whispercpp 1.3.6 → 1.3.7

data/ext/sources/examples/talk-llama/models/delta-net-base.cpp

@@ -1,445 +0,0 @@
-#include "models.h"
-
-#include "llama-impl.h"
-
-// utility to get one slice from the third dimension
-// input dim:  [x, y, c, b]
-// output dim: [x, y, 1, b]
-static ggml_tensor * get_slice_2d(ggml_context * ctx0, ggml_tensor * t, int64_t c) {
-    return ggml_view_4d(ctx0, t, t->ne[0], t->ne[1], 1, t->ne[3],
-        t->nb[1], t->nb[2], t->nb[3], t->nb[2] * c);
-}
-
-llm_build_delta_net_base::llm_build_delta_net_base(const llm_graph_params & params) : llm_graph_context(params) {}
-
-std::pair<ggml_tensor *, ggml_tensor *> llm_build_delta_net_base::build_delta_net_chunking(
-        ggml_tensor * q,
-        ggml_tensor * k,
-        ggml_tensor * v,
-        ggml_tensor * g,
-        ggml_tensor * b,
-        ggml_tensor * s,
-        int           il) {
-    const int64_t S_k      = q->ne[0];
-    const int64_t H_k      = q->ne[1];
-    const int64_t n_tokens = q->ne[2];
-    const int64_t n_seqs   = q->ne[3];
-
-    const int64_t S_v = v->ne[0];
-    const int64_t H_v = v->ne[1];
-    const bool kda = (g->ne[0] == S_k && g->ne[1] == H_k);
-
-    GGML_ASSERT(S_k == S_v);
-    GGML_ASSERT(H_v % H_k == 0);
-
-    GGML_ASSERT(q->ne[0] == S_k && q->ne[1] == H_k && q->ne[2] == n_tokens && q->ne[3] == n_seqs);
-    GGML_ASSERT(k->ne[0] == S_k && k->ne[1] == H_k && k->ne[2] == n_tokens && k->ne[3] == n_seqs);
-    GGML_ASSERT(v->ne[0] == S_v && v->ne[1] == H_v && v->ne[2] == n_tokens && v->ne[3] == n_seqs);
-
-    GGML_ASSERT(g->ne[0] == 1   || g->ne[0] == S_v);
-    GGML_ASSERT(                   g->ne[1] == H_v && g->ne[2] == n_tokens && g->ne[3] == n_seqs);
-    GGML_ASSERT(b->ne[0] == 1   && b->ne[1] == H_v && b->ne[2] == n_tokens && b->ne[3] == n_seqs);
-    GGML_ASSERT(s->ne[0] == S_v && s->ne[1] == S_v && s->ne[2] == H_v      && s->ne[3] == n_seqs);
-
-    const float scale = 1.0f / sqrtf(S_k);
-
-    q = ggml_scale(ctx0, q, scale);
-
-    cb(q, "q_in", il);
-    cb(k, "k_in", il);
-    cb(v, "v_in", il);
-    cb(b, "b_in", il);
-    cb(g, "g_in", il);
-
-    q = ggml_permute(ctx0, q, 0, 2, 1, 3); // [S_k, n_tokens, H_k, n_seqs]
-    k = ggml_permute(ctx0, k, 0, 2, 1, 3); // [S_k, n_tokens, H_k, n_seqs]
-    v = ggml_permute(ctx0, v, 0, 2, 1, 3); // [S_v, n_tokens, H_v, n_seqs]
-    g = ggml_permute(ctx0, g, 0, 2, 1, 3); // [g_0, n_tokens, H_v, n_seqs]
-    b = ggml_permute(ctx0, b, 0, 2, 1, 3); // [  1, n_tokens, H_v, n_seqs]
-
-    const int CS = kda ? 16 : 64; // chunk size
-
-    const int pad = (CS - n_tokens % CS) % CS;
-    const int n_chunks = (n_tokens + pad) / CS;
-
-    q = ggml_pad(ctx0, q, 0, pad, 0, 0);
-    k = ggml_pad(ctx0, k, 0, pad, 0, 0);
-    v = ggml_pad(ctx0, v, 0, pad, 0, 0);
-    g = ggml_pad(ctx0, g, 0, pad, 0, 0);
-    b = ggml_pad(ctx0, b, 0, pad, 0, 0);
