diff --git a/convert_hf_to_gguf.py b/convert_hf_to_gguf.py index 843c00a896..e64756a74a 100755 --- a/convert_hf_to_gguf.py +++ b/convert_hf_to_gguf.py @@ -4102,39 +4102,27 @@ class Qwen2MoeModel(TextModel): # process the experts separately name = name.replace("language_model.", "") # InternVL - # handle aggregated expert tensors - # GGUF stores dimensions reversed from PyTorch, so: - # PyTorch (A,B,C) -> GGUF writes [C,B,A] -> GGML reads ne={C,B,A} - # Input shapes from HF: (n_expert, n_ff_exp, n_embd) or (n_expert, n_embd, n_ff_exp) - # Expected GGML ne: {n_embd, n_ff_exp, n_expert} for gate/up, {n_ff_exp, n_embd, n_expert} for down + # handle pre-packed expert tensors (e.g. Qwen3.5 MoE, Qwen3Next) + # HF stores these using nn.Linear convention: [n_expert, out_features, in_features] + # This matches the individual expert stacking path below (which stacks + # per-expert [out, in] weights into [n_expert, out, in]), so no permute is needed. if name.endswith("mlp.experts.down_proj") or name.endswith("mlp.experts.down_proj.weight"): mapped = f"{name}.weight" if not name.endswith(".weight") else name - # Input: (n_expert=128, n_ff_exp=768, n_embd=2048) - # Want GGML ne: {n_ff_exp, n_embd, n_expert} = {768, 2048, 128} - # Need PyTorch: (128, 2048, 768) [reversed of GGML] - # So: permute(0, 2, 1): (128, 768, 2048) -> (128, 2048, 768) - permuted = data_torch.permute(0, 2, 1).contiguous() - yield from super().modify_tensors(permuted, mapped, bid) + # HF: [n_expert, n_embd, n_ff] → GGML: {n_ff, n_embd, n_expert} ✓ + yield from super().modify_tensors(data_torch, mapped, bid) return if name.endswith("mlp.experts.gate_up_proj") or name.endswith("mlp.experts.gate_up_proj.weight"): - if data_torch.ndim < 3 or data_torch.shape[-1] % 2 != 0: - raise ValueError(f"Unexpected gate_up_proj shape for {name}: {tuple(data_torch.shape)}") - split_dim = data_torch.shape[-1] // 2 - gate = data_torch[..., :split_dim].contiguous() - up = data_torch[..., split_dim:].contiguous() - # Input gate/up: (n_expert=128, n_embd=2048, n_ff_exp=768) - # Want GGML ne: {n_embd, n_ff_exp, n_expert} = {2048, 768, 128} - # Need PyTorch: (128, 768, 2048) [reversed of GGML] - # So: permute(0, 2, 1): (128, 2048, 768) -> (128, 768, 2048) - base_name = name.removesuffix(".weight") - base = base_name.rsplit('.', 1)[0] - mapped_gate = f"{base}.gate_proj.weight" - mapped_up = f"{base}.up_proj.weight" - perm_gate = gate.permute(0, 2, 1).contiguous() - perm_up = up.permute(0, 2, 1).contiguous() - yield from super().modify_tensors(perm_gate, mapped_gate, bid) - yield from super().modify_tensors(perm_up, mapped_up, bid) + # HF: [n_expert, 2*n_ff, n_embd] → split on dim=1 + n_ff = data_torch.shape[1] // 2 + gate = data_torch[:, :n_ff, :].contiguous() + up = data_torch[:, n_ff:, :].contiguous() + # gate/up: [n_expert, n_ff, n_embd] → GGML: {n_embd, n_ff, n_expert} ✓ + base_name = name.removesuffix(".weight").removesuffix(".gate_up_proj") + mapped_gate = f"{base_name}.gate_proj.weight" + mapped_up = f"{base_name}.up_proj.weight" + yield from super().modify_tensors(gate, mapped_gate, bid) + yield from super().modify_tensors(up, mapped_up, bid) return if name.startswith("mlp") or name.startswith("vision_model") or name.startswith("model.vision_tower") or name.startswith("model.multi_modal_projector") or name.startswith("model.visual"): @@ -4344,6 +4332,40 @@ class Qwen3NextModel(Qwen2MoeModel): yield from super().modify_tensors(data_torch, name, bid) +@ModelBase.register("Qwen3_5ForCausalLM", "Qwen3_5TextForCausalLM") +class Qwen3_5Model(Qwen3NextModel): + model_arch = gguf.MODEL_ARCH.QWEN3_5 + + # Stores whichever of in_proj_a/in_proj_b is seen first, keyed by layer + _pending_ba: dict[int | None, tuple[str, Tensor]] = {} + + def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]: + # Handle split in_proj_b + in_proj_a → concatenated SSM_BETA_ALPHA + # safetensors sorts alphabetically so in_proj_a arrives before in_proj_b + if "in_proj_a.weight" in name or "in_proj_b.weight" in name: + which = "a" if "in_proj_a" in name else "b" + if bid not in self._pending_ba: + self._pending_ba[bid] = (which, data_torch) + return + prev_which, prev_tensor = self._pending_ba.pop(bid) + assert prev_which != which, f"duplicate in_proj_{which} for layer {bid}" + b_tensor = prev_tensor if prev_which == "b" else data_torch + a_tensor = prev_tensor if prev_which == "a" else data_torch + ba_combined = torch.cat([b_tensor, a_tensor], dim=0) + yield (self.format_tensor_name(gguf.MODEL_TENSOR.SSM_BETA_ALPHA, bid, ".weight"), ba_combined) + return + else: + # Qwen3Next uses .qkvz tensor, so we use the super to get the other functionalities + # (norm correction, A_log to A etc.) for free + # Qwen2Moe already does the gate_up conversion properly, just use that + yield from super().modify_tensors(data_torch, name, bid) + + +@ModelBase.register("Qwen3_5MoeForCausalLM", "Qwen3_5MoeTextForCausalLM") +class Qwen3_5MoeModel(Qwen3_5Model): + model_arch = gguf.MODEL_ARCH.QWEN3_5_MOE + + @ModelBase.register("RND1") class RND1Model(Qwen2MoeModel): model_arch = gguf.MODEL_ARCH.RND1 diff --git a/gguf-py/gguf/constants.py b/gguf-py/gguf/constants.py index 3af4fffe95..8a3fab1e1c 100644 --- a/gguf-py/gguf/constants.py +++ b/gguf-py/gguf/constants.py @@ -382,6 +382,8 @@ class MODEL_ARCH(IntEnum): QWEN3 = auto() QWEN3MOE = auto() QWEN3NEXT = auto() + QWEN3_5 = auto() + QWEN3_5_MOE = auto() QWEN3VL = auto() QWEN3VLMOE = auto() PHI2 = auto() @@ -812,6 +814,8 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = { MODEL_ARCH.QWEN3: "qwen3", MODEL_ARCH.QWEN3MOE: "qwen3moe", MODEL_ARCH.QWEN3NEXT: "qwen3next", + MODEL_ARCH.QWEN3_5: "qwen3_5", + MODEL_ARCH.QWEN3_5_MOE: "qwen3_5moe", MODEL_ARCH.QWEN3VL: "qwen3vl", MODEL_ARCH.QWEN3VLMOE: "qwen3vlmoe", MODEL_ARCH.PHI2: "phi2", @@ -1784,6 +1788,61 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.SSM_BETA_ALPHA, MODEL_TENSOR.SSM_OUT ], + MODEL_ARCH.QWEN3_5: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_Q_NORM, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_K_NORM, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.ATTN_POST_NORM, + MODEL_TENSOR.ATTN_GATE, + MODEL_TENSOR.ATTN_QKV, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + MODEL_TENSOR.SSM_A, + MODEL_TENSOR.SSM_CONV1D, + MODEL_TENSOR.SSM_DT, + MODEL_TENSOR.SSM_NORM, + MODEL_TENSOR.SSM_IN, + MODEL_TENSOR.SSM_BETA_ALPHA, + MODEL_TENSOR.SSM_OUT, + ], + MODEL_ARCH.QWEN3_5_MOE: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_Q_NORM, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_K_NORM, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.ATTN_POST_NORM, + MODEL_TENSOR.ATTN_GATE, + MODEL_TENSOR.ATTN_QKV, + MODEL_TENSOR.FFN_GATE_INP, + MODEL_TENSOR.FFN_GATE_INP_SHEXP, + MODEL_TENSOR.FFN_UP_SHEXP, + MODEL_TENSOR.FFN_DOWN_SHEXP, + MODEL_TENSOR.FFN_GATE_SHEXP, + MODEL_TENSOR.FFN_DOWN_EXP, + MODEL_TENSOR.FFN_UP_EXP, + MODEL_TENSOR.FFN_GATE_EXP, + MODEL_TENSOR.SSM_A, + MODEL_TENSOR.SSM_CONV1D, + MODEL_TENSOR.SSM_DT, + MODEL_TENSOR.SSM_NORM, + MODEL_TENSOR.SSM_IN, + MODEL_TENSOR.SSM_BETA_ALPHA, + MODEL_TENSOR.SSM_OUT, + ], MODEL_ARCH.QWEN3VL: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, diff --git a/gguf-py/gguf/tensor_mapping.py b/gguf-py/gguf/tensor_mapping.py index 167ade7803..43f32c7b52 100644 --- a/gguf-py/gguf/tensor_mapping.py +++ b/gguf-py/gguf/tensor_mapping.py @@ -228,6 +228,7 @@ class TensorNameMap: "transformer_encoder.{bid}.qkv", # neobert "layers.{bid}.attn.Wqkv", # modern-bert "model.layers.{bid}.self_attn.language_expert_query_key_value", # cogvlm + "model.layers.{bid}.linear_attn.in_proj_qkv", # qwen3.5 ), # Attention query @@ -358,8 +359,9 @@ class TensorNameMap: ), MODEL_TENSOR.ATTN_GATE: ( - "model.layers.