diff --git a/examples/perplexity/perplexity.cpp b/examples/perplexity/perplexity.cpp index ae21bfbaf..e1ea57224 100644 --- a/examples/perplexity/perplexity.cpp +++ b/examples/perplexity/perplexity.cpp @@ -400,6 +400,7 @@ static results_perplexity perplexity_v2(llama_context * ctx, const gpt_params & // clear the KV cache llama_kv_cache_clear(ctx); + for (int j = 0; j < num_batches; ++j) { const int batch_start = start + j * n_batch; const int batch_size = std::min(end - batch_start, n_batch); diff --git a/include/llama.h b/include/llama.h index e942c9d3f..ed5f9bd42 100644 --- a/include/llama.h +++ b/include/llama.h @@ -230,7 +230,7 @@ extern "C" { llama_pos * pos; int32_t * n_seq_id; llama_seq_id ** seq_id; - int8_t * output; // Previously named 'logits', renamed to 'output' now. + int8_t * output; // Previously named "logits", renamed to "output" now. // NOTE: helpers for smooth API transition - can be deprecated in the future // for future-proof code, use the above fields instead and ignore everything below