Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-24 13:28:50 +01:00
Code Issues Packages Projects Releases Wiki Activity

simple : fix batch handling (#3803)

Browse Source
This commit is contained in:
Thibault Terrasson 2023-10-27 16:37:41 +02:00 committed by GitHub
parent 2f9ec7e271
commit c8d6a1f34a
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
examples/simple/simple.cpp
View File

@ -95,13 +95,8 @@ int main(int argc, char ** argv) {
llama_batch batch = llama_batch_init(512, 0, 1);
// evaluate the initial prompt
batch.n_tokens = tokens_list.size();
for (int32_t i = 0; i < batch.n_tokens; i++) {
batch.token[i] = tokens_list[i];
batch.pos[i] = i;
batch.seq_id[i] = 0;
batch.logits[i] = false;
for (size_t i = 0; i < tokens_list.size(); i++) {
llama_batch_add(batch, tokens_list[i], i, { 0 }, false);
}
// llama_decode will output logits only for the last token of the prompt
@ -148,15 +143,10 @@ int main(int argc, char ** argv) {
fflush(stdout);
// prepare the next batch
batch.n_tokens = 0;
llama_batch_clear(batch);
// push this new token for next evaluation
batch.token [batch.n_tokens] = new_token_id;
batch.pos [batch.n_tokens] = n_cur;
batch.seq_id[batch.n_tokens] = 0;
batch.logits[batch.n_tokens] = true;
batch.n_tokens += 1;
llama_batch_add(batch, new_token_id, n_cur, { 0 }, true);
n_decode += 1;
}