matt/ollama37
1
0
Fork 0
mirror of https://github.com/dogkeeper886/ollama37.git synced 2025-12-19 04:07:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama: update to commit de4c07f93 (#10655)

Browse Source
This commit is contained in:
Jeffrey Morgan
2025-05-12 12:17:26 -07:00
committed by GitHub
parent ad035ad595
commit 0cefd46f23
113 changed files with 8097 additions and 4383 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
llama/patches/0005-solar-pro.patch
View File

@@ -138,7 +138,7 @@ index 7ee6a5b7..48dce407 100644
};
diff --git a/src/llama-model-loader.cpp b/src/llama-model-loader.cpp
index ea73a8a7..a012aeae 100644
index 4cce5166..7f6617fa 100644
--- a/src/llama-model-loader.cpp
+++ b/src/llama-model-loader.cpp
@@ -439,6 +439,7 @@ namespace GGUFMeta {
@@ -150,10 +150,10 @@ index ea73a8a7..a012aeae 100644
llama_model_loader::llama_model_loader(
const std::string & fname,
diff --git a/src/llama-model.cpp b/src/llama-model.cpp
index 822e2bb2..572378c9 100644
index 3a4e72a3..831b68c0 100644
--- a/src/llama-model.cpp
+++ b/src/llama-model.cpp
@@ -1386,6 +1386,21 @@ void llama_model::load_hparams(llama_model_loader & ml) {
@@ -1402,6 +1402,21 @@ void llama_model::load_hparams(llama_model_loader & ml) {
default: type = LLM_TYPE_UNKNOWN;
}
} break;
@@ -175,7 +175,7 @@ index 822e2bb2..572378c9 100644
case LLM_ARCH_WAVTOKENIZER_DEC:
{
ml.get_key(LLM_KV_ATTENTION_LAYERNORM_EPS, hparams.f_norm_eps);
@@ -3741,6 +3756,34 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
@@ -3774,6 +3789,34 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
@@ -210,7 +210,7 @@ index 822e2bb2..572378c9 100644
layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);
layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0);
layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0);
@@ -12342,6 +12385,165 @@ struct llm_build_chameleon : public llm_graph_context {
@@ -12397,6 +12440,165 @@ struct llm_build_chameleon : public llm_graph_context {
}
};
@@ -376,7 +376,7 @@ index 822e2bb2..572378c9 100644
struct llm_build_wavtokenizer_dec : public llm_graph_context {
llm_build_wavtokenizer_dec(const llama_model & model, const llm_graph_params & params, ggml_cgraph * gf) : llm_graph_context(params) {
ggml_tensor * cur;
@@ -13092,6 +13294,10 @@ llm_graph_result_ptr llama_model::build_graph(
@@ -13157,6 +13359,10 @@ llm_graph_result_ptr llama_model::build_graph(
{
llm = std::make_unique<llm_build_chameleon>(*this, params, gf);
} break;
@@ -387,7 +387,7 @@ index 822e2bb2..572378c9 100644
case LLM_ARCH_WAVTOKENIZER_DEC:
{
llm = std::make_unique<llm_build_wavtokenizer_dec>(*this, params, gf);
@@ -13238,6 +13444,7 @@ llama_rope_type llama_model_rope_type(const llama_model * model) {
@@ -13301,6 +13507,7 @@ llama_rope_type llama_model_rope_type(const llama_model * model) {
case LLM_ARCH_GRANITE:
case LLM_ARCH_GRANITE_MOE:
case LLM_ARCH_CHAMELEON:
@@ -396,10 +396,10 @@ index 822e2bb2..572378c9 100644
return LLAMA_ROPE_TYPE_NORM;
diff --git a/src/llama-model.h b/src/llama-model.h
index 95eca002..856e6042 100644
index 6bdec263..43746c7d 100644
--- a/src/llama-model.h
+++ b/src/llama-model.h
@@ -64,6 +64,7 @@ enum llm_type {
@@ -65,6 +65,7 @@ enum llm_type {
LLM_TYPE_15B,
LLM_TYPE_16B,
LLM_TYPE_20B,
@@ -407,7 +407,7 @@ index 95eca002..856e6042 100644
LLM_TYPE_27B,
LLM_TYPE_30B,
LLM_TYPE_32B,
@@ -311,6 +312,8 @@ struct llama_layer {
@@ -315,6 +316,8 @@ struct llama_layer {
struct ggml_tensor * ffn_up_scale = nullptr;
struct ggml_tensor * ffn_down_scale = nullptr;