From c4aca655822d08a7e5736dd39653ef84ecbb1bf0 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Thu, 13 Mar 2025 19:26:09 +0200 Subject: [PATCH] hparams : add SWA rope parameters ggml-ci --- src/llama-context.cpp | 14 +++++--------- src/llama-graph.cpp | 4 ++-- src/llama-hparams.cpp | 2 +- src/llama-hparams.h | 4 +++- src/llama-model.cpp | 22 +++++++++++++++------- 5 files changed, 26 insertions(+), 20 deletions(-) diff --git a/src/llama-context.cpp b/src/llama-context.cpp index 89fb33cbc..4df6b18ec 100644 --- a/src/llama-context.cpp +++ b/src/llama-context.cpp @@ -537,16 +537,12 @@ llm_graph_result_ptr llama_context::build_kv_self_shift( const int64_t n_head_kv = hparams.n_head_kv(il); const int64_t n_embd_k_gqa = hparams.n_embd_k_gqa(il); - float freq_base_l = cparams.rope_freq_base; - float freq_scale_l = cparams.rope_freq_scale; + const bool is_swa = hparams.is_swa(il); - // TODO: improve - if (model.arch == LLM_ARCH_GEMMA3) { - const bool is_sliding = hparams.is_sliding(il); - - freq_base_l = is_sliding ? 10000.0f : cparams.rope_freq_base; - freq_scale_l = is_sliding ? 1.0f : cparams.rope_freq_scale; - } + // note: the swa rope params could become part of the cparams in the future + // if we decide to make them configurable, like the non-sliding ones + const float freq_base_l = is_swa ? hparams.rope_freq_base_train_swa : cparams.rope_freq_base; + const float freq_scale_l = is_swa ? hparams.rope_freq_scale_train_swa : cparams.rope_freq_scale; ggml_tensor * rope_factors = kv_self->cbs.get_rope_factors(n_ctx_per_seq(), il); diff --git a/src/llama-graph.cpp b/src/llama-graph.cpp index 4a53e8392..1041ba29f 100644 --- a/src/llama-graph.cpp +++ b/src/llama-graph.cpp @@ -1403,9 +1403,9 @@ ggml_tensor * llm_graph_context::build_attn( ggml_build_forward_expand(gf, ggml_cpy(ctx0, v_cur, v_cache_view)); } - const bool is_sliding = hparams.is_sliding(il); + const bool is_swa = hparams.is_swa(il); - const auto & kq_mask = is_sliding ? inp->get_kq_mask_swa() : inp->get_kq_mask(); + const auto & kq_mask = is_swa ? inp->get_kq_mask_swa() : inp->get_kq_mask(); const auto n_kv = kv_self->n; diff --git a/src/llama-hparams.cpp b/src/llama-hparams.cpp index 58e98bf23..90dfe7a7f 100644 --- a/src/llama-hparams.cpp +++ b/src/llama-hparams.cpp @@ -70,7 +70,7 @@ uint32_t llama_hparams::n_embd_v_s() const { return ssm_d_state * ssm_d_inner; } -bool llama_hparams::is_sliding(uint32_t il) const { +bool llama_hparams::is_swa(uint32_t il) const { if (il < n_layer) { return n_swa > 0 && n_swa_pattern > 0 && il % n_swa_pattern < (n_swa_pattern - 1); } diff --git a/src/llama-hparams.h b/src/llama-hparams.h index e3091c812..dbb7abd31 100644 --- a/src/llama-hparams.h +++ b/src/llama-hparams.h @@ -79,7 +79,9 @@ struct llama_hparams { float rope_attn_factor = 1.0f; float rope_freq_base_train; + float rope_freq_base_train_swa; float rope_freq_scale_train; + float rope_freq_scale_train_swa; uint32_t n_ctx_orig_yarn; float rope_yarn_log_mul; @@ -135,7 +137,7 @@ struct llama_hparams { // dimension of the recurrent state embeddings uint32_t n_embd_v_s() const; - bool is_sliding(uint32_t il) const; + bool is_swa(uint32_t il) const; }; static_assert(std::is_trivially_copyable::value, "llama_hparams must be trivially copyable"); diff --git a/src/llama-model.cpp b/src/llama-model.cpp index 5647d2ad6..cce943df0 100644 --- a/src/llama-model.cpp +++ b/src/llama-model.cpp @@ -475,6 +475,10 @@ void llama_model::load_hparams(llama_model_loader & ml) { } hparams.rope_freq_scale_train = ropescale == 0.0f ? 