tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-29 04:35:05 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix llama_batch_ext_init_from_text

Browse Source
This commit is contained in:
Xuan Son Nguyen
2025-03-13 23:09:27 +01:00
parent 65f0184517
commit c3dd79007b
18 changed files with 40 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
common/common.cpp
View File

@ -1014,7 +1014,7 @@ struct common_init_result common_init_from_params(common_params & params) {
}
if (llama_model_has_encoder(model)) {
llama_batch_ext_ptr batch(llama_batch_ext_init_from_text(tmp.data(), tmp.size(), 0, 0));
llama_batch_ext_ptr batch(llama_batch_ext_init_from_text(tmp.data(), tmp.size(), 0, 0, true));
llama_encode_ext(lctx, batch.get());
llama_token decoder_start_token_id = llama_model_decoder_start_token(model);
if (decoder_start_token_id == LLAMA_TOKEN_NULL) {
@ -1024,7 +1024,7 @@ struct common_init_result common_init_from_params(common_params & params) {
tmp.push_back(decoder_start_token_id);
}
if (llama_model_has_decoder(model)) {
llama_batch_ext_ptr batch(llama_batch_ext_init_from_text(tmp.data(), std::min(tmp.size(), (size_t) params.n_batch), 0, 0));
llama_batch_ext_ptr batch(llama_batch_ext_init_from_text(tmp.data(), std::min(tmp.size(), (size_t) params.n_batch), 0, 0, true));
llama_decode_ext(lctx, batch.get());
}
llama_kv_self_clear(lctx);