Merge branch 'master' into concedo_experimental

# Conflicts:
#	Makefile
#	tests/test-grad0.cpp
#	tests/test-quantize-perf.cpp
This commit is contained in:
Concedo 2023-12-13 14:49:03 +08:00
commit c2c238b4f3
20 changed files with 44 additions and 38 deletions

View file

@ -1,6 +1,6 @@
# llama.cpp/examples/speculative
Demonstartion of speculative decoding and tree-based speculative decoding techniques
Demonstration of speculative decoding and tree-based speculative decoding techniques
More info:

View file

@ -430,7 +430,7 @@ int main(int argc, char ** argv) {
++n_past_tgt;
}
// the first token is always proposed by the traget model before the speculation loop so we erase it here
// the first token is always proposed by the target model before the speculation loop so we erase it here
for (int s = 0; s < n_seq_dft; ++s) {
if (!drafts[s].active) {
continue;