Brunobkr/llama.cpp_AlgMor24_github

1#include "models.h"2 3void llama_model_afmoe::load_arch_hparams(llama_model_loader & ml) {4    ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);5    ml.get_key(LLM_KV_LEADING_DENSE_BLOCK_COUNT,   hparams.n_layer_dense_lead, false);6    ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH,  hparams.n_ff_exp);7    ml.get_key(LLM_KV_EXPERT_SHARED_COUNT,         hparams.n_expert_shared);8    ml.get_key(LLM_KV_EXPERT_GATING_FUNC,          hparams.expert_gating_func, false);9    ml.get_key(LLM_KV_EXPERT_WEIGHTS_SCALE,        hparams.expert_weights_scale, false);10    ml.get_key(LLM_KV_EXPERT_WEIGHTS_NORM,         hparams.expert_weights_norm, false);11    ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW,    hparams.n_swa, false);12 13    // Set up interleaved sliding window attention (ISWA)14    // Pattern: 3 sliding - 1 full (global_attn_every_n_layers = 4)15    if (hparams.n_swa > 0) {16        hparams.swa_type = LLAMA_SWA_TYPE_STANDARD;17        uint32_t swa_period = 4;18        ml.get_key_or_arr(LLM_KV_ATTENTION_SLIDING_WINDOW_PATTERN, swa_period, false);19        hparams.set_swa_pattern(swa_period);20 21        hparams.rope_freq_base_train_swa  = hparams.rope_freq_base_train;22        hparams.rope_freq_scale_train_swa = hparams.rope_freq_scale_train;23        ml.get_key(LLM_KV_ROPE_FREQ_BASE_SWA, hparams.rope_freq_base_train_swa, false);24    } else {25        hparams.swa_type = LLAMA_SWA_TYPE_NONE;26    }27 28    // Default to sigmoid if not set29    if (hparams.expert_gating_func == LLAMA_EXPERT_GATING_FUNC_TYPE_NONE) {30        hparams.expert_gating_func = LLAMA_EXPERT_GATING_FUNC_TYPE_SIGMOID;31    }32 33    switch (hparams.n_layer()) {34        case 56: type = LLM_TYPE_6B; break;35        case 32: type = LLM_TYPE_26B; break;36        default: type = LLM_TYPE_UNKNOWN;37    }38}39 40void llama_model_afmoe::load_arch_tensors(llama_model_loader &) {41    LLAMA_LOAD_LOCALS;42    const int64_t n_expert_shared = hparams.n_expert_shared;43 44    tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);45 46    // output47    output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);48    output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED);49 50    // if output is NULL, init from the input tok embed51    if (output == NULL) {52        output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED);53    }54 55    const int64_t n_ff_exp = hparams.n_ff_exp;56 57    for (int i = 0; i < n_layer; ++i) {58        auto & layer = layers[i];59 60        // dual attention normalization61        layer.attn_norm      = create_tensor(tn(LLM_TENSOR_ATTN_NORM,      "weight", i), {n_embd}, 0);62        layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), {n_embd}, 0);63 64        // attention projections65        create_tensor_qkv(layer, i, n_embd, n_embd_head_k * n_head, n_embd_k_gqa, n_embd_v_gqa, 0);66        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);67 68        // Q/K normalization69        layer.attn_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k}, 0);70        layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k}, 0);71 72        // attention gating73        layer.wqkv_gate = create_tensor(tn(LLM_TENSOR_ATTN_GATE, "weight", i), {n_embd, n_embd_head_k * n_head}, 0);74 75        // dual ffn normalization76        layer.ffn_norm      = create_tensor(tn(LLM_TENSOR_FFN_NORM,      "weight", i), {n_embd}, 0);77        layer.ffn_post_norm = create_tensor(tn(LLM_TENSOR_FFN_POST_NORM, "weight", i), {n_embd}, 0);78 79        if (static_cast<uint32_t>(i) >= hparams.n_layer_dense_lead) {80            // MoE layers81            layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert}, 0);82            layer.ffn_exp_probs_b = create_tensor(tn(LLM_TENSOR_FFN_EXP_PROBS_B, "bias", i), {n_expert}, 0);83 84            // grouped expert weights85            layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff_exp, n_expert}, 0);86            layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), {n_ff_exp, n_embd, n_expert}, 0);87            layer.ffn_up_exps   = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS,   "weight", i), {n_embd, n_ff_exp, n_expert}, 0);88 89            // shared expert90            if (n_expert_shared > 0) {91                const int64_t n_ff_shexp = n_ff_exp * n_expert_shared;92                layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), {n_embd, n_ff_shexp}, 0);93                layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), {n_ff_shexp, n_embd}, 0);94                layer.ffn_up_shexp   = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP,   "weight", i), {n_embd, n_ff_shexp}, 0);95            }96        } else {97            // Dense layers98            layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);99            layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {n_ff, n_embd}, 0);100            layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd, n_ff}, 0);101        }102    }103}104 105std::unique_ptr<llm_graph_context> llama_model_afmoe::build_arch_graph(const llm_graph_params & params) const {106    return std::make_unique<graph>(*this, params);107}108 109llama_model_afmoe::graph::graph(const llama_model & model, const llm_graph_params & params) : llm_graph_context(params) {110    const int64_t n_embd_head = hparams.n_embd_head_v();111    GGML_ASSERT(n_embd_head == hparams.n_embd_head_k());112 113    ggml_tensor * cur;114    ggml_tensor * inpL;115 116    inpL = build_inp_embd(model.tok_embd);117 118    // MuP scaling: embeddings * sqrt(hidden_size)119    // mup_enabled = true, hidden_size = 1024, scale = 32.0120    inpL = ggml_scale(ctx0, inpL, sqrtf(float(n_embd)));121    cb(inpL, "inp_embd_scaled", -1);122 123    // inp_pos - contains the positions124    ggml_tensor * inp_pos = build_inp_pos();125    auto * inp_attn = build_attn_inp_kv_iswa();126    ggml_tensor * inp_out_ids = build_inp_out_ids();127 128    const float kq_scale = 1.0f/sqrtf(float(n_embd_head));129 130    for (int il = 0; il < n_layer; ++il) {131        const float freq_base_l  = model.get_rope_freq_base (cparams, il);132        const float freq_scale_l = model.get_rope_freq_scale(cparams, il);133 134        ggml_tensor * inpSA = inpL;135 136        // This overlaps with SWA layers in current models, so get_rope_freq_base/scale may be superfluous137        const bool use_rope = hparams.n_no_rope_layer_step > 0 &&138                              (il + 1) % hparams.n_no_rope_layer_step != 0;139 140        // dual attention normalization (pre)141        cur = build_norm(inpL,142                model.layers[il].attn_norm, NULL,143                LLM_NORM_RMS, il);144        cb(cur, "attn_norm", il);145 146        // self-attention147        {148            ggml_tensor * attn_inp = cur;  // save input for gate computation149 150            auto [Qcur, Kcur, Vcur] = build_qkv(model.layers[il], cur,151                    n_embd_head, n_head, n_head_kv, il);152 153            // compute gate from input154            ggml_tensor * gate = build_lora_mm(model.layers[il].wqkv_gate, attn_inp);155            cb(gate, "attn_gate_proj", il);156 157            // Q/K normalization158            Qcur = build_norm(Qcur, model.layers[il].attn_q_norm, NULL, LLM_NORM_RMS, il);159            Kcur = build_norm(Kcur, model.layers[il].attn_k_norm, NULL, LLM_NORM_RMS, il);160            cb(Qcur, "Qcur_normed", il);161            cb(Kcur, "Kcur_normed", il);162 163            if (use_rope) {164                Qcur = ggml_rope_ext(165                        ctx0, Qcur, inp_pos, nullptr,166                        n_rot, rope_type, n_ctx_orig, freq_base_l, freq_scale_l,167                        ext_factor, attn_factor, beta_fast, beta_slow);168                cb(Qcur, "Qcur_rope", il);169 170                Kcur = ggml_rope_ext(171                        ctx0, Kcur, inp_pos, nullptr,172                        n_rot, rope_type, n_ctx_orig, freq_base_l, freq_scale_l,173                        ext_factor, attn_factor, beta_fast, beta_slow);174                cb(Kcur, "Kcur_rope", il);175            }176 177            cur = build_attn(inp_attn,178                    NULL, NULL, NULL,  // wo will be applied after gating179                    Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, kq_scale, il);180            cb(cur, "attn_out", il);181 182            // attention gating: attn_out * sigmoid(gate) BEFORE o_proj183            gate = ggml_sigmoid(ctx0, gate);184            cb(gate, "attn_gate_sig", il);185            cur = ggml_mul(ctx0, cur, gate);186            cb(cur, "attn_gated", il);187 188            // now apply output projection189            cur = build_lora_mm(model.layers[il].wo, cur, model.layers[il].wo_s);190            cb(cur, "attn_o_proj", il);191        }192 193        // dual attention normalization (post)194        cur = build_norm(cur,195                model.layers[il].attn_post_norm, NULL,196                LLM_NORM_RMS, il);197        cb(cur, "attn_post_norm", il);198 199        if (il == n_layer - 1 && inp_out_ids) {200            cur   = ggml_get_rows(ctx0,   cur, inp_out_ids);201            inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);202        }203 204        ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);205        cb(ffn_inp, "ffn_inp", il);206 207        // dual ffn normalization (pre)208        cur = build_norm(ffn_inp,209                model.layers[il].ffn_norm, NULL,210                LLM_NORM_RMS, il);211        cb(cur, "ffn_norm", il);212 213        // MoE or dense FFN214        if ((uint32_t)il >= hparams.n_layer_dense_lead) {215            // MoE layer with sigmoid routing, normalization, and scaling216            ggml_tensor * moe_out = build_moe_ffn(cur,217                    model.layers[il].ffn_gate_inp,218                    model.layers[il].ffn_up_exps,219                    model.layers[il].ffn_gate_exps,220                    model.layers[il].ffn_down_exps,221                    model.layers[il].ffn_exp_probs_b,222                    n_expert, n_expert_used,223                    LLM_FFN_SILU,224                    hparams.expert_weights_norm,           // norm_w (route_norm=True)225                    hparams.expert_weights_scale,          // w_scale (route_scale=2.826)226                    (llama_expert_gating_func_type) hparams.expert_gating_func,227                    il);228            cb(moe_out, "ffn_moe_out", il);229 230            // shared expert231            if (hparams.n_expert_shared > 0) {232                ggml_tensor * ffn_shexp = build_ffn(cur,233                        model.layers[il].ffn_up_shexp,   NULL, NULL,234                        model.layers[il].ffn_gate_shexp, NULL, NULL,235                        model.layers[il].ffn_down_shexp, NULL, NULL,236                        NULL,237                        LLM_FFN_SILU, LLM_FFN_PAR, il);238                cb(ffn_shexp, "ffn_shexp", il);239 240                cur = ggml_add(ctx0, moe_out, ffn_shexp);241                cb(cur, "ffn_out", il);242            } else {243                cur = moe_out;244            }245        } else {246            // dense layer247            cur = build_ffn(cur,248                    model.layers[il].ffn_up,   NULL, NULL,249                    model.layers[il].ffn_gate, NULL, NULL,250                    model.layers[il].ffn_down, NULL, NULL,251                    NULL,252                    LLM_FFN_SILU, LLM_FFN_PAR, il);253            cb(cur, "ffn_out", il);254        }255 256        // dual ffn normalization (post)257        cur = build_norm(cur,258                model.layers[il].ffn_post_norm, NULL,259                LLM_NORM_RMS, il);260        cb(cur, "ffn_post_norm", il);261 262        cur = ggml_add(ctx0, cur, ffn_inp);263        cur = build_cvec(cur, il);264        cb(cur, "l_out", il);265 266        // input for next layer267        inpL = cur;268    }269 270    cur = inpL;271 272    cur = build_norm(cur,273            model.output_norm, NULL,274            LLM_NORM_RMS, -1);275    cb(cur, "result_norm", -1);276 277    res->t_embd = cur;278 279    // lm_head280    cur = build_lora_mm(model.output, cur, model.output_s);281    cb(cur, "result_output", -1);282    res->t_logits = cur;283 284    ggml_build_forward_expand(gf, cur);285}286