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

llama : disambiguate API

Browse source 
ggml-ci
This commit is contained in:
Georgi Gerganov 2024-09-02 10:06:42 +03:00
parent 8f1d81a0b6
commit 086e7f6ebc
No known key found for this signature in database
GPG key ID: BF970631944C16B7
9 changed files with 22 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

4
examples/embedding/embedding.cpp
View file

@ -31,7 +31,7 @@ static void batch_add_seq(llama_batch & batch, const std::vector<int32_t> & toke
}
static void batch_decode(llama_context * ctx, llama_batch & batch, float * output, int n_seq, int n_embd, int embd_norm) {
const enum llama_pooling_type pooling_type = llama_pooling_type(ctx);
const enum llama_pooling_type pooling_type = llama_get_pooling_type(ctx);
const struct llama_model * model = llama_get_model(ctx);
// clear previous kv_cache values (irrelevant for embeddings)
@ -114,7 +114,7 @@ int main(int argc, char ** argv) {
const int n_ctx_train = llama_n_ctx_train(model);
const int n_ctx = llama_n_ctx(ctx);
const enum llama_pooling_type pooling_type = llama_pooling_type(ctx);
const enum llama_pooling_type pooling_type = llama_get_pooling_type(ctx);
if (llama_model_has_encoder(model) && llama_model_has_decoder(model)) {
fprintf(stderr, "%s: error: computing embeddings in encoder-decoder models is not supported\n", __func__);