From 1510016ea4af6de424f191d700eaff900933c695 Mon Sep 17 00:00:00 2001 From: Gabe Goodhart Date: Thu, 12 Jun 2025 14:00:53 -0600 Subject: [PATCH] fix: Remove logits_all after rebase Branch: HybridRecurrentCache Signed-off-by: Gabe Goodhart --- src/llama-kv-cache-hybrid-recurrent.cpp | 4 ++-- src/llama-kv-cache-hybrid-recurrent.h | 3 +-- 2 files changed, 3 insertions(+), 4 deletions(-) diff --git a/src/llama-kv-cache-hybrid-recurrent.cpp b/src/llama-kv-cache-hybrid-recurrent.cpp index 889f43025..49a7c35ab 100644 --- a/src/llama-kv-cache-hybrid-recurrent.cpp +++ b/src/llama-kv-cache-hybrid-recurrent.cpp @@ -56,10 +56,10 @@ llama_kv_cache_hybrid_recurrent::llama_kv_cache_hybrid_recurrent( n_seq_max )) {} -llama_memory_state_ptr llama_kv_cache_hybrid_recurrent::init_batch(const llama_batch & batch, uint32_t n_ubatch, bool embd_pooled, bool logits_all) { +llama_memory_state_ptr llama_kv_cache_hybrid_recurrent::init_batch(const llama_batch & batch, uint32_t n_ubatch, bool embd_pooled) { // since this includes a recurrent cache, we cannot use split_simple - auto sbatch = llama_sbatch(batch, hparams.n_embd, false, logits_all); + auto sbatch = llama_sbatch(batch, hparams.n_embd, false); // follow the recurrent pattern for creating the ubatch splits std::vector ubatches; diff --git a/src/llama-kv-cache-hybrid-recurrent.h b/src/llama-kv-cache-hybrid-recurrent.h index 444e87e10..d6678eb21 100644 --- a/src/llama-kv-cache-hybrid-recurrent.h +++ b/src/llama-kv-cache-hybrid-recurrent.h @@ -53,8 +53,7 @@ public: llama_memory_state_ptr init_batch( const llama_batch & batch, uint32_t n_ubatch, - bool embd_pooled, - bool logits_all) override; + bool embd_pooled) override; llama_memory_state_ptr init_full() override;