1.0f : 1.0f/ropescale; + // by default assume that the sliding-window layers use the same scaling type as the non-sliding-window layers + hparams.rope_freq_base_train_swa = hparams.rope_freq_base_train; + hparams.rope_freq_scale_train_swa = hparams.rope_freq_scale_train; + ml.get_key(LLM_KV_ROPE_SCALING_ATTN_FACTOR, hparams.rope_attn_factor, false); // non-transformer models do not have attention heads @@ -877,6 +881,9 @@ void llama_model::load_hparams(llama_model_loader & ml) { { hparams.n_swa_pattern = 6; + hparams.rope_freq_base_train_swa = 10000.0f; + hparams.rope_freq_scale_train_swa = 1.0f; + ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa); ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); @@ -1346,13 +1353,14 @@ bool llama_model::load_tensors(llama_model_loader & ml) { const int i_gpu_start = std::max((int) hparams.n_layer - n_gpu_layers, (int) 0); const int act_gpu_layers = devices.empty() ? 0 : std::min(n_gpu_layers, (int)n_layer + 1); auto get_layer_buft_list = [&](int il) -> llama_model::impl::layer_dev { + const bool is_swa = il < (int) hparams.n_layer && hparams.is_swa(il); if (il < i_gpu_start || (il - i_gpu_start) >= act_gpu_layers) { - LLAMA_LOG_DEBUG("load_tensors: layer %3d assigned to device %s\n", il, ggml_backend_dev_name(cpu_dev)); + LLAMA_LOG_DEBUG("load_tensors: layer %3d assigned to device %s, is_swa = %d\n", il, ggml_backend_dev_name(cpu_dev), is_swa); return {cpu_dev, &pimpl->cpu_buft_list}; } const int layer_gpu = std::upper_bound(splits.begin(), splits.begin() + n_devices(), float(il - i_gpu_start)/act_gpu_layers) - splits.begin(); auto * dev = devices.at(layer_gpu); - LLAMA_LOG_DEBUG("load_tensors: layer %3d assigned to device %s\n", il, ggml_backend_dev_name(dev)); + LLAMA_LOG_DEBUG("load_tensors: layer %3d assigned to device %s, is_swa = %d\n", il, ggml_backend_dev_name(dev), is_swa); return {dev, &pimpl->gpu_buft_list.at(dev)}; }; @@ -7381,10 +7389,10 @@ struct llm_build_gemma3 : public llm_graph_context { auto * inp_attn = build_attn_inp_kv_unified(true, true); for (int il = 0; il < n_layer; ++il) { - const bool is_sliding = hparams.is_sliding(il); + const bool is_swa = hparams.is_swa(il); - const float freq_base_l = is_sliding ? 10000.0f : freq_base; - const float freq_scale_l = is_sliding ? 1.0f : freq_scale; + const float freq_base_l = is_swa ? hparams.rope_freq_base_train_swa : cparams.rope_freq_base; + const float freq_scale_l = is_swa ? hparams.rope_freq_scale_train_swa : cparams.rope_freq_scale; // norm cur = build_norm(inpL, model.layers[il].attn_norm, NULL, LLM_NORM_RMS, il); @@ -7973,7 +7981,7 @@ struct llm_build_cohere2 : public llm_graph_context { auto * inp_attn = build_attn_inp_kv_unified(true, true); for (int il = 0; il < n_layer; ++il) { - const bool is_sliding = hparams.is_sliding(il); + const bool is_swa = hparams.is_swa(il); // norm cur = build_norm(inpL, model.layers[il].attn_norm, NULL, LLM_NORM, il); @@ -8007,7 +8015,7 @@ struct llm_build_cohere2 : public llm_graph_context { cb(Vcur, "Vcur", il); } - if (is_sliding) { + if (is_swa) { Qcur = ggml_rope_ext(ctx0, ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, rope_factors, n_rot, rope_type, n_ctx_orig, freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow);