llama.cpp/src/models/gemma3.cpp

#include "models.h"

void llama_model_gemma3::load_arch_hparams(llama_model_loader & ml) {
    const bool found_swa = ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa, false);
    if (found_swa && hparams.n_swa > 0) {
        hparams.swa_type = LLAMA_SWA_TYPE_STANDARD;
        uint32_t swa_period = 6;
        ml.get_key_or_arr(LLM_KV_ATTENTION_SLIDING_WINDOW_PATTERN, swa_period, false);
        hparams.set_swa_pattern(swa_period);

        ml.get_key(LLM_KV_ROPE_FREQ_BASE_SWA, hparams.rope_freq_base_train_swa, false);
    } else {
        hparams.swa_type = LLAMA_SWA_TYPE_NONE;
    }

    hparams.f_final_logit_softcapping = 0.0f;
    ml.get_key(LLM_KV_FINAL_LOGIT_SOFTCAPPING, hparams.f_final_logit_softcapping, false);
    ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);

    switch (hparams.n_layer()) {
        case 18: type = LLM_TYPE_270M; break;
        case 26: type = LLM_TYPE_1B; break;
        case 32: type = LLM_TYPE_8B; break; // Rnj-1
        case 34: type = LLM_TYPE_4B; break;
        case 48: type = LLM_TYPE_12B; break;
        case 62: type = LLM_TYPE_27B; break;
        default: type = LLM_TYPE_UNKNOWN;
    }

    // ref: https://github.com/google/gemma_pytorch/blob/014acb7ac4563a5f77c76d7ff98f31b568c16508/gemma/config.py#L289
    hparams.f_attention_scale = type == LLM_TYPE_27B
        ? 1.0f / std::sqrt(float(hparams.n_embd / hparams.n_head(0)))
        : 1.0f / std::sqrt(float(hparams.n_embd_head_k()));
}

void llama_model_gemma3::load_arch_tensors(llama_model_loader &) {
    LLAMA_LOAD_LOCALS;

    tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);

    // output
    output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
    output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED);

    // if output is NULL, init from the input tok embed
    if (output == NULL) {
        output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD,   "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED);
    }

    // Dense linear weights
    dense_2_out_layers = create_tensor(tn(LLM_TENSOR_DENSE_2_OUT, "weight"), {n_embd, hparams.dense_2_feat_out}, TENSOR_NOT_REQUIRED);
    dense_3_out_layers = create_tensor(tn(LLM_TENSOR_DENSE_3_OUT, "weight"), {hparams.dense_3_feat_in, n_embd}, TENSOR_NOT_REQUIRED);


    for (int i = 0; i < n_layer; ++i) {
        auto & layer = layers[i];

        layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);

        create_tensor_qkv(layer, i, n_embd, n_embd_head_k * n_head, n_embd_k_gqa, n_embd_v_gqa, 0);
        layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);

        layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), {n_embd}, 0);
        layer.attn_k_norm    = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM,    "weight", i), {n_embd_head_k}, 0);
        layer.attn_q_norm    = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM,    "weight", i), {n_embd_head_k}, 0);

        layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
        layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd,   n_ff}, 0);
        layer.ffn_up   = create_tensor(tn(LLM_TENSOR_FFN_UP,   "weight", i), {n_embd,   n_ff}, 0);
        layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), {  n_ff, n_embd}, 0);
        layer.ffn_post_norm = create_tensor(tn(LLM_TENSOR_FFN_POST_NORM, "weight", i), {n_embd}, 0);
    }
}

std::unique_ptr<llm_graph_context> llama_model_gemma3::build_arch_graph(const llm_graph_params & params) const {
    if (hparams.swa_type == LLAMA_SWA_TYPE_STANDARD) {
        return std::make_unique<graph<true>>(*this, params);
    } else {
        return std::make_unique<graph<false>>(*this, params);
    }
}

template <bool iswa>
llama_model_gemma3::graph<iswa>::graph(const llama_model & model, const llm_graph_params & params) : llm_graph_context(params) {
    const int64_t n_embd_head = hparams.n_embd_head_k();

    ggml_tensor * cur;
    ggml_tensor * inpL;

    inpL = build_inp_embd(model.tok_embd);

    // important: do not normalize weights for raw embeddings input (i.e. encoded image embeddings)
    inpL = ggml_scale(ctx0, inpL, ubatch.token ? sqrtf(n_embd) : 1.0f);
    cb(inpL, "inp_scaled", -1);

    // inp_pos - contains the positions
    ggml_tensor * inp_pos = build_inp_pos();

