TENSOR.OUTPUT_NORM, MODEL_TENSOR.OUTPUT_NORM,
MODEL_TENSOR.ATTN_Q, MODEL_TENSOR.ATTN_Q,
MODEL_TENSOR.ATTN_Q_NORM, MODEL_TENSOR.ATTN_Q_NORM,

1
src/llama-arch.cpp
View File

@@ -778,6 +778,7 @@ static const std::map<llm_arch, std::map<llm_tensor, const char *>> LLM_TENSOR_N
{ {
{ LLM_TENSOR_TOKEN_EMBD, "token_embd" }, { LLM_TENSOR_TOKEN_EMBD, "token_embd" },
{ LLM_TENSOR_OUTPUT_NORM, "output_norm" }, { LLM_TENSOR_OUTPUT_NORM, "output_norm" },
{ LLM_TENSOR_OUTPUT, "output" },
{ LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" }, { LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" },
{ LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" }, { LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" },
{ LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" }, { LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" },

7
src/llama-model.cpp
View File

@@ -2571,7 +2571,12 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
// output // output
output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0); output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED); // same as tok_embd, duplicated to allow offloading output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED);
// if output is NULL, init from the input tok embed
if (output == NULL) {
output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED);
}
for (int i = 0; i < n_layer; ++i) { for (int i = 0; i < n_layer; ++i) {
auto & layer = layers[i]; auto & layer = layers[i];