tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-01 05:05:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

context : remove logits_all flag

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-05-03 19:21:10 +03:00
parent 8733e0cf6e
commit 6c0501adf7
8 changed files with 6 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/llama-context.cpp
View File

@ -116,8 +116,6 @@ llama_context::llama_context(
__func__, n_ctx_per_seq, hparams.n_ctx_train);
}
logits_all = params.logits_all;
if (!hparams.vocab_only) {
// GPU backends
for (auto * dev : model.devices) {
@ -890,7 +888,7 @@ int llama_context::decode(llama_batch & inp_batch) {
for (uint32_t i = 0; i < n_tokens_all; ++i) {
n_outputs_all += batch.logits[i] != 0;
}
} else if (logits_all || embd_pooled) {
} else if (embd_pooled) {
n_outputs_all = n_tokens_all;
} else {
// keep last output only