vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

address build warnings on llama.cpp

Browse source
This commit is contained in:
liuwei 2024-05-01 18:50:10 +00:00 • committed by Georgi Gerganov
parent 56d9fa72de
commit cc19780a55
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
1 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
llama.cpp
View file

@ -3345,7 +3345,7 @@ struct llama_model_loader {
GGML_ASSERT((arr_info.gt != GGUF_TYPE_INT32 || std::is_same<T, int>::value));
result.resize(arr_info.length);
result.assign((T*)arr_info.data, (T*)arr_info.data + arr_info.length);
result.assign((const T*)arr_info.data, (const T*)arr_info.data + arr_info.length);
return true;
}
@ -10979,7 +10979,7 @@ static void llama_set_inputs(llama_context & lctx, const llama_batch & batch) {
GGML_ASSERT(hparams.rope_short_factors.size() == freq_dim);
auto max_pos = batch.n_tokens > 0 && batch.pos != nullptr ? *std::max_element(batch.pos, batch.pos + batch.n_tokens) : batch.n_tokens - 1;
if (max_pos + 1 > hparams.n_yarn_orig_ctx) {
if ((uint32_t)(max_pos + 1) > hparams.n_yarn_orig_ctx) {
ggml_backend_tensor_set(lctx.freq_factors, hparams.rope_long_factors.data(), 0, freq_dim * ggml_element_size(lctx.freq_factors));
}
else {