From f7b0233ecac4a3100c20d377fcb1582abaa4bda9 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Sat, 16 Nov 2024 10:04:49 +0200 Subject: [PATCH] wip --- examples/llama-bench/llama-bench.cpp | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/examples/llama-bench/llama-bench.cpp b/examples/llama-bench/llama-bench.cpp index 8f4e0e206..8a01b453b 100644 --- a/examples/llama-bench/llama-bench.cpp +++ b/examples/llama-bench/llama-bench.cpp @@ -1337,7 +1337,14 @@ static void test_prompt(llama_context * ctx, int n_prompt, int n_batch, int n_th for (int i = 1; i < n_tokens; i++) { tokens[i] = std::rand() % n_vocab; } - llama_decode(ctx, llama_batch_get_one(tokens.data(), n_tokens)); + auto batch = llama_batch_get_one(tokens.data(), n_tokens); + int8_t logits[512]; + for (int i = 0; i < n_tokens; i++) { + logits[i] = 1; + } + batch.logits = logits; + + llama_decode(ctx, batch); n_processed += n_tokens; }