public_git/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/whisper.cpp.git synced 2026-10-08 07:21:21 +02:00
Code Issues Packages Projects Releases Wiki Activity

talk-llama : sync llama.cpp

Browse Source
This commit is contained in:
Georgi Gerganov
2026-01-14 09:11:59 +02:00
parent 13dc9a912b
commit ecfcc65fbf
43 changed files with 3208 additions and 668 deletions
Download Patch File Download Diff File Expand all files Collapse all files
examples/talk-llama/llama-arch.cpp
+22
View File
@@ -118,6 +118,7 @@ static const std::map<llm_arch, const char *> LLM_ARCH_NAMES = {
{ LLM_ARCH_MISTRAL3, "mistral3" },
{ LLM_ARCH_MIMO2, "mimo2" },
{ LLM_ARCH_LLAMA_EMBED, "llama-embed" },
{ LLM_ARCH_MAINCODER, "maincoder" },
{ LLM_ARCH_UNKNOWN, "(unknown)" },
};
@@ -151,6 +152,7 @@ static const std::map<llm_kv, const char *> LLM_KV_NAMES = {
{ LLM_KV_VOCAB_SIZE, "%s.vocab_size" },
{ LLM_KV_CONTEXT_LENGTH, "%s.context_length" },
{ LLM_KV_EMBEDDING_LENGTH, "%s.embedding_length" },
{ LLM_KV_EMBEDDING_LENGTH_OUT, "%s.embedding_length_out" },
{ LLM_KV_FEATURES_LENGTH, "%s.features_length" },
{ LLM_KV_BLOCK_COUNT, "%s.block_count" },
{ LLM_KV_LEADING_DENSE_BLOCK_COUNT, "%s.leading_dense_block_count" },
@@ -948,6 +950,8 @@ static std::set<llm_tensor> llm_get_tensor_names(llm_arch arch) {
LLM_TENSOR_ATTN_K_NORM,
LLM_TENSOR_ATTN_V,
LLM_TENSOR_ATTN_OUT,
LLM_TENSOR_ATTN_QKV,
LLM_TENSOR_ATTN_GATE,
LLM_TENSOR_FFN_NORM,
LLM_TENSOR_FFN_GATE_INP,
LLM_TENSOR_FFN_GATE_EXPS,
@@ -2074,6 +2078,7 @@ static std::set<llm_tensor> llm_get_tensor_names(llm_arch arch) {
LLM_TENSOR_TOKEN_EMBD,
LLM_TENSOR_OUTPUT_NORM_LFM2,
LLM_TENSOR_OUTPUT,
LLM_TENSOR_DENSE_2_OUT,
};
case LLM_ARCH_LFM2MOE:
return {
@@ -2234,6 +2239,23 @@ static std::set<llm_tensor> llm_get_tensor_names(llm_arch arch) {
return {
LLM_TENSOR_TOKEN_EMBD,
};
case LLM_ARCH_MAINCODER:
return {
LLM_TENSOR_TOKEN_EMBD,
LLM_TENSOR_OUTPUT_NORM,
LLM_TENSOR_OUTPUT,
LLM_TENSOR_ATTN_NORM,
LLM_TENSOR_ATTN_Q,
LLM_TENSOR_ATTN_Q_NORM,
LLM_TENSOR_ATTN_K,
LLM_TENSOR_ATTN_K_NORM,
LLM_TENSOR_ATTN_V,
LLM_TENSOR_ATTN_OUT,
LLM_TENSOR_FFN_NORM,
LLM_TENSOR_FFN_GATE,
LLM_TENSOR_FFN_DOWN,
LLM_TENSOR_FFN_UP,
};
default:
GGML_ABORT("unknown architecture for tensor mapping");
}