tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-27 20:05:20 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : add support for DistilBert (#13907)

Browse Source 
* add distilbert

* small fixes

* add note for LLM_ARCH_DISTIL_BERT

* Use MODEL_ARCH.BERT for DistilBert

---------

Co-authored-by: dinhhuy <huy.dinh@brains-tech.co.jp>
This commit is contained in:
Đinh Trọng Huy
2025-05-30 18:56:02 +09:00
committed by GitHub
parent 2c90da4c7e
commit 291f2b6913
3 changed files with 37 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/llama-model.cpp
View File

@ -2114,7 +2114,7 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
case LLM_ARCH_NOMIC_BERT_MOE:
{
tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
type_embd = create_tensor(tn(LLM_TENSOR_TOKEN_TYPES, "weight"), {n_embd, n_token_types}, 0);
type_embd = create_tensor(tn(LLM_TENSOR_TOKEN_TYPES, "weight"), {n_embd, n_token_types}, TENSOR_NOT_REQUIRED);
if (arch == LLM_ARCH_BERT) {
pos_embd = create_tensor(tn(LLM_TENSOR_POS_EMBD, "weight"), {n_embd, n_ctx_train}, 0);
@ -5885,8 +5885,10 @@ struct llm_build_bert : public llm_graph_context {
inpL = build_inp_embd(model.tok_embd);
// token types are hardcoded to zero ("Sentence A")
ggml_tensor * type_row0 = ggml_view_1d(ctx0, model.type_embd, n_embd, 0);
inpL = ggml_add(ctx0, inpL, type_row0);
if (model.type_embd) {
ggml_tensor * type_row0 = ggml_view_1d(ctx0, model.type_embd, n_embd, 0);
inpL = ggml_add(ctx0, inpL, type_row0);
}
if (model.arch == LLM_ARCH_BERT) {
inpL = ggml_add(ctx0, ggml_get_rows(ctx0, model.pos_embd, inp_pos), inpL);
}