    // TODO: is causal == true correct? might need some changes
    using inp_attn_type = std::conditional_t<iswa, llm_graph_input_attn_kv_iswa, llm_graph_input_attn_kv>;
    inp_attn_type * inp_attn = nullptr;

    if constexpr (iswa) {
        inp_attn = build_attn_inp_kv_iswa();
    } else {
        inp_attn = build_attn_inp_kv();
    }

    ggml_tensor * inp_out_ids = build_inp_out_ids();

    for (int il = 0; il < n_layer; ++il) {
        float freq_base_l  = 0.0f;
        float freq_scale_l = 0.0f;

        if constexpr (iswa) {
            freq_base_l  = model.get_rope_freq_base (cparams, il);
            freq_scale_l = model.get_rope_freq_scale(cparams, il);
        } else {
            freq_base_l  = freq_base;
            freq_scale_l = freq_scale;
        }

        // norm
        cur = build_norm(inpL, model.layers[il].attn_norm, NULL, LLM_NORM_RMS, il);
        cb(cur, "attn_norm", il);

        // self-attention
        {
            // compute Q and K and RoPE them
            auto [Qcur, Kcur, Vcur] = build_qkv(model.layers[il], cur,
                    n_embd_head, n_head, n_head_kv, il);

            Qcur = build_norm(Qcur, model.layers[il].attn_q_norm, NULL, LLM_NORM_RMS, il);
            cb(Qcur, "Qcur_normed", il);

            Qcur = ggml_rope_ext(
                    ctx0, Qcur, inp_pos, nullptr,
                    n_rot, rope_type, n_ctx_orig, freq_base_l, freq_scale_l,
                    ext_factor, attn_factor, beta_fast, beta_slow);

            Kcur = build_norm(Kcur, model.layers[il].attn_k_norm, NULL, LLM_NORM_RMS, il);
            cb(Kcur, "Kcur_normed", il);

            Kcur = ggml_rope_ext(
                    ctx0, Kcur, inp_pos, nullptr,
                    n_rot, rope_type, n_ctx_orig, freq_base_l, freq_scale_l,
                    ext_factor, attn_factor, beta_fast, beta_slow);

            cb(Qcur, "Qcur", il);
            cb(Kcur, "Kcur", il);
            cb(Vcur, "Vcur", il);

            // ref: https://github.com/google/gemma_pytorch/blob/014acb7ac4563a5f77c76d7ff98f31b568c16508/gemma/model.py#L315
            Qcur = ggml_scale(ctx0, Qcur, hparams.f_attention_scale);

            cur = build_attn(inp_attn,
                    model.layers[il].wo, NULL, model.layers[il].wo_s,
                    Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, 1.0f, il);
        }
        if (il == n_layer - 1 && inp_out_ids) {
            cur  = ggml_get_rows(ctx0,  cur, inp_out_ids);
            inpL = ggml_get_rows(ctx0, inpL, inp_out_ids);
        }
        cur = build_norm(cur,
                model.layers[il].attn_post_norm, NULL,
                LLM_NORM_RMS, il);
        cb(cur, "attn_post_norm", il);

        ggml_tensor * sa_out = ggml_add(ctx0, cur, inpL);
        cb(sa_out, "sa_out", il);

        cur = build_norm(sa_out,
                model.layers[il].ffn_norm, NULL,
                LLM_NORM_RMS, il);
        cb(cur, "ffn_norm", il);

        // feed-forward network
        {
            cur = build_ffn(cur,
                    model.layers[il].ffn_up,   NULL, NULL,
                    model.layers[il].ffn_gate, NULL, NULL,
                    model.layers[il].ffn_down, NULL, NULL,
                    NULL,
                    LLM_FFN_GELU, LLM_FFN_PAR, il);
            cb(cur, "ffn_out", il);
        }
        cur = build_norm(cur,
                model.layers[il].ffn_post_norm, NULL,
                LLM_NORM_RMS, -1);
        cb(cur, "ffn_post_norm", il);

        cur = ggml_add(ctx0, cur, sa_out);

        cur = build_cvec(cur, il);
        cb(cur, "l_out", il);

        // input for next layer
        inpL = cur;
    }
    cur = inpL;

    cur = build_norm(cur,
            model.output_norm, NULL,
            LLM_NORM_RMS, -1);

    cb(cur, "result_norm", -1);
    res->t_embd = cur;

    // lm_head
    cur = build_lora_mm(model.output, cur, model.output_s);

    if (hparams.f_final_logit_softcapping) {
        cur = ggml_scale(ctx0, cur, 1.0f / hparams.f_final_logit_softcapping);
        cur = ggml_tanh(ctx0, cur);
        cur = ggml_scale(ctx0, cur, hparams.f_final_logit_softcapping);
    }

    cb(cur, "result_output", -1);
    res->t_logits = cur;

    ggml_build_forward_expand(gf, cur);
}

template struct llama_model_gemma3::graph<false>;
template struct llama_model_gemma3::graph<true>;