-
-    ggml_tensor * v_b = ggml_mul(ctx0, v, b);
-    ggml_tensor * k_b = ggml_mul(ctx0, k, b);
-
-    cb(v_b, "v_b", il);
-    cb(k_b, "k_b", il);
-
-    q   = ggml_reshape_4d(ctx0, q,   S_k, CS, n_chunks, H_k * n_seqs);
-    k   = ggml_reshape_4d(ctx0, k,   S_k, CS, n_chunks, H_k * n_seqs);
-    k_b = ggml_reshape_4d(ctx0, k_b, S_k, CS, n_chunks, H_v * n_seqs);
-    v   = ggml_reshape_4d(ctx0, v,   S_v, CS, n_chunks, H_v * n_seqs);
-    v_b = ggml_reshape_4d(ctx0, v_b, S_v, CS, n_chunks, H_v * n_seqs);
-
-    g = ggml_reshape_4d(ctx0, g, g->ne[0], CS, n_chunks, H_v * n_seqs);
-    b = ggml_reshape_4d(ctx0, b, 1,        CS, n_chunks, H_v * n_seqs);
-
-    // [CS, g_0, n_chunks, H_v * n_seqs]
-    // TODO: extend ggml_cumsum with axis parameter to avoid transpose
-    ggml_tensor * g_cs = ggml_cumsum(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, g)));
-    cb(g_cs, "g_cs", il);
-
-    ggml_tensor * kb = nullptr;
-    ggml_tensor * kq = nullptr;
-    if (kda) {
-        const int64_t CHB = n_chunks * H_k * n_seqs;
-
-        ggml_tensor * g_cs_i = ggml_reshape_4d(ctx0, g_cs, CS, 1, S_k, CHB);  // [chunk_size, 1, S_k, CHB]
-        ggml_tensor * g_cs_j = ggml_reshape_4d(ctx0, g_cs, 1, CS, S_k, CHB);  // [1, chunk_size, S_k, CHB]
-
-        g_cs_j = ggml_repeat_4d(ctx0, g_cs_j, CS, CS, S_k, CHB);  // [1, chunk_size, S_k, CHB] -> [chunk_size, chunk_size, S_k, CHB]
-
-        // decay_mask [chunk_size,chunk_size,S_k,CHB]
-        ggml_tensor * decay_mask;
-        decay_mask = ggml_sub(ctx0, g_cs_j, g_cs_i);
-        decay_mask = ggml_tri(ctx0, decay_mask, GGML_TRI_TYPE_LOWER_DIAG);
-        decay_mask = ggml_exp(ctx0, decay_mask);
-        cb(decay_mask, "decay_mask", il);
-
-        // decay_mask [S_k,BT_j,BT_i,CHB] *Note* second and third chunk_sizes are switched
-        decay_mask = ggml_cont_4d(ctx0, ggml_permute(ctx0, decay_mask, 2, 1, 0, 3), S_k, CS, CS, CHB);
-
-        ggml_tensor * k_b_i = ggml_reshape_4d(ctx0, k_b, S_k, CS,  1, CHB);
-        ggml_tensor * k_j   = ggml_reshape_4d(ctx0, k,   S_k,  1, CS, CHB);
-        ggml_tensor * q_i   = ggml_reshape_4d(ctx0, q,   S_k, CS,  1, CHB);
-
-        ggml_tensor * decay_k_b_i = ggml_mul(ctx0, decay_mask, k_b_i);
-        ggml_tensor * decay_q_i   = ggml_mul(ctx0, decay_mask, q_i);
-
-        // decay_k_b_i [S,BT,BT,CHB] @ k_j [S,1,BT,CHB] = Akk [BT,1,BT,CHB]
-        kb = ggml_mul_mat(ctx0, decay_k_b_i, k_j);
-        kq = ggml_mul_mat(ctx0, decay_q_i,   k_j);
-
-        kb = ggml_cont(ctx0, ggml_transpose(ctx0, ggml_reshape_4d(ctx0, kb, CS, CS, n_chunks, H_v * n_seqs)));
-        kq = ggml_cont(ctx0, ggml_transpose(ctx0, ggml_reshape_4d(ctx0, kq, CS, CS, n_chunks, H_v * n_seqs)));
-    } else {
-        ggml_tensor * g_cs_i = g_cs;
-        ggml_tensor * g_cs_j = ggml_reshape_4d(ctx0, g_cs, 1, CS, n_chunks, H_v * n_seqs);
