From 8d608a81b7bd170f700648f8214e6f3279d4d715 Mon Sep 17 00:00:00 2001 From: l3utterfly Date: Thu, 2 May 2024 04:27:41 +0900 Subject: [PATCH] main : fix off by one error for context shift (#6921) --- examples/main/main.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/main/main.cpp b/examples/main/main.cpp index 5c693657c..eabbc2db3 100644 --- a/examples/main/main.cpp +++ b/examples/main/main.cpp @@ -544,7 +544,7 @@ int main(int argc, char ** argv) { // if we run out of context: // - take the n_keep first tokens from the original prompt (via n_past) // - take half of the last (n_ctx - n_keep) tokens and recompute the logits in batches - if (n_past + (int) embd.size() + std::max(0, guidance_offset) > n_ctx) { + if (n_past + (int) embd.size() + std::max(0, guidance_offset) >= n_ctx) { if (params.n_predict == -2) { LOG_TEE("\n\n%s: context full and n_predict == -%d => stopping\n", __func__, params.n_predict); break;