echodict/llama.cpp

1#include "models.h"2 3llm_build_llada_moe::llm_build_llada_moe(const llama_model & model, const llm_graph_params & params) : llm_graph_context(params) {4    const int64_t n_embd_head = hparams.n_embd_head_v();5 6    GGML_ASSERT(n_embd_head == hparams.n_embd_head_k());7    GGML_ASSERT(n_embd_head == n_rot);8 9    ggml_tensor * cur;10    ggml_tensor * inpL;11 12    inpL = build_inp_embd(model.tok_embd);13 14    // inp_pos - contains the positions15    ggml_tensor * inp_pos = build_inp_pos();16 17    auto * inp_attn = build_attn_inp_no_cache();18 19    ggml_tensor * inp_out_ids = build_inp_out_ids();20 21    for (int il = 0; il < n_layer; ++il) {22        ggml_tensor * inpSA = inpL;23 24        // norm25        cur = build_norm(inpL,26                model.layers[il].attn_norm, NULL,27                LLM_NORM_RMS, il);28        cb(cur, "attn_norm", il);29 30        // self_attention31        {32            // compute Q and K and RoPE them33            auto [Qcur, Kcur, Vcur] = build_qkv(model.layers[il], cur,34                    n_embd_head, n_head, n_head_kv, il);35 36            Qcur = build_norm(Qcur, model.layers[il].attn_q_norm, NULL, LLM_NORM_RMS, il);37            cb(Qcur, "Qcur_normed", il);38 39            Kcur = build_norm(Kcur, model.layers[il].attn_k_norm, NULL, LLM_NORM_RMS, il);40            cb(Kcur, "Kcur_normed", il);41 42            Qcur = ggml_rope_ext(43                    ctx0, Qcur, inp_pos, nullptr,44                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,45                    ext_factor, attn_factor, beta_fast, beta_slow46                    );47 48            Kcur = ggml_rope_ext(49                    ctx0, Kcur, inp_pos, nullptr,50                    n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,51                    ext_factor, attn_factor, beta_fast, beta_slow52                    );53 54            cb(Qcur, "Qcur", il);55            cb(Kcur, "Kcur", il);56            cb(Vcur, "Vcur", il);57 58            cur = build_attn(inp_attn,59                    model.layers[il].wo, NULL, model.layers[il].wo_s,60                    Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, 1.0f/sqrtf(float(n_embd_head)), il);61        }62        if (il == n_layer - 1 && inp_out_ids) {63            cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);64            inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);65        }66        ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);67        cb(ffn_inp, "ffn_inp", il);68 69        // MoE branch70        cur = build_norm(ffn_inp,71                model.layers[il].ffn_norm, NULL,72                LLM_NORM_RMS, il);73        cb(cur, "ffn_norm", il);74 75        cur = build_moe_ffn(cur,76                model.layers[il].ffn_gate_inp,77                model.layers[il].ffn_up_exps,78                model.layers[il].ffn_gate_exps,79                model.layers[il].ffn_down_exps,80                nullptr,81                n_expert, n_expert_used,82                LLM_FFN_SILU, false,83                hparams.expert_weights_scale,84                LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,85                il);86        cb(cur, "ffn_moe_out", il);87 88        cur = ggml_add(ctx0, cur, ffn_inp);89 90        cur = build_cvec(cur, il);91        cb(cur, "l_out", il);92 93        // input for next layer94        inpL = cur;95    }96    cur = inpL;97 98    cur = build_norm(cur,99            model.output_norm, NULL,100            LLM_NORM_RMS, -1);101 102    cb(cur, "result_norm", -1);103    res->t_embd = cur;104 105    // lm_head106    cur = build_lora_mm(model.output, cur);107 108    cb(cur, "result_output", -1);109    res->t_logits = cur;110 111    ggml_build_forward_expand(gf, cur);112}113