{bid}.self_attn.gate_proj", # afmoe - "model.layers.{bid}.self_attn.g_proj", # step3.5 head-wise attention gate + "model.layers.{bid}.self_attn.gate_proj", # afmoe + "model.layers.{bid}.self_attn.g_proj", # step3.5 head-wise attention gate + "model.layers.{bid}.linear_attn.in_proj_z", # qwen3.5 ), # Feed-forward norm diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt index 2115fc4255..0c164617a1 100644 --- a/src/CMakeLists.txt +++ b/src/CMakeLists.txt @@ -57,6 +57,7 @@ add_library(llama models/deci.cpp models/deepseek.cpp models/deepseek2.cpp + models/delta.cpp models/dots1.cpp models/dream.cpp models/ernie4-5-moe.cpp @@ -122,6 +123,8 @@ add_library(llama models/qwen3vl-moe.cpp models/qwen3moe.cpp models/qwen3next.cpp + models/qwen3-5.cpp + models/qwen3-5moe.cpp models/refact.cpp models/rnd1.cpp models/rwkv6-base.cpp diff --git a/src/llama-arch.cpp b/src/llama-arch.cpp index bd78f1e556..fce46772d7 100644 --- a/src/llama-arch.cpp +++ b/src/llama-arch.cpp @@ -35,6 +35,8 @@ static const std::map LLM_ARCH_NAMES = { { LLM_ARCH_QWEN3, "qwen3" }, { LLM_ARCH_QWEN3MOE, "qwen3moe" }, { LLM_ARCH_QWEN3NEXT, "qwen3next" }, + { LLM_ARCH_QWEN3_5, "qwen3_5" }, + { LLM_ARCH_QWEN3_5_MOE, "qwen3_5moe" }, { LLM_ARCH_QWEN3VL, "qwen3vl" }, { LLM_ARCH_QWEN3VLMOE, "qwen3vlmoe" }, { LLM_ARCH_PHI2, "phi2" }, @@ -985,6 +987,63 @@ static std::set llm_get_tensor_names(llm_arch arch) { LLM_TENSOR_SSM_NORM, LLM_TENSOR_SSM_OUT, }; + case LLM_ARCH_QWEN3_5: + return { + LLM_TENSOR_TOKEN_EMBD, + LLM_TENSOR_OUTPUT_NORM, + LLM_TENSOR_OUTPUT, + LLM_TENSOR_ATTN_NORM, + LLM_TENSOR_ATTN_POST_NORM, + LLM_TENSOR_ATTN_Q, + LLM_TENSOR_ATTN_Q_NORM, + LLM_TENSOR_ATTN_K, + LLM_TENSOR_ATTN_K_NORM, + LLM_TENSOR_ATTN_V, + LLM_TENSOR_ATTN_OUT, + LLM_TENSOR_ATTN_QKV, + LLM_TENSOR_ATTN_GATE, + LLM_TENSOR_FFN_GATE, + LLM_TENSOR_FFN_DOWN, + LLM_TENSOR_FFN_UP, + LLM_TENSOR_SSM_A_NOSCAN, + LLM_TENSOR_SSM_CONV1D, + LLM_TENSOR_SSM_DT, + LLM_TENSOR_SSM_BETA_ALPHA, + LLM_TENSOR_SSM_IN, + LLM_TENSOR_SSM_NORM, + LLM_TENSOR_SSM_OUT, + }; + case LLM_ARCH_QWEN3_5_MOE: + return { + LLM_TENSOR_TOKEN_EMBD, + LLM_TENSOR_OUTPUT_NORM, + LLM_TENSOR_OUTPUT, + LLM_TENSOR_ATTN_NORM, + LLM_TENSOR_ATTN_POST_NORM, + LLM_TENSOR_ATTN_Q, + LLM_TENSOR_ATTN_Q_NORM, + LLM_TENSOR_ATTN_K, + LLM_TENSOR_ATTN_K_NORM, + LLM_TENSOR_ATTN_V, + LLM_TENSOR_ATTN_OUT, + LLM_TENSOR_ATTN_QKV, + LLM_TENSOR_ATTN_GATE, + LLM_TENSOR_FFN_GATE_INP, + LLM_TENSOR_FFN_GATE_EXPS, + LLM_TENSOR_FFN_DOWN_EXPS, + LLM_TENSOR_FFN_UP_EXPS, + LLM_TENSOR_FFN_GATE_INP_SHEXP, + LLM_TENSOR_FFN_GATE_SHEXP, + LLM_TENSOR_FFN_DOWN_SHEXP, + LLM_TENSOR_FFN_UP_SHEXP, + LLM_TENSOR_SSM_A_NOSCAN, + LLM_TENSOR_SSM_CONV1D, + LLM_TENSOR_SSM_DT, + LLM_TENSOR_SSM_BETA_ALPHA, + LLM_TENSOR_SSM_IN, + LLM_TENSOR_SSM_NORM, + LLM_TENSOR_SSM_OUT, + }; case LLM_ARCH_QWEN3VL: case LLM_ARCH_CHAMELEON: case LLM_ARCH_HUNYUAN_DENSE: @@ -2674,6 +2733,8 @@ bool llm_arch_is_hybrid(const llm_arch & arch) { case LLM_ARCH_NEMOTRON_H: case LLM_ARCH_NEMOTRON_H_MOE: case LLM_ARCH_QWEN3NEXT: + case LLM_ARCH_QWEN3_5: + case LLM_ARCH_QWEN3_5_MOE: case LLM_ARCH_KIMI_LINEAR: return true; default: diff --git a/src/llama-arch.h b/src/llama-arch.h index e8263369b8..a392ecce2b 100644 --- a/src/llama-arch.h +++ b/src/llama-arch.h @@ -39,6 +39,8 @@ enum llm_arch { LLM_ARCH_QWEN3, LLM_ARCH_QWEN3MOE, LLM_ARCH_QWEN3NEXT, + LLM_ARCH_QWEN3_5, + LLM_ARCH_QWEN3_5_MOE, LLM_ARCH_QWEN3VL, LLM_ARCH_QWEN3VLMOE, LLM_ARCH_PHI2, diff --git a/src/llama-context.cpp b/src/llama-context.cpp index a6df893a31..80b9a7d46a 100644 --- a/src/llama-context.cpp +++ b/src/llama-context.cpp @@ -2013,7 +2013,7 @@ void llama_context::output_reorder() { // uint32_t llama_context::graph_max_nodes(uint32_t n_tokens) const { - if (model.arch == LLM_ARCH_QWEN3NEXT || model.arch == LLM_ARCH_KIMI_LINEAR) { + if (model.arch == LLM_ARCH_QWEN3NEXT || model.arch == LLM_ARCH_QWEN3_5 || model.arch == LLM_ARCH_QWEN3_5_MOE || model.arch == LLM_ARCH_KIMI_LINEAR) { return std::max(n_tokens * 40, 32u * model.n_tensors()); } uint32_t res = std::max(1024u, 8u*model.n_tensors()); diff --git a/src/llama-model.cpp b/src/llama-model.cpp index 674d06c891..8fc61aee37 100644 --- a/src/llama-model.cpp +++ b/src/llama-model.cpp @@ -2412,6 +2412,25 @@ void llama_model::load_hparams(llama_model_loader & ml) { default: type = LLM_TYPE_UNKNOWN; } } break; + case LLM_ARCH_QWEN3_5: + case LLM_ARCH_QWEN3_5_MOE: + { + ml.get_key(LLM_KV_EXPERT_FEED_FORWARD_LENGTH, hparams.n_ff_exp, false); + ml.get_key(LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH, hparams.n_ff_shexp, false); + ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); + + // Load linear attention (gated delta net) parameters + ml.get_key(LLM_KV_SSM_CONV_KERNEL, hparams.ssm_d_conv); + ml.get_key(LLM_KV_SSM_INNER_SIZE, hparams.ssm_d_inner); + ml.get_key(LLM_KV_SSM_STATE_SIZE, hparams.ssm_d_state); + ml.get_key(LLM_KV_SSM_TIME_STEP_RANK, hparams.ssm_dt_rank); + ml.get_key(LLM_KV_SSM_GROUP_COUNT, hparams.ssm_n_group); + + // Mark recurrent layers (linear attention layers) + for (uint32_t i = 0; i < hparams.n_layer; ++i) { + hparams.recurrent_layer_arr[i] = ((i + 1) % 4 != 0); + } + } break; case LLM_ARCH_MISTRAL3: { ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps); @@ -7094,6 +7113,129 @@ bool llama_model::load_tensors(llama_model_loader & ml) { layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff_exp, n_embd, n_expert }, 0); layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert }, 0); + // Shared experts + layer.ffn_gate_inp_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP_SHEXP, "weight", i), { n_embd }, 0); + layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), { n_embd, hparams.n_ff_shexp }, 0); + layer.ffn_up_shexp = create_tensor(tn(LLM_TENSOR_FFN_UP_SHEXP, "weight", i), { n_embd, hparams.n_ff_shexp }, 0); + layer.ffn_down_shexp = create_tensor(tn(LLM_TENSOR_FFN_DOWN_SHEXP, "weight", i), { hparams.n_ff_shexp, n_embd }, 0); + } + } break; + case LLM_ARCH_QWEN3_5: + { + 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 == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), { n_embd, n_vocab }, TENSOR_DUPLICATED); + } + + // Calculate dimensions from hyperparameters + const int64_t head_k_dim = hparams.ssm_d_state; + const int64_t head_v_dim = hparams.ssm_d_state; + const int64_t n_k_heads = hparams.ssm_n_group; + const int64_t n_v_heads = hparams.ssm_dt_rank; + const int64_t key_dim = head_k_dim * n_k_heads; + const int64_t value_dim = head_v_dim * n_v_heads; + const int64_t conv_dim = key_dim * 2 + value_dim; + + const int64_t ba_dim = n_v_heads * 2; + + 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); + layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), { n_embd }, 0); + + if (!hparams.is_recurrent(i)) { + // Full attention layers + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), { n_embd, n_embd_head_k * n_head * 2 }, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), { n_embd, n_embd_k_gqa }, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), { n_embd, 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_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), { n_embd_head_k }, 0); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), { n_embd_head_k }, 0); + } else { + // Linear attention (gated delta net) specific tensors + layer.ssm_in = create_tensor(tn(LLM_TENSOR_SSM_IN, "weight", i), { n_embd, key_dim * 2 + value_dim * 2 }, TENSOR_NOT_REQUIRED); + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), { n_embd, key_dim * 2 + value_dim }, TENSOR_NOT_REQUIRED); + layer.wqkv_gate = create_tensor(tn(LLM_TENSOR_ATTN_GATE, "weight", i), { n_embd, value_dim }, TENSOR_NOT_REQUIRED); + layer.ssm_conv1d = create_tensor(tn(LLM_TENSOR_SSM_CONV1D, "weight", i), { hparams.ssm_d_conv, conv_dim }, 0); + layer.ssm_dt = create_tensor(tn(LLM_TENSOR_SSM_DT, "bias", i), { hparams.ssm_dt_rank }, 0); + layer.ssm_a = create_tensor(tn(LLM_TENSOR_SSM_A_NOSCAN, i), { hparams.ssm_dt_rank }, 0); + layer.ssm_beta_alpha = create_tensor(tn(LLM_TENSOR_SSM_BETA_ALPHA, "weight", i), { n_embd, ba_dim }, 0); + layer.ssm_norm = create_tensor(tn(LLM_TENSOR_SSM_NORM, "weight", i), { head_v_dim }, 0); + layer.ssm_out = create_tensor(tn(LLM_TENSOR_SSM_OUT, "weight", i), { value_dim, n_embd }, 0); + } + + // Dense FFN for all layers + 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); + } + } break; + case LLM_ARCH_QWEN3_5_MOE: + { + 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 == NULL) { + output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), { n_embd, n_vocab }, TENSOR_DUPLICATED); + } + + const int64_t n_ff_exp = hparams.n_ff_exp ? hparams.n_ff_exp : n_ff / n_expert_used; + + // Calculate dimensions from hyperparameters + const int64_t head_k_dim = hparams.ssm_d_state; + const int64_t head_v_dim = hparams.ssm_d_state; + const int64_t n_k_heads = hparams.ssm_n_group; + const int64_t n_v_heads = hparams.ssm_dt_rank; + const int64_t key_dim = head_k_dim * n_k_heads; + const int64_t value_dim = head_v_dim * n_v_heads; + const int64_t conv_dim = key_dim * 2 + value_dim; + + const int64_t ba_dim = n_v_heads * 2; + + 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); + layer.attn_post_norm = create_tensor(tn(LLM_TENSOR_ATTN_POST_NORM, "weight", i), { n_embd }, 0); + + if (!hparams.is_recurrent(i)) { + // Full attention layers + layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), { n_embd, n_embd_head_k * n_head * 2 }, 0); + layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), { n_embd, n_embd_k_gqa }, 0); + layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), { n_embd, 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_q_norm = create_tensor(tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), { n_embd_head_k }, 0); + layer.attn_k_norm = create_tensor(tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), { n_embd_head_k }, 0); + } else { + // Linear attention (gated delta net) specific tensors + layer.ssm_in = create_tensor(tn(LLM_TENSOR_SSM_IN, "weight", i), { n_embd, key_dim * 2 + value_dim * 2 }, TENSOR_NOT_REQUIRED); + layer.wqkv = create_tensor(tn(LLM_TENSOR_ATTN_QKV, "weight", i), { n_embd, key_dim * 2 + value_dim }, TENSOR_NOT_REQUIRED); + layer.wqkv_gate = create_tensor(tn(LLM_TENSOR_ATTN_GATE, "weight", i), { n_embd, value_dim }, TENSOR_NOT_REQUIRED); + layer.ssm_conv1d = create_tensor(tn(LLM_TENSOR_SSM_CONV1D, "weight", i), { hparams.ssm_d_conv, conv_dim }, 0); + layer.ssm_dt = create_tensor(tn(LLM_TENSOR_SSM_DT, "bias", i), { hparams.ssm_dt_rank }, 0); + layer.ssm_a = create_tensor(tn(LLM_TENSOR_SSM_A_NOSCAN, i), { hparams.ssm_dt_rank }, 0); + layer.ssm_beta_alpha = create_tensor(tn(LLM_TENSOR_SSM_BETA_ALPHA, "weight", i), { n_embd, ba_dim }, 0); + layer.ssm_norm = create_tensor(tn(LLM_TENSOR_SSM_NORM, "weight", i), { head_v_dim }, 0); + layer.ssm_out = create_tensor(tn(LLM_TENSOR_SSM_OUT, "weight", i), { value_dim, n_embd }, 0); + } + + // MoE FFN + layer.ffn_gate_inp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), { n_embd, n_expert }, 0); + layer.ffn_gate_exps = create_tensor(tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert }, 0); + layer.ffn_down_exps = create_tensor(tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff_exp, n_embd, n_expert }, 0); + layer.ffn_up_exps = create_tensor(tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), { n_embd, n_ff_exp, n_expert }, 0); + // Shared experts layer.ffn_gate_inp_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_INP_SHEXP, "weight", i), { n_embd }, 0); layer.ffn_gate_shexp = create_tensor(tn(LLM_TENSOR_FFN_GATE_SHEXP, "weight", i), { n_embd, hparams.n_ff_shexp }, 0); @@ -7545,6 +7687,8 @@ void llama_model::print_info() const { arch == LLM_ARCH_PLAMO2 || arch == LLM_ARCH_GRANITE_HYBRID || arch == LLM_ARCH_QWEN3NEXT || + arch == LLM_ARCH_QWEN3_5 || + arch == LLM_ARCH_QWEN3_5_MOE || arch == LLM_ARCH_NEMOTRON_H || arch == LLM_ARCH_NEMOTRON_H_MOE) { LLAMA_LOG_INFO("%s: ssm_d_conv = %u\n", __func__, hparams.ssm_d_conv); @@ -8343,6 +8487,14 @@ ggml_cgraph * llama_model::build_graph(const llm_graph_params & params) const { { llm = std::make_unique(*this, params); } break; + case LLM_ARCH_QWEN3_5: + { + llm = std::make_unique(*this, params); + } break; + case LLM_ARCH_QWEN3_5_MOE: + { + llm = std::make_unique(*this, params); + } break; case LLM_ARCH_MISTRAL3: { llm = std::make_unique(*this, params); @@ -8603,6 +8755,8 @@ llama_rope_type llama_model_rope_type(const llama_model * model) { case LLM_ARCH_PANGU_EMBED: case LLM_ARCH_AFMOE: case LLM_ARCH_QWEN3NEXT: + case LLM_ARCH_QWEN3_5: + case LLM_ARCH_QWEN3_5_MOE: case LLM_ARCH_MIMO2: case LLM_ARCH_STEP35: return LLAMA_ROPE_TYPE_NEOX; diff --git a/src/models/delta.cpp b/src/models/delta.cpp new file mode 100644 index 0000000000..d1d9837d09 --- /dev/null +++ b/src/models/delta.cpp @@ -0,0 +1,618 @@ +#include "models.h" +#include "ggml.h" +#include +#include +#include + +llm_graph_context_delta::llm_graph_context_delta(const llm_graph_params & params) : llm_graph_context_mamba(params) {} + +/** + * Unified Delta Net implementation supporting both GDA and KDA modes. + * + * GDA (Gated Delta Attention): g has shape [H, T, B] in GGML (PyTorch: [B, T, H]) + * - Per-head gating, broadcasts over K dimension + * + * KDA (Key-wise Delta Attention): g has shape [K, H, T, B] in GGML (PyTorch: [B, T, H, K]) + * - Per-key gating + * + * The mode is auto-detected based on g's dimensionality. + * + * Tensor dimension convention: + * GGML: ne[0] is innermost (fastest varying), ne[3] is outermost + * PyTorch: dim 0 is outermost, dim -1 is innermost + * So GGML [A, B, C, D] corresponds to PyTorch [D, C, B, A] + */ + +// Helper to get a slice along dimension 2 (n_chunks dimension) +static ggml_tensor * get_slice_2d(ggml_context * ctx, ggml_tensor * t, int64_t chunk) { + return ggml_view_4d(ctx, t, + t->ne[0], t->ne[1], 1, t->ne[3], + t->nb[1], t->nb[2], t->nb[3], + chunk * t->nb[2]); +} + +/** + * Unified chunked Delta Net implementation. + * + * Input tensor format matches qwen3next conventions: + * @param q Query tensor [S_k, H_k, n_tokens, n_seqs] + * @param k Key tensor [S_k, H_k, n_tokens, n_seqs] + * @param v Value tensor [S_v, H_v, n_tokens, n_seqs] + * @param g Gate tensor: + * GDA: [H_v, n_tokens, n_seqs] + * KDA: [S_k, H_v, n_tokens, n_seqs] + * @param beta Beta tensor [H_v, 1, n_tokens, n_seqs] + * @param state State tensor [S_v, S_v * H_v, 1, n_seqs] + * @param causal_mask Lower triangular mask [chunk_size, chunk_size] + * @param identity Identity matrix [chunk_size, chunk_size] + * @param diag_mask Diagonal mask [chunk_size, chunk_size] + * @param il Layer index (for debugging callbacks) + * @param chunk_size Chunk size for chunked processing + * @param eps_norm Epsilon for L2 normalization + * + * @return Pair of (output_tokens, new_state) + */ +std::pair llm_graph_context_delta::build_delta_net_unified_chunking( + ggml_context * ctx0, + ggml_tensor * q, + ggml_tensor * k, + ggml_tensor * v, + ggml_tensor * g, + ggml_tensor * beta, + ggml_tensor * state_reshaped, + ggml_tensor * causal_mask, + ggml_tensor * identity, + ggml_tensor * diag_mask, + int il, + int64_t chunk_size, + float eps_norm) { + + // Input format: [S, H, n_tokens, n_seqs] (matching qwen3next convention) + 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]; + + // Detect KDA vs GDA based on g's shape + // GDA: g has shape [H_v, n_tokens, n_seqs] + // KDA: g has shape [S_k, H_v, n_tokens, n_seqs] (4D with ne[0]=S_k) + const bool is_kda = (g->ne[0] == S_k && g->ne[1] == H_v); + + // Validate tensor shapes + GGML_ASSERT(v->ne[2] == n_tokens); + GGML_ASSERT(k->ne[2] == n_tokens); + GGML_ASSERT(state_reshaped->ne[0] == S_v && state_reshaped->ne[1] == S_v && state_reshaped->ne[2] == H_v && state_reshaped->ne[3] == n_seqs); + 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(beta->ne[0] == H_v && beta->ne[2] == n_tokens && beta->ne[3] == n_seqs); + GGML_ASSERT(H_k == H_v); + + if (is_kda) { + // KDA: g shape [S_k, H_v, n_tokens, n_seqs] + GGML_ASSERT(g->ne[0] == S_k && g->ne[1] == H_v && g->ne[2] == n_tokens && g->ne[3] == n_seqs); + } else { + // GDA: g shape [H_v, n_tokens, n_seqs] + GGML_ASSERT(g->ne[0] == H_v && g->ne[1] == n_tokens && g->ne[2] == n_seqs); + } + + // L2 normalize q and k + q = ggml_l2_norm(ctx0, q, eps_norm); + k = ggml_l2_norm(ctx0, k, eps_norm); + + const float scale = 1.0f / sqrtf((float)S_v); + q = ggml_scale(ctx0, q, scale); + + beta = ggml_sigmoid(ctx0, beta); + + cb(q, "q_in", il); + cb(k, "k_in", il); + cb(v, "v_in", il); + cb(beta, "beta_in", il); + cb(g, "g_in", il); + + // Permute tensors to working format [S, n_tokens, H, n_seqs] + // Input: [S, H, n_tokens, n_seqs] -> permute(0, 2, 1, 3) -> [S, n_tokens, H, n_seqs] + q = ggml_cont_4d(ctx0, ggml_permute(ctx0, q, 0, 2, 1, 3), S_k, n_tokens, H_k, n_seqs); + k = ggml_cont_4d(ctx0, ggml_permute(ctx0, k, 0, 2, 1, 3), S_k, n_tokens, H_k, n_seqs); + v = ggml_cont_4d(ctx0, ggml_permute(ctx0, v, 0, 2, 1, 3), S_v, n_tokens, H_v, n_seqs); + if (is_kda) { + g = ggml_cont_4d(ctx0, ggml_permute(ctx0, g, 0, 2, 1, 3), S_k, n_tokens, H_k, n_seqs); + } else { + g = ggml_cont_4d(ctx0, ggml_permute(ctx0, g, 2, 0, 3, 1), n_tokens, 1, H_k, n_seqs); + } + beta = ggml_cont(ctx0, ggml_permute(ctx0, beta, 2, 0, 1, 3)); + + cb(q, "q_perm", il); + cb(k, "k_perm", il); + cb(v, "v_perm", il); + cb(beta, "beta_perm", il); + cb(g, "g_perm", il); + cb(state_reshaped, "state_in", il); + + // Padding for chunk processing + const int64_t pad = (chunk_size - n_tokens % chunk_size) % chunk_size; + const int64_t n_chunks = (n_tokens + pad) / chunk_size; + + 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); + beta = ggml_pad(ctx0, beta, 0, pad, 0, 0); + g = ggml_pad(ctx0, g, pad, 0, 0, 0); + + + cb(q, "q_pad", il); + cb(k, "k_pad", il); + cb(v, "v_pad", il); + cb(beta, "beta_pad", il); + cb(g, "g_pad", il); + + ggml_tensor * v_beta = ggml_mul(ctx0, v, beta); + ggml_tensor * k_beta = ggml_mul(ctx0, k, beta); + + cb(v_beta, "v_beta", il); + cb(k_beta, "k_beta", il); + + // Reshape to chunks + q = ggml_reshape_4d(ctx0, q, S_k, chunk_size, n_chunks, H_k * n_seqs); + k = ggml_reshape_4d(ctx0, k, S_k, chunk_size, n_chunks, H_k * n_seqs); + k_beta = ggml_reshape_4d(ctx0, k_beta, S_k, chunk_size, n_chunks, H_k * n_seqs); + v = ggml_reshape_4d(ctx0, v, S_v, chunk_size, n_chunks, H_v * n_seqs); + v_beta = ggml_reshape_4d(ctx0, v_beta, S_v, chunk_size, n_chunks, H_v * n_seqs); + beta = ggml_reshape_4d(ctx0, beta, 1, chunk_size, n_chunks, H_k * n_seqs); + + // Reshape g for chunks + ggml_tensor * g_cumsum; + ggml_tensor * g_cumsum_t; + if (is_kda) { + // KDA: g [S_k, n_tokens+pad, H_k, n_seqs] -> [S_k, chunk_size, n_chunks, H_k * n_seqs] + g = ggml_reshape_4d(ctx0, g, S_k, chunk_size, n_chunks, H_k * n_seqs); + // Cumsum along chunk_size dimension (ne[1]) + // GGML cumsum operates on ne[0], so we need to transpose, cumsum, transpose back + g = ggml_cont(ctx0, ggml_transpose(ctx0, g)); // [chunk_size, S_k, n_chunks, H_k * n_seqs] + g_cumsum_t = ggml_cumsum(ctx0, g); + g_cumsum = ggml_cont(ctx0, ggml_transpose(ctx0, g_cumsum_t)); // [S_k, chunk_size, n_chunks, H_k * n_seqs] + } else { + // GDA: g [n_tokens+pad, 1, H_k, n_seqs] -> [chunk_size, 1, n_chunks, H_k * n_seqs] + g = ggml_reshape_4d(ctx0, g, chunk_size, 1, n_chunks, H_k * n_seqs); + g_cumsum = ggml_cumsum(ctx0, g); + g_cumsum_t = ggml_reshape_4d(ctx0, g_cumsum, 1, chunk_size, n_chunks, H_k * n_seqs); + } + + cb(g_cumsum, "g_cumsum", il); + + // Build attention matrix A for the WY representation solve + // For GDA: A[j,i] = sum_k(k[j,k] * exp(g[j] - g[i]) * k[i,k]) = (k @ k^T) * exp(g[j] - g[i]) + // For KDA: A[j,i] = sum_k(k_beta[j,k] * exp(g[j,k] - g[i,k]) * k[i,k]) + // KDA uses decay mask with S_k packed into batch to compute exp(g[j,k] - g[i,k]) per-key + + ggml_tensor * k_decay; + ggml_tensor * decay_mask = nullptr; + ggml_tensor * g_exp_pos = nullptr; + + if (is_kda) { + // KDA: Use decay mask with S_k in leading dimension for efficient mul_mat reduction + // A[j,i] = sum_k(k_beta[j,k] * exp(g[j,k] - g[i,k]) * k[i,k]) + // By putting S_k in dim 0, mul_mat implicitly sums over it + + const int64_t CHB = n_chunks * H_k * n_seqs; + + // g_cumsum_t is [chunk_size, S_k, n_chunks, H_k * n_seqs] + // Reshape to [chunk_size, S_k, CHB] then build decay mask + ggml_tensor * gcs = ggml_reshape_3d(ctx0, g_cumsum_t, chunk_size, S_k, CHB); + ggml_tensor * gcs_i = ggml_reshape_4d(ctx0, gcs, chunk_size, 1, S_k, CHB); + ggml_tensor * gcs_j = ggml_reshape_4d(ctx0, gcs, 1, chunk_size, S_k, CHB); + + // Build decay mask: [chunk_size, chunk_size, S_k, CHB] + ggml_tensor * gcs_j_bc = ggml_repeat_4d(ctx0, gcs_j, chunk_size, chunk_size, S_k, CHB); + decay_mask = ggml_sub(ctx0, gcs_j_bc, gcs_i); + + cb(decay_mask, "decay_mask_kda", il); + + decay_mask = ggml_mul(ctx0, decay_mask, diag_mask); + decay_mask = ggml_exp(ctx0, decay_mask); + decay_mask = ggml_mul(ctx0, decay_mask, diag_mask); + + // Permute to [S_k, chunk_size_j, chunk_size_i, CHB] for mul_mat reduction over S_k + decay_mask = ggml_cont_4d(ctx0, ggml_permute(ctx0, decay_mask, 2, 1, 0, 3), S_k, chunk_size, chunk_size, CHB); + + // Reshape k and k_beta for broadcasting with decay_mask + // k_i: indexed at position i (dim 2 of decay_mask) + // k_beta_j: indexed at position j (dim 1 of decay_mask) + ggml_tensor * k_i = ggml_reshape_4d(ctx0, k, S_k, 1, chunk_size, CHB); + ggml_tensor * k_beta_j = ggml_reshape_4d(ctx0, k_beta, S_k, chunk_size, 1, CHB); + + // decay_k_beta_j[s,j,i,b] = decay[s,j,i,b] * k_beta[s,j,b] + ggml_tensor * decay_k_beta_j = ggml_mul(ctx0, decay_mask, k_beta_j); + + // mul_mat sums over S_k: result[j,1,i,CHB] = sum_s decay_k_beta_j[s,j,i,b] * k_i[s,1,i,b] + k_decay = ggml_mul_mat(ctx0, decay_k_beta_j, k_i); + k_decay = ggml_cont(ctx0, ggml_transpose(ctx0, ggml_reshape_4d(ctx0, k_decay, chunk_size, chunk_size, n_chunks, H_k * n_seqs))); + + // g_exp_pos is still needed for later (kbeta_gexp, etc.) + g_exp_pos = ggml_exp(ctx0, g_cumsum); + } else { + // GDA: Use decay mask approach (g broadcasts over K dimension) + // g_cumsum [chunk_size, 1, n_chunks, H_v * n_seqs] + ggml_tensor * gcs_i = g_cumsum; + ggml_tensor * gcs_j = g_cumsum_t; + g_exp_pos = ggml_exp(ctx0, g_cumsum_t); + ggml_tensor * gcs_j_broadcast = ggml_repeat_4d(ctx0, gcs_j, chunk_size, chunk_size, n_chunks, H_v * n_seqs); + decay_mask = ggml_sub(ctx0, gcs_j_broadcast, gcs_i); + + cb(decay_mask, "decay_mask", il); + + decay_mask = ggml_mul(ctx0, decay_mask, diag_mask); + decay_mask = ggml_exp(ctx0, decay_mask); + decay_mask = ggml_mul(ctx0, decay_mask, diag_mask); + + ggml_tensor * kmulkbeta = ggml_mul_mat(ctx0, k, k_beta); + k_decay = ggml_mul(ctx0, kmulkbeta, decay_mask); + } + + ggml_tensor * attn = ggml_neg(ctx0, ggml_mul(ctx0, k_decay, causal_mask)); + + cb(attn, "attn_pre_solve", il); + + // Solve triangular system: (I + L) @ X = I, where L is strictly lower triangular + ggml_tensor * attn_lower = ggml_mul(ctx0, attn, causal_mask); + ggml_tensor * lhs = ggml_sub(ctx0, ggml_repeat(ctx0, identity, attn_lower), attn_lower); + ggml_tensor * lin_solve = ggml_solve_tri(ctx0, lhs, attn, true, true, false); + attn = ggml_mul(ctx0, lin_solve, causal_mask); + attn = ggml_add(ctx0, attn, identity); + + cb(attn, "attn_solved", il); + + // Compute u = A @ v and w = A @ (g.exp() * k) + v = ggml_mul_mat(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, v_beta)), attn); + + ggml_tensor * kbeta_gexp = ggml_mul(ctx0, k_beta, g_exp_pos); + cb(kbeta_gexp, "kbeta_gexp", il); + + ggml_tensor * k_cumdecay = ggml_cont(ctx0, ggml_transpose(ctx0, + ggml_mul_mat(ctx0, attn, ggml_cont(ctx0, ggml_transpose(ctx0, kbeta_gexp))))); + cb(k_cumdecay, "k_cumdecay", il); + + // Attention scores q @ k^T with decay + // For GDA: attn_kq[j,i] = sum_k(q[j,k] * exp(g[j] - g[i]) * k[i,k]) + // For KDA: attn_kq[j,i] = sum_k(q[j,k] * exp(g[j,k] - g[i,k]) * k[i,k]) + ggml_tensor * attn_kq; + if (is_kda) { + // KDA: Same approach as k_decay - use decay_mask with S_k in leading dim + const int64_t CHB = n_chunks * H_k * n_seqs; + + // Rebuild decay mask (same structure as k_decay) + ggml_tensor * gcs = ggml_reshape_3d(ctx0, g_cumsum_t, chunk_size, S_k, CHB); + ggml_tensor * gcs_i = ggml_reshape_4d(ctx0, gcs, chunk_size, 1, S_k, CHB); + ggml_tensor * gcs_j = ggml_reshape_4d(ctx0, gcs, 1, chunk_size, S_k, CHB); + ggml_tensor * gcs_j_bc = ggml_repeat_4d(ctx0, gcs_j, chunk_size, chunk_size, S_k, CHB); + ggml_tensor * decay_mask_kq = ggml_sub(ctx0, gcs_j_bc, gcs_i); + + decay_mask_kq = ggml_mul(ctx0, decay_mask_kq, diag_mask); + decay_mask_kq = ggml_exp(ctx0, decay_mask_kq); + decay_mask_kq = ggml_mul(ctx0, decay_mask_kq, diag_mask); + + // Permute to [S_k, chunk_size_j, chunk_size_i, CHB] + decay_mask_kq = ggml_cont_4d(ctx0, ggml_permute(ctx0, decay_mask_kq, 2, 1, 0, 3), S_k, chunk_size, chunk_size, CHB); + + // q_j: indexed at position j, k_i: indexed at position i + ggml_tensor * q_j = ggml_reshape_4d(ctx0, q, S_k, chunk_size, 1, CHB); + ggml_tensor * k_i = ggml_reshape_4d(ctx0, k, S_k, 1, chunk_size, CHB); + + // decay_q_j[s,j,i,b] = decay[s,j,i,b] * q[s,j,b] + ggml_tensor * decay_q_j = ggml_mul(ctx0, decay_mask_kq, q_j); + + // mul_mat sums over S_k + attn_kq = ggml_mul_mat(ctx0, decay_q_j, k_i); + attn_kq = ggml_cont(ctx0, ggml_transpose(ctx0, ggml_reshape_4d(ctx0, attn_kq, chunk_size, chunk_size, n_chunks, H_k * n_seqs))); + } else { + // GDA: Use decay mask + attn_kq = ggml_mul_mat(ctx0, k, q); + attn_kq = ggml_mul(ctx0, attn_kq, decay_mask); + attn_kq = ggml_mul(ctx0, attn_kq, diag_mask); + } + cb(attn_kq, "attn_kq", il); + + // Compute g_last and g_diff for state updates + ggml_tensor * g_last; + ggml_tensor * g_diff_exp; + ggml_tensor * g_last_exp; + + if (is_kda) { + // KDA: g_cumsum [S_k, chunk_size, n_chunks, H_k * n_seqs] + // Get last element along chunk_size dimension (ne[1]) + g_last = ggml_view_4d(ctx0, g_cumsum, + g_cumsum->ne[0], 1, g_cumsum->ne[2], g_cumsum->ne[3], + g_cumsum->nb[1], g_cumsum->nb[2], g_cumsum->nb[3], + (g_cumsum->ne[1] - 1) * g_cumsum->nb[1]); + g_last = ggml_cont(ctx0, g_last); + g_last_exp = ggml_exp(ctx0, g_last); + + // g_diff = g_last - g_cumsum + ggml_tensor * g_last_broadcast = ggml_repeat_4d(ctx0, g_last, + g_cumsum->ne[0], g_cumsum->ne[1], g_cumsum->ne[2], g_cumsum->ne[3]); + ggml_tensor * g_diff = ggml_sub(ctx0, g_last_broadcast, g_cumsum); + g_diff_exp = ggml_exp(ctx0, g_diff); + } else { + // GDA: g_cumsum [chunk_size, 1, n_chunks, H_k * n_seqs] + g_last = ggml_view_4d(ctx0, g_cumsum, + 1, 1, g_cumsum->ne[2], g_cumsum->ne[3], + g_cumsum->nb[1], g_cumsum->nb[2], g_cumsum->nb[3], + (g_cumsum->ne[0] - 1) * ggml_element_size(g_cumsum)); + g_last = ggml_cont(ctx0, g_last); + g_last_exp = ggml_exp(ctx0, g_last); + + ggml_tensor * g_diff = ggml_neg(ctx0, ggml_sub(ctx0, g_cumsum, g_last)); + g_diff_exp = ggml_exp(ctx0, g_diff); + } + + cb(g_last, "g_last", il); + cb(g_last_exp, "g_last_exp", il); + + ggml_tensor * key_gdiff = ggml_mul(ctx0, k, g_diff_exp); + cb(key_gdiff, "key_gdiff", il); + + // Process chunks + ggml_tensor * new_state = state_reshaped; + ggml_tensor * core_attn_out = nullptr; + + for (int64_t chunk = 0; chunk < n_chunks; chunk++) { + ggml_tensor * q_chunk = get_slice_2d(ctx0, q, chunk); + ggml_tensor * v_chunk = get_slice_2d(ctx0, v, chunk); + ggml_tensor * k_cumdecay_chunk = get_slice_2d(ctx0, k_cumdecay, chunk); + ggml_tensor * attn_chunk = get_slice_2d(ctx0, attn_kq, chunk); + ggml_tensor * gexp_chunk = get_slice_2d(ctx0, g_exp_pos, chunk); + + cb(attn_chunk, "attn_chunk", il); + + ggml_tensor * state_t = ggml_cont_4d(ctx0, ggml_permute(ctx0, new_state, 1, 0, 2, 3), + S_v, S_v, 1, H_v * n_seqs); + + // v_prime = k_cumdecay @ state + ggml_tensor * v_prime = ggml_mul_mat(ctx0, state_t, k_cumdecay_chunk); + cb(v_prime, "v_prime_chunk", il); + + // v_new = v - v_prime + ggml_tensor * v_new = ggml_sub(ctx0, ggml_repeat(ctx0, v_chunk, v_prime), v_prime); + ggml_tensor * v_new_t = ggml_cont(ctx0, ggml_transpose(ctx0, v_new)); + cb(v_new, "v_new_chunk", il); + + // attn_inter = (q * g.exp()) @ state + ggml_tensor * q_g_exp = ggml_mul(ctx0, q_chunk, gexp_chunk); + ggml_tensor * attn_inter = ggml_mul_mat(ctx0, state_t, q_g_exp); + cb(attn_inter, "attn_inter_chunk", il); + + // output = attn_inter + attn @ v_new + ggml_tensor * v_attn = ggml_mul_mat(ctx0, v_new_t, attn_chunk); + cb(v_attn, "v_attn_chunk", il); + + ggml_tensor * core_attn_out_chunk = ggml_add(ctx0, attn_inter, v_attn); + cb(core_attn_out_chunk, "core_attn_out_chunk", il); + + core_attn_out = core_attn_out == nullptr + ? core_attn_out_chunk + : ggml_concat(ctx0, core_attn_out, core_attn_out_chunk, 2); + + // State update: state = state * g_last_exp + key_gdiff^T @ v_new + ggml_tensor * k_gdiff = ggml_cont(ctx0, get_slice_2d(ctx0, key_gdiff, chunk)); + ggml_tensor * kgdmulvnew = ggml_mul_mat(ctx0, v_new_t, ggml_cont(ctx0, ggml_transpose(ctx0, k_gdiff))); + + ggml_tensor * gexp_last_chunk = ggml_cont(ctx0, get_slice_2d(ctx0, g_last_exp, chunk)); + + if (is_kda) { + // KDA: g_last_exp [S_k, 1, n_chunks, H_k * n_seqs] + // State: [S_v, S_v, H_v, n_seqs] + // Need to reshape g_last_exp to broadcast correctly over V dimension only + gexp_last_chunk = ggml_reshape_4d(ctx0, gexp_last_chunk, + 1, gexp_last_chunk->ne[0], H_v, n_seqs); // [1, S_k, H_v, n_seqs] + // Transpose to [S_k, 1, H_v, n_seqs] then broadcast + gexp_last_chunk = ggml_cont(ctx0, ggml_permute(ctx0, gexp_last_chunk, 1, 0, 2, 3)); + } else { + // GDA: g_last_exp [1, 1, n_chunks, H_k * n_seqs] + // Broadcasts over both K and V dimensions + gexp_last_chunk = ggml_reshape_4d(ctx0, gexp_last_chunk, + gexp_last_chunk->ne[0], gexp_last_chunk->ne[1], H_v, n_seqs); + } + + new_state = ggml_add(ctx0, + ggml_mul(ctx0, new_state, gexp_last_chunk), + ggml_reshape_4d(ctx0, kgdmulvnew, kgdmulvnew->ne[0], kgdmulvnew->ne[1], H_v, n_seqs)); + } + + // Truncate padding and permute back + ggml_tensor * output_tokens = ggml_view_4d(ctx0, core_attn_out, + S_v, n_tokens, H_v, n_seqs, + ggml_row_size(core_attn_out->type, S_v), + ggml_row_size(core_attn_out->type, S_v * chunk_size * n_chunks), + ggml_row_size(core_attn_out->type, S_v * chunk_size * n_chunks * H_v), 0); + output_tokens = ggml_cont(ctx0, output_tokens); + + cb(output_tokens, "output_tokens", il); + + output_tokens = ggml_permute(ctx0, output_tokens, 0, 2, 1, 3); + output_tokens = ggml_cont(ctx0, output_tokens); + + return {output_tokens, new_state}; +} + + +/** + * Unified autoregressive Delta Net implementation (single token processing). + * + * This implementation uses matrix multiplication instead of elementwise operations + summation, + * which is more efficient and mathematically equivalent. See inline comments for equivalences. + * + * Input tensor format matches qwen3next conventions: + * @param q Query tensor [S_k, H_k, 1, n_seqs] + * @param