-
-        g_cs_j = ggml_repeat_4d(ctx0, g_cs_j, CS, CS, n_chunks, H_v * n_seqs);
-
-        // [CS, CS, n_chunks, H_v * n_seqs]
-        ggml_tensor * decay_mask;
-        decay_mask = ggml_sub(ctx0, g_cs_j, g_cs_i);
-        decay_mask = ggml_tri(ctx0, decay_mask, GGML_TRI_TYPE_LOWER_DIAG);
-        decay_mask = ggml_exp(ctx0, decay_mask);
-        cb(decay_mask, "decay_mask", il);
-
-        // [CS, CS, n_chunks, H_k * n_seqs]
-        kb = ggml_mul_mat(ctx0, k,  k_b);
-        kb = ggml_mul    (ctx0, kb, decay_mask);
-
-        // [CS, CS, n_chunks, H_k * n_seqs]
-        kq = ggml_mul_mat(ctx0, k, q);
-        kq = ggml_mul(ctx0, kq, decay_mask);
-    }
-
-    kq = ggml_tri(ctx0, kq, GGML_TRI_TYPE_LOWER_DIAG);
-    cb(kq, "kq", il);
-
-    // [CS, CS, n_chunks, H_k * n_seqs]
-    ggml_tensor * attn;
-    attn = ggml_tri(ctx0, kb, GGML_TRI_TYPE_LOWER);
-    cb(attn, "attn", il);
-
-    ggml_tensor * identity;
-    identity = ggml_view_1d(ctx0, attn, CS, 0);
-    identity = ggml_fill   (ctx0, identity, 1.0f);
-    identity = ggml_diag   (ctx0, identity);
-
-    ggml_tensor * lhs = ggml_add(ctx0, attn, identity);
-    cb(lhs, "dnet_add_ch_lhs", il);
-
-    attn = ggml_neg(ctx0, attn);
-    cb(attn, "attn_pre_solve", il);
-
-    ggml_tensor * lin_solve = ggml_solve_tri(ctx0, lhs, attn, true, true, false);
-    attn = ggml_add(ctx0, lin_solve, identity);
-    cb(attn, "dnet_add_ch_attn_solved", il); // [CS, CS, n_chunks, H_k * n_seqs]
-
-    // [S_v, CS, n_chunks, H_v * n_seqs]
-    v = ggml_mul_mat(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, v_b)), attn);
-
-    // [CS, 1, n_chunks, H_v * n_seqs] KDA: [CS, S_k, n_chunks, H_v * n_seqs]
-    ggml_tensor * g_exp = ggml_exp(ctx0, g_cs);
-
-    k_b = ggml_cont(ctx0, ggml_transpose(ctx0, k_b));
-
-    // [CS, S_k, n_chunks, H_k * n_seqs]
-    ggml_tensor * kbg = ggml_mul(ctx0, k_b, g_exp);
-    cb(kbg, "k_beta_g_exp", il);
-
-    // [S_k, CS, n_chunks, H_k * n_seqs]
-    ggml_tensor * k_cd = ggml_mul_mat(ctx0, kbg, attn);
-    cb(k_cd, "k_cumdecay", il);
-
-    // [1, CS, n_chunks, H_k * n_seqs] KDA: [S_k, CS, n_chunks, H_k * n_seqs]
-    ggml_tensor * g_exp_t = ggml_cont(ctx0, ggml_transpose(ctx0, g_exp));
-    ggml_tensor * q_g_exp = ggml_mul(ctx0, q, g_exp_t);
-
-    // vectorized calculation of key_gdiff
-    // improved from the chunked version:
-    //   g_last = torch.clamp(g_cum[:, :, -1], max=50.0).exp().unsqueeze(-1).unsqueeze(-1)
-    //   g_diff = torch.clamp(g_cum[:, :, -1:] - g_cum, max=50.0).exp()
-    //   key_gdiff = key * g_diff.unsqueeze(-1)
-    //   kgdmulvnew = (key_gdiff).transpose(-1, -2) @ v_new
-    //   last_recurrent_state = last_recurrent_state * g_last + kgdmulvnew
-
-    // get last element in g_cumsum along CS dimension (ne0)
-    // example: [[x, y, z, ..., last], ...] -> [[last], ...]
-    // [1, 1, n_chunks, H_v * n_seqs] KDA: [1, S_k, n_chunks, H_v * n_seqs]
-    ggml_tensor * g_last = ggml_view_4d(ctx0, g_cs, 1, g_cs->ne[1], g_cs->ne[2], g_cs->ne[3],
-            g_cs->nb[1],
-            g_cs->nb[2],
-            g_cs->nb[3],
-            ggml_row_size(g_cs->type, g_cs->ne[0] - 1));
-    cb(g_last, "g_last", il);
-
-    // TODO: remove this cont when CUDA supports non-cont unary ops
-    g_last = ggml_cont(ctx0, g_last);
-
-    // [1, 1, n_chunks, H_v * n_seqs] KDA: [S_k, 1, n_chunks, H_v * n_seqs]
-    ggml_tensor * g_last_exp_t = ggml_transpose(ctx0, ggml_exp(ctx0, g_last));
-    cb(g_last_exp_t, "g_last_exp_t", il);
-
-    // [CS, 1, n_chunks, H_v * n_seqs] KDA: [CS, S_k, n_chunks, H_v * n_seqs]
-    ggml_tensor * g_diff = ggml_neg(ctx0, ggml_sub(ctx0, g_cs, g_last));
-    cb(g_diff, "g_diff", il);
-
-    ggml_tensor * g_diff_exp_t = ggml_cont(ctx0, ggml_transpose(ctx0, ggml_exp(ctx0, g_diff)));
-
-    // [S_k, CS, n_chunks, H_v * n_seqs]
-    ggml_tensor * kg = ggml_mul(ctx0, k, g_diff_exp_t);
-    cb(kg, "key_gdiff", il);
-
-    // [CS, S_k, n_chunks, H_v * n_seqs]
-    ggml_tensor * kg_t = ggml_cont(ctx0, ggml_transpose(ctx0, kg));
-    cb(kg_t, "key_gdiff_t", il);
-
-    s = ggml_reshape_4d(ctx0, s, S_v, S_v, 1, H_v * n_seqs);
-    cb(s, "dnet_add_ch_state", il);
-
-    // [CS, S_v, n_chunks, H_v * n_seqs]
-    ggml_tensor * v_t = ggml_cont(ctx0, ggml_transpose(ctx0, v));
-
-    for (int64_t chunk = 0; chunk < n_chunks; chunk++) {
-        ggml_tensor * ch_k_cd    = get_slice_2d(ctx0, k_cd,    chunk); // [S_k,  CS, 1, H_k * n_seqs]
-        ggml_tensor * ch_v_t     = get_slice_2d(ctx0, v_t,     chunk); // [ CS, S_v, 1, H_v * n_seqs]
-        ggml_tensor * ch_kq      = get_slice_2d(ctx0, kq,      chunk); // [ CS,  CS, 1, H_k * n_seqs]
-        ggml_tensor * ch_q_g_exp = get_slice_2d(ctx0, q_g_exp, chunk); // [S_k,  CS, 1, H_k * n_seqs]
-        ggml_tensor * ch_kg_t    = get_slice_2d(ctx0, kg_t,    chunk); // [ CS, S_k, 1, H_v * n_seqs]
-
-        // [CS, S_v, 1, H_v * n_seqs]
-        ggml_tensor * v_t_p = ggml_mul_mat(ctx0, ch_k_cd, s);
-        cb(v_t_p, "v_prime", il);
-
-        // [CS, S_v, 1, H_v * n_seqs]
-        ggml_tensor * v_t_new = ggml_sub(ctx0, ch_v_t, v_t_p);
-        cb(v_t_new, "v_t_new", il);
-
-        // [S_v, CS, 1, H_v * n_seqs]
-        ggml_tensor * v_attn = ggml_mul_mat(ctx0, v_t_new, ch_kq);
-        cb(v_attn, "v_attn", il);
-
-        // [S_v, CS, 1, H_v * n_seqs]
-        ggml_tensor * attn_inter = ggml_mul_mat(ctx0, s, ch_q_g_exp);
-        cb(attn_inter, "attn_inter", il);
-
-        // [S_v, CS, 1, H_v * n_seqs]
-        ggml_tensor * o_ch = ggml_add(ctx0, attn_inter, v_attn);
-        cb(o_ch, "dnet_add_ch_attn_out", il);
-
-        v = ggml_set_inplace(ctx0, v, o_ch, v->nb[1], v->nb[2], v->nb[3], chunk * v->nb[2]);
-
-        // kgdmulvnew = (key_gdiff).transpose(-1, -2) @ v_new
-        // TODO: head broadcast might not work here - probably will need a transpose
-        ggml_tensor * kgv = ggml_mul_mat(ctx0, ch_kg_t, v_t_new); // [S_k, S_v, 1, H_k * n_seqs]
-
-        // last_recurrent_state = last_recurrent_state * g_last + kgdmulvnew
-        ggml_tensor * ch_g_last_exp_t = get_slice_2d(ctx0, g_last_exp_t, chunk);
-
-        s = ggml_mul(ctx0, s, ch_g_last_exp_t);
-        s = ggml_add(ctx0, s, kgv);
