graph : simplify attention api

ggml-ci

src/llama-context.cpp

@@ -2567,23 +2567,29 @@ void llama_context_kv_self::build_attn_inp(
     }
 }
 
-void llama_context_kv_self::build_attn_kv_store(
+ggml_tensor * llama_context_kv_self::build_attn(
         ggml_context * ctx0,
          ggml_cgraph * gf,
+         ggml_tensor * wo,
+         ggml_tensor * wo_b,
          ggml_tensor * k_cur,
          ggml_tensor * v_cur,
+         ggml_tensor * q_cur,
              int32_t   n_tokens,
-             int64_t   il,
+             float     kq_scale,
+             int       il,
              bool      worst_case) {
     const auto & hparams = model.hparams;
 
     const auto & n_ctx = cparams.n_ctx;
 
-    const auto kv_head = worst_case ? (kv_self.recurrent ? 0 : kv_self.size - n_tokens) : kv_self.head;
-
     const int64_t n_embd_k_gqa = hparams.n_embd_k_gqa(il);
     const int64_t n_embd_v_gqa = hparams.n_embd_v_gqa(il);
 
+    // store to KV cache
+    {
+        const auto kv_head = worst_case ? (kv_self.recurrent ? 0 : kv_self.size - n_tokens) : kv_self.head;
+
         GGML_ASSERT(kv_self.size == n_ctx);
 
         struct ggml_tensor * k_cache_view = ggml_view_1d(ctx0, kv_self.k_l[il], n_tokens*n_embd_k_gqa, ggml_row_size(kv_self.k_l[il]->type, n_embd_k_gqa)*kv_head);
@@ -2609,21 +2615,8 @@ void llama_context_kv_self::build_attn_kv_store(
         //cb(v_cache_view, "v_cache_view", il);
 
         ggml_build_forward_expand(gf, ggml_cpy(ctx0, v_cur, v_cache_view));
-}
+    }
 
-ggml_tensor * llama_context_kv_self::build_attn_qkv(
-        ggml_context * ctx0,
-         ggml_cgraph * gf,
-         ggml_tensor * wo,
-         ggml_tensor * wo_b,
-         ggml_tensor * q_cur,
-             int32_t   n_tokens,
-             float     kq_scale,
-             int       il,
-             bool      worst_case) {
-    const auto & hparams = model.hparams;
-
-    const auto & n_ctx         = cparams.n_ctx;
     const auto & n_embd_head_k = hparams.n_embd_head_k;
     const auto & n_embd_head_v = hparams.n_embd_head_v;
 
@@ -2657,8 +2650,6 @@ ggml_tensor * llama_context_kv_self::build_attn_qkv(
 
     const int64_t n_head       = hparams.n_head(il);
     const int64_t n_head_kv    = hparams.n_head_kv(il);
-    const int64_t n_embd_k_gqa = hparams.n_embd_k_gqa(il);
-    const int64_t n_embd_v_gqa = hparams.n_embd_v_gqa(il);
 
     struct ggml_tensor * q = ggml_permute(ctx0, q_cur, 0, 2, 1, 3);
     //cb(q, "q", il);

src/llama-context.h

@@ -376,20 +376,13 @@ public:
                     bool   swa,
                     bool   worst_case) override;
 
-    virtual void build_attn_kv_store(
+    virtual ggml_tensor * build_attn(
-            ggml_context * ctx0,
-             ggml_cgraph * gf,
-             ggml_tensor * k_cur,
-             ggml_tensor * v_cur,
-                 int32_t   n_tokens,
-                 int64_t   il,
-                 bool      worst_case) override;
-
-    virtual ggml_tensor * build_attn_qkv(
             ggml_context * ctx0,
              ggml_cgraph * gf,
              ggml_tensor * wo,
              ggml_tensor * wo_b,
+             ggml_tensor * k_cur,
+             ggml_tensor * v_cur,
              ggml_tensor * q_cur,
                  int32_t   n_tokens,
                  float     kq_scale,
@@ -443,6 +436,7 @@ protected:
 
 // a recurrent transformer (ie.e RWKV, Mamba)
 // TODO: temporary reuse kv_self, but in the future, implement recurrent-specific context with specific cache
+//class llama_context_recurrent : public llama_context {
 class llama_context_recurrent : public llama_context_kv_self {
 public:
     llama_context_recurrent(

src/llama-graph.h

@@ -88,20 +88,13 @@ public:
                     bool   swa,
                     bool   worst_case) = 0;
 
-    virtual void build_attn_kv_store(
+    virtual ggml_tensor * build_attn(
-            ggml_context * ctx0,
-             ggml_cgraph * gf,
-             ggml_tensor * k_cur,
-             ggml_tensor * v_cur,
-                 int32_t   n_tokens,
-                 int64_t   il,
-                 bool      worst_case) = 0;
-
-    virtual ggml_tensor * build_attn_qkv(
             ggml_context * ctx0,
              ggml_cgraph * gf,
              ggml_tensor * wo,
              ggml_tensor * wo_b,
+             ggml_tensor * k_cur,
+             ggml_tensor * v_cur,
              ggml_tensor * q_cur,
                  int32_t   n_tokens,
                  float     kq_scale,

src/llama-model.cpp

@@ -4258,13 +4258,7 @@ struct llm_build_context {
         ggml_build_forward_expand(gf, k_cur);
         ggml_build_forward_expand(gf, v_cur);
 
-        //build_kv_store(gf, k_cur, v_cur, il);
+        ggml_tensor * cur = lgf->build_attn(ctx0, gf, wo, wo_b, k_cur, v_cur, q_cur, n_tokens, kq_scale, il, worst_case);
-        lgf->build_attn_kv_store(ctx0, gf, k_cur, v_cur, n_tokens, il, worst_case);
-
-        struct ggml_tensor * cur;
-
-        //cur = build_kqv(gf, wo, wo_b, q_cur, kq_mask, kq_scale, il);
-        cur = lgf->build_attn_qkv(ctx0, gf, wo, wo_b, q_cur, n_tokens, kq_scale, il, worst_case);
         cb(cur, "kqv_out", il);
 
         return cur;