k Key tensor [S_k, H_k, 1, n_seqs] + * @param v Value tensor [S_v, H_v, 1, n_seqs] + * @param g Gate tensor: + * GDA: [H_v, 1, n_seqs] + * KDA: [S_k, H_v, 1, n_seqs] + * @param beta Beta tensor [H_v, 1, 1, n_seqs] + * @param state State tensor [S_v, S_v * H_v, 1, n_seqs] + * @param il Layer index (for debugging callbacks) + * @param eps_norm Epsilon for L2 normalization + * + * @return Pair of (output_tokens, new_state) + */ +std::pair llm_graph_context_delta::build_delta_net_unified_autoregressive( + ggml_context * ctx0, + ggml_tensor * q, + ggml_tensor * k, + ggml_tensor * v, + ggml_tensor * g, + ggml_tensor * beta, + ggml_tensor * state, + int il, + float eps_norm) { + + // Input format: [S, H, n_tokens, n_seqs] (matching qwen3next convention) + 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); // Autoregressive mode is for single token + + // Detect KDA vs GDA based on g's shape + // GDA: g has shape [H_v, 1, n_seqs] or [H_v, n_tokens, n_seqs] + // KDA: g has shape [S_k, H_v, 1, n_seqs] or [S_k, H_v, n_tokens, n_seqs] + const bool is_kda = (g->ne[0] == S_k && g->ne[1] == H_v); + + // Validate shapes + GGML_ASSERT(v->ne[2] == n_tokens); + GGML_ASSERT(k->ne[2] == n_tokens); + GGML_ASSERT(state->ne[0] == S_v && state->ne[1] == S_v && state->ne[2] == H_v && state->ne[3] == n_seqs); + 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(beta->ne[0] == H_v && beta->ne[2] == n_tokens && beta->ne[3] == n_seqs); + GGML_ASSERT(H_k == H_v); + + if (is_kda) { + GGML_ASSERT(g->ne[0] == S_k && g->ne[1] == H_v); + } else { + GGML_ASSERT(g->ne[0] == H_v); + } + + // L2 normalize q and k + q = ggml_l2_norm(ctx0, q, eps_norm); + k = ggml_l2_norm(ctx0, k, eps_norm); + + const float scale = 1.0f / sqrtf((float)S_v); + q = ggml_scale(ctx0, q, scale); + beta = ggml_sigmoid(ctx0, beta); + + cb(q, "q_in", il); + cb(k, "k_in", il); + cb(v, "v_in", il); + cb(beta, "beta_in", il); + cb(g, "g_in", il); + + // Reshape g and beta for broadcasting + ggml_tensor * g_t; + ggml_tensor * beta_t; + + if (is_kda) { + // KDA: g [S_k, H_v, 1, n_seqs] -> [S_k, 1, H_k, n_seqs] + // For state multiplication, need [1, S_k, H_v, n_seqs] to broadcast over V only + g_t = ggml_reshape_4d(ctx0, g, S_k, 1, H_k, n_seqs); + } else { + // GDA: g [H_v, 1, n_seqs] -> [1, 1, H_k, n_seqs] + // For state multiplication, broadcasts over both K and V + g_t = ggml_reshape_4d(ctx0, ggml_transpose(ctx0, g), 1, 1, H_k, n_seqs); + } + + beta_t = ggml_reshape_4d(ctx0, ggml_transpose(ctx0, beta), 1, 1, H_k, n_seqs); + + // Apply exponential to g_t + g_t = ggml_exp(ctx0, g_t); + + // State decay: state = state * exp(g) + if (is_kda) { + // KDA: g_t [S_k, 1, H_k, n_seqs], state [S_v, S_v, H_v, n_seqs] + // Need to broadcast g_t over V dimension (ne[0] of state) + // Permute g_t to [1, S_k, H_k, n_seqs] for correct broadcasting + ggml_tensor * g_broadcast = ggml_cont(ctx0, ggml_permute(ctx0, g_t, 1, 0, 2, 3)); + state = ggml_mul(ctx0, state, g_broadcast); + } else { + // GDA: g_t [1, 1, H_k, n_seqs] broadcasts over both dimensions + state = ggml_mul(ctx0, state, g_t); + } + + // Equivalence to previous version: + // Previous: kv_mem = sum_k(state * k) using elementwise mult + sum_rows + // Current: k_state = state_t @ k_t using matrix multiplication + // These are equivalent because: sum_k(A * B) = A @ B when dimensions align + ggml_tensor * state_t = ggml_cont(ctx0, ggml_transpose(ctx0, state)); + ggml_tensor * k_t = ggml_reshape_4d(ctx0, k, S_k, 1, H_k, n_seqs); + ggml_tensor * k_state = ggml_mul_mat(ctx0, state_t, k_t); + + // v_diff = v - k_state (equivalent to v - kv_mem in previous version) + ggml_tensor * v_t = ggml_reshape_4d(ctx0, v, S_v, 1, H_v, n_seqs); + ggml_tensor * v_diff = ggml_sub(ctx0, v_t, k_state); + ggml_tensor * k_beta = ggml_mul(ctx0, k_t, beta_t); + + // Equivalence to previous version: + // Previous: state += k.unsqueeze(-1) * delta where delta = (v - kv_mem) * beta + // Current: state += v_diff^T @ k_beta^T using matrix multiplication + // These are equivalent because: outer_product(k, v_diff * beta) = v_diff^T @ k^T + state = ggml_add(ctx0, state, ggml_mul_mat(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, v_diff)), ggml_cont(ctx0, ggml_transpose(ctx0, k_beta)))); + + // Equivalence to previous version: + // Previous: core_attn_out = sum_k(state * q) using elementwise mult + sum_rows + // Current: core_attn_out = state_t @ q using matrix multiplication + // These are equivalent because: sum_k(A * B) = A @ B when dimensions align + q = ggml_reshape_4d(ctx0, q, S_k, 1, H_k, n_seqs); + state_t = ggml_cont(ctx0, ggml_transpose(ctx0, state)); + ggml_tensor * core_attn_out = ggml_mul_mat(ctx0, state_t, q); + // core_attn_out should be [S_v, 1, H_v, n_seqs] after this + cb(core_attn_out, "output_tokens", il); + cb(state, "new_state", il); + + return {core_attn_out, state}; +} + + +/** + * Main entry point that dispatches to chunked or autoregressive based on n_tokens. + * + * Input tensor format matches qwen3next conventions: + * @param q Query tensor [S_k, H_k, n_tokens, n_seqs] + * @param k Key tensor [S_k, H_k, n_tokens, n_seqs] + * @param v Value tensor [S_v, H_v, n_tokens, n_seqs] + * @param g Gate tensor (GDA: [H_v, n_tokens, n_seqs], KDA: [S_k, H_v, n_tokens, n_seqs]) + * @param beta Beta tensor [H_v, 1, n_tokens, n_seqs] + * @param state State tensor [S_v, S_v * H_v, 1, n_seqs] + */ +std::pair llm_graph_context_delta::build_delta_net_unified( + ggml_context * ctx0, + ggml_tensor * q, + ggml_tensor * k, + ggml_tensor * v, + ggml_tensor * g, + ggml_tensor * beta, + ggml_tensor * state, + ggml_tensor * causal_mask, + ggml_tensor * identity, + ggml_tensor * diag_mask, + int il, + int64_t chunk_size, + float eps_norm) { + + // Input format: [S, H, n_tokens, n_seqs] (matching qwen3next convention) + const int64_t n_tokens = q->ne[2]; + + if (n_tokens == 1) { + return build_delta_net_unified_autoregressive( + ctx0, q, k, v, g, beta, state, il, eps_norm); + } + return build_delta_net_unified_chunking( + ctx0, q, k, v, g, beta, state, causal_mask, identity, diag_mask, + il, chunk_size, eps_norm); +} diff --git a/src/models/kimi-linear.cpp b/src/models/kimi-linear.cpp index 0f037d1a39..d9ee698075 100644 --- a/src/models/kimi-linear.cpp +++ b/src/models/kimi-linear.cpp @@ -1,5 +1,4 @@ #include "models.h" -#include "ggml.h" #define CHUNK_SIZE 64 diff --git a/src/models/models.h b/src/models/models.h index cfcbb9aaa5..2a750c168e 100644 --- a/src/models/models.h +++ b/src/models/models.h @@ -17,6 +17,53 @@ struct llm_graph_context_mamba : public llm_graph_context { }; +struct llm_graph_context_delta : public llm_graph_context_mamba { + llm_graph_context_delta(const llm_graph_params & params); + + virtual ~llm_graph_context_delta() = default; + + std::pair build_delta_net_unified_chunking( + ggml_context * ctx0, + ggml_tensor * q, + ggml_tensor * k, + ggml_tensor * v, + ggml_tensor * g, + ggml_tensor * beta, + ggml_tensor * state, + ggml_tensor * causal_mask, + ggml_tensor * identity, + ggml_tensor * diag_mask, + int il, + int64_t chunk_size, + float eps_norm); + + std::pair build_delta_net_unified_autoregressive( + ggml_context * ctx0, + ggml_tensor * q, + ggml_tensor * k, + ggml_tensor * v, + ggml_tensor * g, + ggml_tensor * beta, + ggml_tensor * state, + int il, + float eps_norm); + + std::pair build_delta_net_unified( + ggml_context * ctx0, + ggml_tensor * q, + ggml_tensor * k, + ggml_tensor * v, + ggml_tensor * g, + ggml_tensor * beta, + ggml_tensor * state, + ggml_tensor * causal_mask, + ggml_tensor * identity, + ggml_tensor * diag_mask, + int il, + int64_t chunk_size, + float eps_norm); +}; + // Base class for RWKV-related models struct llm_build_rwkv6_base : public