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

fix style

Browse source
This commit is contained in:
Minsoo Cheong 2024-02-27 15:29:14 +09:00
parent fb18827b4e
commit 34b942a429
1 changed files with 5 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

4
examples/speculative/speculative.cpp
View file

@ -210,10 +210,9 @@ int main(int argc, char ** argv) {
// stochastic verification
llama_token_data_array dist_tgt = llama_sampling_probability_distribution(ctx_sampling, ctx_tgt, NULL, drafts[s_keep].i_batch_tgt[i_dft]);
float p_tgt, p_dft;
// GGML_ASSERT(dist_tgt.size() == dist_dft.size());
// GGML_ASSERT(dist_tgt.size() == dist_dft.size());
for (int s = 0; s < n_seq_dft; ++s) {
if (!drafts[s].active) {
continue;
@ -234,7 +233,6 @@ int main(int argc, char ** argv) {
llama_token_data_array dist_dft = drafts[s].dist[i_dft];
// acquire the probability of the token from the draft model
for (int i = 0; i < dist_tgt.size; i++) {
if (dist_tgt.data[i].id == drafts[s].tokens[i_dft]) {
p_tgt = dist_tgt.data[i].p;
}