echodict/llama.cpp

1#include "models.h"2 3llm_build_qwen3vlmoe::llm_build_qwen3vlmoe(const llama_model & model, const llm_graph_params & params) : llm_graph_context(params) {4    const size_t n_deepstack_layers = hparams.n_deepstack_layers;5 6    const int64_t n_embd      = hparams.n_embd;7    const int64_t n_embd_head = hparams.n_embd_head_v();8 9    GGML_ASSERT(n_embd_head == hparams.n_embd_head_k());10    GGML_ASSERT(n_embd_head == n_rot);11 12    ggml_tensor * cur;13    ggml_tensor * inpL;14 15    inpL = build_inp_embd(model.tok_embd);16 17    int sections[4];18    std::copy(std::begin(hparams.rope_sections), std::begin(hparams.rope_sections) + 4, sections);19 20    // inp_pos - contains the positions21    ggml_tensor * inp_pos = build_inp_pos();22 23    auto * inp_attn = build_attn_inp_kv();24 25    ggml_tensor * inp_out_ids = build_inp_out_ids();26 27    for (int il = 0; il < n_layer; ++il) {28        ggml_tensor * inpSA = inpL;29 30        // norm31        cur = build_norm(inpL,32                model.layers[il].attn_norm, NULL,33                LLM_NORM_RMS, il);34        cb(cur, "attn_norm", il);35 36        // self_attention37        {38            // compute Q and K and RoPE them39            auto [Qcur, Kcur, Vcur] = build_qkv(model.layers[il], cur,40                    n_embd_head, n_head, n_head_kv, il);41 42            Qcur = build_norm(Qcur, model.layers[il].attn_q_norm, NULL, LLM_NORM_RMS, il);43            cb(Qcur, "Qcur_normed", il);44 45            Qcur = ggml_rope_multi(46                    ctx0, Qcur, inp_pos, nullptr,47                    n_rot, sections, rope_type, n_ctx_orig, freq_base, freq_scale,48                    ext_factor, attn_factor, beta_fast, beta_slow49                    );50 51            Kcur = build_norm(Kcur, model.layers[il].attn_k_norm, NULL, LLM_NORM_RMS, il);52            cb(Kcur, "Kcur_normed", il);53 54            Kcur = ggml_rope_multi(55                    ctx0, Kcur, inp_pos, nullptr,56                    n_rot, sections, rope_type, n_ctx_orig, freq_base, freq_scale,57                    ext_factor, attn_factor, beta_fast, beta_slow58                    );59 60            cb(Qcur, "Qcur", il);61            cb(Kcur, "Kcur", il);62            cb(Vcur, "Vcur", il);63 64            cur = build_attn(inp_attn,65                    model.layers[il].wo, model.layers[il].wo_b, model.layers[il].wo_s,66                    Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, 1.0f/sqrtf(float(n_embd_head)), il);67        }68 69        if (il == n_layer - 1 && inp_out_ids) {70            cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);71            inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);72        }73 74        ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);75        cb(ffn_inp, "ffn_inp", il);76 77        // MoE branch78        cur = build_norm(ffn_inp,79                model.layers[il].ffn_norm, NULL,80                LLM_NORM_RMS, il);81        cb(cur, "ffn_norm", il);82 83        ggml_tensor * moe_out =84            build_moe_ffn(cur,85                    model.layers[il].ffn_gate_inp,86                    model.layers[il].ffn_up_exps,87                    model.layers[il].ffn_gate_exps,88                    model.layers[il].ffn_down_exps,89                    nullptr,90                    n_expert, n_expert_used,91                    LLM_FFN_SILU, true,92                    hparams.expert_weights_scale,93                    LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,94                    il);95        cb(moe_out, "ffn_moe_out", il);96        cur = moe_out;97 98        cur = ggml_add(ctx0, cur, ffn_inp);99 100        cur = build_cvec(cur, il);101        cb(cur, "l_out", il);102 103        if (il < (int) n_deepstack_layers) {104            ggml_tensor * ds = ggml_view_2d(ctx0, res->t_inp_embd, n_embd, n_tokens, res->t_inp_embd->nb[1], (il + 1) * n_embd * sizeof(float));105            cur = ggml_add(ctx0, cur, ds);106            cb(cur, "deepstack_out", il);107        }108 109        // input for next layer110        inpL = cur;111    }112 113    cur = inpL;114 115    cur = build_norm(cur,116            model.output_norm, NULL,117            LLM_NORM_RMS, -1);118 119    cb(cur, "result_norm", -1);120    res->t_embd = cur;121 122    // lm_head123    cur = build_lora_mm(model.output, cur);124 125    cb(cur, "result_output", -1);126    res->t_logits = cur;127 128    ggml_build_forward_expand(gf, cur);129}130 131