-        cb(s, "dnet_add_ch_state", il);
-    }
-
-    // truncate padded tokens
-    ggml_tensor * o = ggml_view_4d(ctx0, v,
-            S_v, n_tokens, H_v, n_seqs,
-            ggml_row_size(v->type, S_v),
-            ggml_row_size(v->type, S_v * CS * n_chunks),
-            ggml_row_size(v->type, S_v * CS * n_chunks * H_v), 0);
-    o = ggml_permute  (ctx0, o, 0, 2, 1, 3); // [S_v, H_v, n_tokens, n_seqs]
-    s = ggml_reshape_4d(ctx0, s, S_v, S_v, H_v, n_seqs);
-    cb(s, "output_state", il);
-
-    return {o, s};
-}
-
-std::pair<ggml_tensor *, ggml_tensor *> llm_build_delta_net_base::build_delta_net_autoregressive(
-        ggml_tensor * q,
-        ggml_tensor * k,
-        ggml_tensor * v,
-        ggml_tensor * g,
-        ggml_tensor * b, // beta
-        ggml_tensor * s, // state
-        int           il) {
-    const int64_t S_k      = q->ne[0];
-    const int64_t H_k      = q->ne[1];
-    const int64_t n_tokens = q->ne[2];
-    const int64_t n_seqs   = q->ne[3];
-
-    const int64_t S_v = v->ne[0];
-    const int64_t H_v = v->ne[1];
-
-    GGML_ASSERT(n_tokens == 1);
-
-    GGML_ASSERT(S_k == S_v);
-    GGML_ASSERT(H_v % H_k == 0);
-
-    GGML_ASSERT(q->ne[0] == S_k && q->ne[1] == H_k && q->ne[2] == n_tokens && q->ne[3] == n_seqs);
-    GGML_ASSERT(k->ne[0] == S_k && k->ne[1] == H_k && k->ne[2] == n_tokens && k->ne[3] == n_seqs);
-    GGML_ASSERT(v->ne[0] == S_v && v->ne[1] == H_v && v->ne[2] == n_tokens && v->ne[3] == n_seqs);
-
-    GGML_ASSERT(g->ne[0] == 1   || g->ne[0] == S_v);
-    GGML_ASSERT(                   g->ne[1] == H_v && g->ne[2] == n_tokens && g->ne[3] == n_seqs);
-    GGML_ASSERT(b->ne[0] == 1   && b->ne[1] == H_v && b->ne[2] == n_tokens && b->ne[3] == n_seqs);
-    GGML_ASSERT(s->ne[0] == S_v && s->ne[1] == S_v && s->ne[2] == H_v      && s->ne[3] == n_seqs);
-
-    const float scale = 1.0f / sqrtf(S_k);
-
-    q = ggml_scale(ctx0, q, scale);
-
-    q = ggml_permute(ctx0, q, 0, 2, 1, 3); // [S_k, n_tokens, H_k, n_seqs]
-    k = ggml_permute(ctx0, k, 0, 2, 1, 3); // [S_k, n_tokens, H_k, n_seqs]
-    v = ggml_permute(ctx0, v, 0, 2, 1, 3); // [S_v, n_tokens, H_v, n_seqs]
-
-    cb(q, "q_in", il);
-    cb(k, "k_in", il);
-    cb(v, "v_in", il);
-    cb(b, "b_in", il);
-    cb(g, "g_in", il);
-
-    // GDA: [1,  1,  H_v, n_seqs]
-    // KDA: [1, S_k, H_v, n_seqs]
-    g = ggml_reshape_4d(ctx0, g, 1, g->ne[0], H_v, n_seqs);
-    b = ggml_reshape_4d(ctx0, b, 1,        1, H_v, n_seqs);
-
-    // [S_v, S_v, H_v, n_seqs]
-    g = ggml_exp(ctx0, g);
-    s = ggml_mul(ctx0, s, g);
-
-    // [1, S_v, H_v, n_seqs]
-    ggml_tensor * sk;
-    sk = ggml_mul     (ctx0, s, k);
-    sk = ggml_sum_rows(ctx0, sk);
-
-    // [S_v, 1, H_v, n_seqs]
-    ggml_tensor * d;
-    d = ggml_sub(ctx0, v, ggml_transpose(ctx0, sk));
-    d = ggml_mul(ctx0, d, b);
-
-    // [1, S_v, H_v, n_seqs]
-    ggml_tensor * d_t;
-    d_t = ggml_transpose(ctx0, d);
-
-    // [S_v, S_v, H_v, n_seqs]
-    ggml_tensor * kd;
-    k  = ggml_repeat(ctx0, k, s);
-    kd = ggml_mul   (ctx0, k, d_t);
-
-    s = ggml_add(ctx0, s, kd);
-
-    cb(s, "dnet_add_ar_state", il);
-
-    ggml_tensor * s_q = ggml_mul     (ctx0, s, q);
-    ggml_tensor * o   = ggml_sum_rows(ctx0, s_q);
-
-    o = ggml_permute  (ctx0, o, 2, 0, 1, 3); // [S_v, H_v, n_tokens, n_seqs]
-
-    return {o, s};
-}
-
-std::pair<ggml_tensor *, ggml_tensor *> llm_build_delta_net_base::build_delta_net_fused(
-        ggml_tensor * q,
-        ggml_tensor * k,
-        ggml_tensor * v,
-        ggml_tensor * g,
-        ggml_tensor * b,
-        ggml_tensor * s,
-        int           il) {
-    const int64_t S_k      = q->ne[0];
-    const int64_t H_k      = q->ne[1];
-    const int64_t n_tokens = q->ne[2];
-    const int64_t n_seqs   = q->ne[3];
-
-    const int64_t S_v = v->ne[0];
-    const int64_t H_v = v->ne[1];
-
-    GGML_ASSERT(S_k == S_v);
-    GGML_ASSERT(H_v % H_k == 0);
-
-    GGML_ASSERT(q->ne[0] == S_k && q->ne[1] == H_k && q->ne[2] == n_tokens && q->ne[3] == n_seqs);
-    GGML_ASSERT(k->ne[0] == S_k && k->ne[1] == H_k && k->ne[2] == n_tokens && k->ne[3] == n_seqs);
-    GGML_ASSERT(v->ne[0] == S_v && v->ne[1] == H_v && v->ne[2] == n_tokens && v->ne[3] == n_seqs);
-
-    GGML_ASSERT(g->ne[0] == 1   || g->ne[0] == S_v);
-    GGML_ASSERT(                   g->ne[1] == H_v && g->ne[2] == n_tokens && g->ne[3] == n_seqs);
-    GGML_ASSERT(b->ne[0] == 1   && b->ne[1] == H_v && b->ne[2] == n_tokens && b->ne[3] == n_seqs);
-    GGML_ASSERT(s->ne[0] == S_v && s->ne[1] == S_v && s->ne[2] == H_v      && s->ne[3] == n_seqs);
-
-    ggml_tensor * result = ggml_gated_delta_net(ctx0, q, k, v, g, b, s);
-    if (n_tokens == 1) {
-        cb(result, LLAMA_TENSOR_NAME_FGDN_AR, il);
-    } else {
-        cb(result, LLAMA_TENSOR_NAME_FGDN_CH, il);
-    }
-
-    ggml_tensor * output = ggml_view_4d(ctx0, result,
-            S_v, H_v, n_tokens, n_seqs,
-            ggml_row_size(result->type, S_v),
-            ggml_row_size(result->type, S_v * H_v),
-            ggml_row_size(result->type, S_v * H_v * n_tokens), 0);
-
-    ggml_tensor * new_state = ggml_view_4d(ctx0, result,
-            S_v, S_v, H_v, n_seqs,
-            ggml_row_size(result->type, S_v),
-            ggml_row_size(result->type, S_v * S_v),
-            ggml_row_size(result->type, S_v * S_v * H_v),
-            ggml_row_size(result->type, S_v * H_v * n_tokens * n_seqs));
-
-    return {output, new_state};
-}
-
-std::pair<ggml_tensor *, ggml_tensor *> llm_build_delta_net_base::build_delta_net(
-        ggml_tensor * q,
-        ggml_tensor * k,
-        ggml_tensor * v,
-        ggml_tensor * g,
-        ggml_tensor * b,
-        ggml_tensor * s,
-        int           il) {
-    const int64_t n_seq_tokens = q->ne[2];
-
-    if (n_seq_tokens == 1) {
-        if (cparams.fused_gdn_ar) {
-            return build_delta_net_fused(q, k, v, g, b, s, il);
-        }
-        return build_delta_net_autoregressive(q, k, v, g, b, s, il);
-    }
-
-    if (cparams.fused_gdn_ch) {
-        return build_delta_net_fused(q, k, v, g, b, s, il);
-    }
-
-    return build_delta_net_chunking(q, k, v, g, b, s, il);
-}

data/ext/sources/examples/talk-llama/models/dots1.cpp

@@ -1,132 +0,0 @@
-#include "models.h"
-
-llm_build_dots1::llm_build_dots1(const llama_model & model, const llm_graph_params & params) :
-    llm_graph_context(params) {
-    const int64_t n_embd_head = hparams.n_embd_head_v();
-
-    GGML_ASSERT(n_embd_head == hparams.n_embd_head_k());
-    GGML_ASSERT(n_embd_head == n_rot);
-
-    ggml_tensor * cur;
-    ggml_tensor * inpL;
-
-    inpL = build_inp_embd(model.tok_embd);
-
-    // inp_pos - contains the positions
-    ggml_tensor * inp_pos = build_inp_pos();
-
-    auto * inp_attn = build_attn_inp_kv();
-
-    ggml_tensor * inp_out_ids = build_inp_out_ids();
-
-    for (int il = 0; il < n_layer; ++il) {
-        ggml_tensor * inpSA = inpL;
-
-        // 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
-            ggml_tensor * Qcur = build_lora_mm(model.layers[il].wq, cur);
-            cb(Qcur, "Qcur", il);
-
-            ggml_tensor * Kcur = build_lora_mm(model.layers[il].wk, cur);
-            cb(Kcur, "Kcur", il);
-
-            ggml_tensor * Vcur = build_lora_mm(model.layers[il].wv, cur);
-            cb(Vcur, "Vcur", il);
-
-            Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens);
-            Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
-            Vcur = ggml_reshape_3d(ctx0, Vcur, n_embd_head, n_head_kv, n_tokens);
-
-            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, freq_scale,
-                                 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, freq_scale,
-                                 ext_factor, attn_factor, beta_fast, beta_slow);
-
-            cb(Qcur, "Qcur", il);
-            cb(Kcur, "Kcur", il);
-            cb(Vcur, "Vcur", il);
-
-            cur = build_attn(inp_attn,
-                    model.layers[il].wo, model.layers[il].bo,
-                    Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, 1.0f / sqrtf(float(n_embd_head)), il);
-        }
-        if (il == n_layer - 1 && inp_out_ids) {
-            cur   = ggml_get_rows(ctx0, cur, inp_out_ids);
-            inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
-        }
-        ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
-        cb(ffn_inp, "ffn_inp", il);
-
-        // MoE branch
-        cur = build_norm(ffn_inp, model.layers[il].ffn_norm, NULL, LLM_NORM_RMS, il);
-        cb(cur, "ffn_norm", il);
-
-        if ((uint32_t) il < hparams.n_layer_dense_lead) {
-            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_SILU, LLM_FFN_PAR, il);
-            cb(cur, "ffn_out", il);
-        } else {
-            ggml_tensor * moe_out = build_moe_ffn(cur,
-                model.layers[il].ffn_gate_inp,
-                model.layers[il].ffn_up_exps,
-                model.layers[il].ffn_gate_exps,
-                model.layers[il].ffn_down_exps,
-                model.layers[il].ffn_exp_probs_b,
-                n_expert, n_expert_used,
-                LLM_FFN_SILU, hparams.expert_weights_norm,
-                hparams.expert_weights_scale,
-                (llama_expert_gating_func_type) hparams.expert_gating_func,
-                il);
-            cb(moe_out, "ffn_moe_out", il);
-
-            {
-                ggml_tensor * ffn_shexp =
-                    build_ffn(cur,
-                        model.layers[il].ffn_up_shexp, NULL, NULL,
-                        model.layers[il].ffn_gate_shexp, NULL, NULL,
-                        model.layers[il].ffn_down_shexp, NULL, NULL,
-                        NULL, LLM_FFN_SILU, LLM_FFN_PAR, il);
-                cb(ffn_shexp, "ffn_shexp", il);
-
-                cur = ggml_add(ctx0, moe_out, ffn_shexp);
-                cb(cur, "ffn_out", il);
-            }
-        }
-        cur = ggml_add(ctx0, cur, ffn_inp);
-
-        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);
-
-    cb(cur, "result_output", -1);
-    res->t_logits = cur;
-
-    ggml_build_forward_expand(gf, cur);
-}

data/ext/sources/examples/talk-llama/models/dream.cpp

@@ -1,105 +0,0 @@
-#include "models.h"
-
-
-
-llm_build_dream::llm_build_dream(const llama_model & model, const llm_graph_params & params) :
-    llm_graph_context(params) {
-    //copied from qwen2
-    const int64_t n_embd_head = hparams.n_embd_head_v();
-
-    GGML_ASSERT(n_embd_head == hparams.n_embd_head_k());
-    GGML_ASSERT(n_embd_head == n_rot);
-
-    ggml_tensor * cur;
-    ggml_tensor * inpL;
-
-    inpL = build_inp_embd(model.tok_embd);
-
-    // inp_pos - contains the positions
-    ggml_tensor * inp_pos = build_inp_pos();
-
-    auto * inp_attn = build_attn_inp_no_cache();
-
-    ggml_tensor * inp_out_ids = build_inp_out_ids();
-
-    for (int il = 0; il < n_layer; ++il) {
-        ggml_tensor * inpSA = inpL;
-
-        // 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
-            ggml_tensor * Qcur = build_lora_mm(model.layers[il].wq, cur);
-            Qcur               = ggml_add(ctx0, Qcur, model.layers[il].bq);
-            cb(Qcur, "Qcur", il);
-
-            ggml_tensor * Kcur = build_lora_mm(model.layers[il].wk, cur);
-            Kcur               = ggml_add(ctx0, Kcur, model.layers[il].bk);
-            cb(Kcur, "Kcur", il);
-
-            ggml_tensor * Vcur = build_lora_mm(model.layers[il].wv, cur);
-            Vcur               = ggml_add(ctx0, Vcur, model.layers[il].bv);
-            cb(Vcur, "Vcur", il);
-
-            Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens);
-            Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
-            Vcur = ggml_reshape_3d(ctx0, Vcur, n_embd_head, n_head_kv, n_tokens);
-
-            Qcur = ggml_rope_ext(ctx0, Qcur, inp_pos, nullptr, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
-                                 ext_factor, attn_factor, beta_fast, beta_slow);
-
-            Kcur = ggml_rope_ext(ctx0, Kcur, inp_pos, nullptr, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
-                                 ext_factor, attn_factor, beta_fast, beta_slow);
-
-            cb(Qcur, "Qcur", il);
-            cb(Kcur, "Kcur", il);
-            cb(Vcur, "Vcur", il);
-
-            cur = build_attn(inp_attn,
-                    model.layers[il].wo, model.layers[il].bo,
-                    Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, 1.0f / sqrtf(float(n_embd_head)), il);
-        }
-        if (il == n_layer - 1 && inp_out_ids) {
-            cur   = ggml_get_rows(ctx0, cur, inp_out_ids);
-            inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
-        }
-        ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
-        cb(ffn_inp, "ffn_inp", il);
-
-        // feed-forward network
-        cur = build_norm(ffn_inp, model.layers[il].ffn_norm, NULL, LLM_NORM_RMS, il);
-        cb(cur, "ffn_norm", il);
-
-        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_SILU, LLM_FFN_PAR, il);
-        cb(cur, "ffn_out", il);
-
-        cur = ggml_add(ctx0, cur, ffn_inp);
-
-        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);
-
-    cb(cur, "result_output", -1);
-    res->t_logits = cur;
-
-    ggml_build_forward_expand(gf, cur);
-}