llm_graph_context { const llama_model & model; @@ -476,7 +523,7 @@ struct llm_build_qwen3vl : public llm_graph_context { struct llm_build_qwen3vlmoe : public llm_graph_context { llm_build_qwen3vlmoe(const llama_model & model, const llm_graph_params & params); }; -struct llm_build_qwen3next : public llm_graph_context_mamba { +struct llm_build_qwen3next : public llm_graph_context_delta { llm_build_qwen3next(const llama_model & model, const llm_graph_params & params); private: ggml_tensor * build_layer_attn( @@ -534,6 +581,59 @@ private: const llama_model & model; }; +struct llm_build_qwen3_5 : public llm_graph_context_delta { + llm_build_qwen3_5(const llama_model & model, const llm_graph_params & params); + +protected: + // Tag type for subclass constructors that need to call build_graph() themselves + // (to ensure virtual dispatch works correctly) + struct defer_graph_build_t {}; + + llm_build_qwen3_5(const llama_model & model, const llm_graph_params & params, defer_graph_build_t); + + void build_graph(); + + virtual ggml_tensor * build_layer_ffn( + ggml_tensor * cur, + int il); + + const llama_model & model; + +private: + ggml_tensor * build_layer_attn( + llm_graph_input_attn_kv * inp_attn, + ggml_tensor * cur, + ggml_tensor * inp_pos, + int il); + + ggml_tensor * build_layer_attn_linear( + llm_graph_input_rs * inp, + ggml_tensor * cur, + ggml_tensor * causal_mask, + ggml_tensor * identity, + ggml_tensor * diag_mask, + int il); + + ggml_tensor * build_norm_gated( + ggml_tensor * input, + ggml_tensor * weights, + ggml_tensor * gate, + int layer); + + std::pair build_qkvz( + ggml_tensor * input, + int il); +}; + +struct llm_build_qwen3_5_moe : public llm_build_qwen3_5 { + llm_build_qwen3_5_moe(const llama_model & model, const llm_graph_params & params); + +protected: + ggml_tensor * build_layer_ffn( + ggml_tensor * cur, + int il) override; +}; + struct llm_build_qwen : public llm_graph_context { llm_build_qwen(const llama_model & model, const llm_graph_params & params); }; diff --git a/src/models/qwen3-5.cpp b/src/models/qwen3-5.cpp new file mode 100644 index 0000000000..0947299d73 --- /dev/null +++ b/src/models/qwen3-5.cpp @@ -0,0 +1,421 @@ +#include "models.h" + +#define CHUNK_SIZE 64 + +llm_build_qwen3_5::llm_build_qwen3_5(const llama_model & model, const llm_graph_params & params) : + llm_graph_context_delta(params), model(model) { + build_graph(); +} + +// virtual call in constructor fix +llm_build_qwen3_5::llm_build_qwen3_5(const llama_model & model, const llm_graph_params & params, defer_graph_build_t /*tag*/) : + llm_graph_context_delta(params), model(model) { +} + +void llm_build_qwen3_5::build_graph() { + ggml_tensor * cur; + ggml_tensor * inpL; + + inpL = build_inp_embd(model.tok_embd); + cb(inpL, "model.embed_tokens", -1); + + auto * inp = build_inp_mem_hybrid(); + + ggml_tensor * inp_pos = build_inp_pos(); + ggml_tensor * inp_out_ids = build_inp_out_ids(); + + ggml_tensor * causal_mask = + ggml_tri(ctx0, ggml_fill(ctx0, ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, CHUNK_SIZE, CHUNK_SIZE), 1.0f), + GGML_TRI_TYPE_LOWER); + + ggml_tensor * identity = ggml_diag(ctx0, ggml_fill(ctx0, ggml_new_tensor_1d(ctx0, GGML_TYPE_F32, CHUNK_SIZE), 1.0f)); + ggml_tensor * diag_mask = ggml_add(ctx0, causal_mask, identity); + + ggml_build_forward_expand(gf, causal_mask); + ggml_build_forward_expand(gf, identity); + ggml_build_forward_expand(gf, diag_mask); + + for (int il = 0; il < n_layer; ++il) { + ggml_tensor * inpSA = inpL; + + cur = build_norm(inpL, model.layers[il].attn_norm, nullptr, LLM_NORM_RMS, il); + cb(cur, "attn_norm", il); + + if (hparams.is_recurrent(il)) { + cur = build_layer_attn_linear(inp->get_recr(), cur, causal_mask, identity, diag_mask, il); + } else { + cur = build_layer_attn(inp->get_attn(), cur, inp_pos, 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); + } + + cur = ggml_add(ctx0, cur, inpSA); + cb(cur, "attn_residual", il); + + ggml_tensor * ffn_residual = cur; + + ggml_tensor * attn_post_norm = build_norm(cur, model.layers[il].attn_post_norm, nullptr, LLM_NORM_RMS, il); + cb(attn_post_norm, "attn_post_norm", il); + + cur = build_layer_ffn(attn_post_norm, il); + cb(cur, "ffn_out", il); + + cur = ggml_add(ctx0, cur, ffn_residual); + cb(cur, "post_ffn", il); + + inpL = cur; + } + cur = inpL; + + cur = build_norm(cur, model.output_norm, nullptr, LLM_NORM_RMS, -1); + + cb(cur, "result_norm", -1); + res->t_embd = cur; + + cur = build_lora_mm(model.output, cur); + + cb(cur, "result_output", -1); + res->t_logits = cur; + + ggml_build_forward_expand(gf, cur); +} + +ggml_tensor * llm_build_qwen3_5::build_norm_gated( + ggml_tensor * input, + ggml_tensor * weights, + ggml_tensor * gate, + int layer) { + ggml_tensor * normalized = build_norm(input, weights, nullptr, LLM_NORM_RMS, layer); + ggml_tensor * gated_silu = ggml_silu(ctx0, gate); + + return ggml_mul(ctx0, normalized, gated_silu); +} + +ggml_tensor * llm_build_qwen3_5::build_layer_attn( + llm_graph_input_attn_kv * inp, + ggml_tensor * cur, + ggml_tensor * inp_pos, + int il) { + const int64_t n_embd_head = hparams.n_embd_head_v; + GGML_ASSERT(n_embd_head == hparams.n_embd_head_k); + + ggml_tensor * Qcur_full = build_lora_mm(model.layers[il].wq, cur); // [ (n_embd_head * 2) * n_head, n_tokens ] + cb(Qcur_full, "Qcur_full", il); + + ggml_tensor * Qcur = ggml_view_3d(ctx0, Qcur_full, n_embd_head, n_head, n_tokens, + ggml_element_size(Qcur_full) * n_embd_head * 2, + ggml_element_size(Qcur_full) * n_embd_head * 2 * n_head, 0); + cb(Qcur, "Qcur_reshaped", il); + + Qcur = build_norm(Qcur, model.layers[il].attn_q_norm, nullptr, LLM_NORM_RMS, il); + cb(Qcur, "Qcur_normed", 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); + + Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens); + Kcur = build_norm(Kcur, model.layers[il].attn_k_norm, nullptr, LLM_NORM_RMS, il); + cb(Kcur, "Kcur_normed", il); + + ggml_tensor * gate = ggml_view_3d(ctx0, Qcur_full, n_embd_head, n_head, n_tokens, + ggml_element_size(Qcur_full) * n_embd_head * 2, + ggml_element_size(Qcur_full) * n_embd_head * 2 * n_head, + ggml_element_size(Qcur_full) * n_embd_head); + gate = ggml_cont_2d(ctx0, gate, n_embd_head * n_head, n_tokens); + cb(gate, "gate_reshaped", il); + + 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); + + const float kq_scale = hparams.f_attention_scale == 0.0f ? 1.0f / sqrtf(float(n_embd_head)) : hparams.f_attention_scale; + + cur = build_attn(inp, + nullptr, nullptr, + Qcur, Kcur, Vcur, nullptr, nullptr, nullptr, kq_scale, il); + cb(cur, "attn_pregate", il); + + ggml_tensor * gate_sigmoid = ggml_sigmoid(ctx0, gate); + cb(gate_sigmoid, "gate_sigmoid", il); + + cur = ggml_mul(ctx0, cur, gate_sigmoid); + cb(cur, "attn_gated", il); + + cur = build_lora_mm(model.layers[il].wo, cur); + cb(cur, "attn_output", il); + + return cur; +} + +std::pair llm_build_qwen3_5::build_qkvz( + ggml_tensor * input, + int il) { + const int64_t d_inner = hparams.ssm_d_inner; + const int64_t n_seqs = ubatch.n_seqs; + const int64_t head_k_dim = hparams.ssm_d_state; + const int64_t num_k_heads = hparams.ssm_n_group; + const int64_t num_v_heads = hparams.ssm_dt_rank; + const int64_t head_v_dim = d_inner / num_v_heads; + const int64_t n_seq_tokens = ubatch.n_seq_tokens; + + if (model.layers[il].wqkv) { + ggml_tensor * qkv_mixed = build_lora_mm(model.layers[il].wqkv, input); + qkv_mixed = ggml_reshape_3d(ctx0, qkv_mixed, qkv_mixed->ne[0], n_seq_tokens, n_seqs); + cb(qkv_mixed, "linear_attn_qkv_mixed", il); + + ggml_tensor * z = build_lora_mm(model.layers[il].wqkv_gate, input); + cb(z, "z", il); + + return { qkv_mixed, z }; + + } + // legacy path for combined in_proj_qkvz + ggml_tensor * mixed_qkvz = build_lora_mm(model.layers[il].ssm_in, input); + cb(mixed_qkvz, "linear_attn_mixed_qkvz", il); + + int64_t qkvz_new_dim = 2 * head_k_dim + 2 * head_v_dim * (num_v_heads / num_k_heads); + ggml_tensor * mixed_qkvz_reshaped = ggml_reshape_4d(ctx0, mixed_qkvz, qkvz_new_dim, num_k_heads, n_seq_tokens, n_seqs); + + int64_t split_sizes_qkvz[4] = { + head_k_dim, + head_k_dim, + head_v_dim * num_v_heads / num_k_heads, + head_v_dim * num_v_heads / num_k_heads + }; + + ggml_tensor * query = + ggml_view_4d(ctx0, mixed_qkvz_reshaped, split_sizes_qkvz[0], num_k_heads, n_seq_tokens, n_seqs, + mixed_qkvz_reshaped->nb[1], mixed_qkvz_reshaped->nb[2], mixed_qkvz_reshaped->nb[3], 0); + cb(query, "q", il); + + ggml_tensor * key = ggml_view_4d(ctx0, mixed_qkvz_reshaped, split_sizes_qkvz[1], num_k_heads, n_seq_tokens, n_seqs, + mixed_qkvz_reshaped->nb[1], mixed_qkvz_reshaped->nb[2], mixed_qkvz_reshaped->nb[3], + split_sizes_qkvz[0] * ggml_element_size(mixed_qkvz_reshaped)); + cb(key, "k", il); + + ggml_tensor * value = + ggml_view_4d(ctx0, mixed_qkvz_reshaped, split_sizes_qkvz[2], num_k_heads, n_seq_tokens, n_seqs, + mixed_qkvz_reshaped->nb[1], mixed_qkvz_reshaped->nb[2], mixed_qkvz_reshaped->nb[3], + (split_sizes_qkvz[0] + split_sizes_qkvz[1]) * ggml_element_size(mixed_qkvz_reshaped)); + cb(value, "v", il); + + ggml_tensor * z = ggml_view_4d(ctx0, mixed_qkvz_reshaped, split_sizes_qkvz[3], num_k_heads, n_seq_tokens, n_seqs, + mixed_qkvz_reshaped->nb[1], mixed_qkvz_reshaped->nb[2], mixed_qkvz_reshaped->nb[3], + (split_sizes_qkvz[0] + split_sizes_qkvz[1] + split_sizes_qkvz[2]) * ggml_element_size(mixed_qkvz_reshaped)); + z = ggml_cont(ctx0, z); + cb(z, "z", il); + + ggml_tensor * query_flat = ggml_reshape_3d(ctx0, query, head_k_dim * num_k_heads, n_seq_tokens, n_seqs); + cb(query_flat, "query_flat", il); + + ggml_tensor * key_flat = ggml_reshape_3d(ctx0, key, head_k_dim * num_k_heads, n_seq_tokens, n_seqs); + cb(key_flat, "key_flat", il); + + ggml_tensor * value_flat = ggml_reshape_3d(ctx0, value, head_v_dim * num_v_heads, n_seq_tokens, n_seqs); + cb(value_flat, "value_flat", il); + + ggml_tensor * qkv_mixed = ggml_concat(ctx0, query_flat, key_flat, 0); + qkv_mixed = ggml_concat(ctx0, qkv_mixed, value_flat, 0); + cb(qkv_mixed, "qkv_mixed", il); + + return { qkv_mixed, z }; +} + +ggml_tensor * llm_build_qwen3_5::build_layer_attn_linear( + llm_graph_input_rs * inp, + ggml_tensor * cur, + ggml_tensor * causal_mask, + ggml_tensor * identity, + ggml_tensor * diag_mask, + int il) { + const auto * mctx_cur = inp->mctx; + + const int64_t d_inner = hparams.ssm_d_inner; + const int64_t n_seqs = ubatch.n_seqs; + const int64_t head_k_dim = hparams.ssm_d_state; + const int64_t num_k_heads = hparams.ssm_n_group; + const int64_t num_v_heads = hparams.ssm_dt_rank; + const int64_t head_v_dim = d_inner / num_v_heads; + const int64_t n_seq_tokens = ubatch.n_seq_tokens; + + const auto kv_head = mctx_cur->get_head(); + + GGML_ASSERT(n_seqs != 0); + GGML_ASSERT(ubatch.equal_seqs()); + GGML_ASSERT(ubatch.n_tokens == n_seq_tokens * n_seqs); + + auto qkvz = build_qkvz(cur, il); + ggml_tensor * qkv_mixed = qkvz.first; + ggml_tensor * z = qkvz.second; + + ggml_tensor * mixed_ba = build_lora_mm(model.layers[il].ssm_beta_alpha, cur); + cb(mixed_ba, "linear_attn_mixed_ba", il); + + int64_t ba_new_dim = 2 * num_v_heads / num_k_heads; + ggml_tensor * mixed_ba_reshaped = ggml_reshape_4d(ctx0, mixed_ba, ba_new_dim, num_k_heads, n_seq_tokens, n_seqs); + + int64_t split_sizes_ba[2] = { + num_v_heads / num_k_heads, + num_v_heads / num_k_heads + }; + + ggml_tensor * b = ggml_view_4d(ctx0, mixed_ba_reshaped, split_sizes_ba[0], num_k_heads, n_seq_tokens, n_seqs, + mixed_ba_reshaped->nb[1], mixed_ba_reshaped->nb[2], mixed_ba_reshaped->nb[3], 0); + cb(b, "b", il); + + ggml_tensor * a = ggml_view_4d(ctx0, mixed_ba_reshaped, split_sizes_ba[1], num_k_heads, n_seq_tokens, n_seqs, + mixed_ba_reshaped->nb[1], mixed_ba_reshaped->nb[2], mixed_ba_reshaped->nb[3], + split_sizes_ba[0] * ggml_element_size(mixed_ba_reshaped)); + cb(a, "a", il); + + ggml_tensor * beta = ggml_cont_4d(ctx0, b, num_v_heads, 1, n_seq_tokens, n_seqs); + + ggml_tensor * alpha = ggml_cont_3d(ctx0, a, num_v_heads, n_seq_tokens, n_seqs); + + ggml_tensor * alpha_biased = ggml_add(ctx0, alpha, model.layers[il].ssm_dt); + ggml_tensor * alpha_softplus = ggml_softplus(ctx0, alpha_biased); + cb(alpha_softplus, "a_softplus", il); + ggml_tensor * gate = ggml_mul(ctx0, alpha_softplus, model.layers[il].ssm_a); + cb(gate, "gate", il); + + ggml_tensor * conv_states_all = mctx_cur->get_r_l(il); + ggml_tensor * ssm_states_all = mctx_cur->get_s_l(il); + + ggml_tensor * conv_states = build_rs(inp, conv_states_all, hparams.n_embd_r(), n_seqs); + cb(conv_states, "conv_states", il); + + ggml_tensor * conv_kernel = model.layers[il].ssm_conv1d; + const int64_t conv_kernel_size = conv_kernel->ne[0]; + const int64_t conv_channels = d_inner + 2 * hparams.ssm_n_group * hparams.ssm_d_state; + conv_states = ggml_reshape_3d(ctx0, conv_states, conv_kernel_size - 1, conv_channels, n_seqs); + cb(conv_states, "conv_states_reshaped", il); + + qkv_mixed = ggml_permute(ctx0, qkv_mixed, 1, 0, 2, 3); + cb(qkv_mixed, "qkv_mixed_permuted", il); + + ggml_tensor * conv_input = ggml_concat(ctx0, conv_states, qkv_mixed, 0); + cb(conv_input, "conv_input", il); + + ggml_tensor * last_conv_states = + ggml_view_3d(ctx0, conv_input, conv_kernel_size - 1, conv_channels, n_seqs, conv_input->nb[1], + conv_input->nb[2], (conv_input->ne[0] - conv_states->ne[0]) * ggml_element_size(conv_input)); + cb(last_conv_states, "last_conv_states", il); + + ggml_tensor * state_update_target = + ggml_view_1d(ctx0, conv_states_all, (conv_kernel_size - 1) * conv_channels * n_seqs, + kv_head * (conv_kernel_size - 1) * conv_channels * ggml_element_size(conv_states_all)); + cb(state_update_target, "state_update_target", il); + + ggml_build_forward_expand(gf, ggml_cpy(ctx0, last_conv_states, state_update_target)); + cb(conv_states_all, "conv_states_updated", il); + + ggml_tensor * conv_output_proper = ggml_ssm_conv(ctx0, conv_input, conv_kernel); + cb(conv_output_proper, "conv_output_raw", il); + + ggml_tensor * conv_output_silu = ggml_silu(ctx0, conv_output_proper); + cb(conv_output_silu, "conv_output_silu", il); + + ggml_tensor * conv_qkv_mix = conv_output_silu; + + int64_t qkv_dim = head_k_dim * num_k_heads * 2 + head_v_dim * num_v_heads; + int64_t nb1_qkv = ggml_row_size(conv_qkv_mix->type, qkv_dim); + + ggml_tensor * q_conv = + ggml_view_2d(ctx0, conv_qkv_mix, head_k_dim * num_k_heads, n_seq_tokens * n_seqs, nb1_qkv, 0); + cb(q_conv, "q_conv", il); + ggml_tensor * k_conv = + ggml_view_2d(ctx0, conv_qkv_mix, head_k_dim * num_k_heads, n_seq_tokens * n_seqs, nb1_qkv, + head_k_dim * num_k_heads * ggml_element_size(conv_qkv_mix)); + cb(k_conv, "k_conv", il); + ggml_tensor * v_conv = + ggml_view_2d(ctx0, conv_qkv_mix, head_v_dim * num_v_heads, n_seq_tokens * n_seqs, nb1_qkv, + 2 * head_k_dim * num_k_heads * ggml_element_size(conv_qkv_mix)); + cb(v_conv, "v_conv", il); + + q_conv = ggml_cont_4d(ctx0, q_conv, head_k_dim, num_k_heads, n_seq_tokens, n_seqs); + k_conv = ggml_cont_4d(ctx0, k_conv, head_k_dim, num_k_heads, n_seq_tokens, n_seqs); + v_conv = ggml_cont_4d(ctx0, v_conv, head_v_dim, num_v_heads, n_seq_tokens, n_seqs); + + ggml_tensor * state = build_rs(inp, ssm_states_all, hparams.n_embd_s(), n_seqs); + state = ggml_reshape_4d(ctx0, state, head_v_dim, head_v_dim, num_v_heads, n_seqs); + cb(state, "state_predelta", il); + + if (num_k_heads != num_v_heads) { + GGML_ASSERT(num_v_heads % num_k_heads == 0); + int64_t repeat_factor = num_v_heads / num_k_heads; + + ggml_tensor * q_reshaped = ggml_reshape_3d(ctx0, q_conv, head_k_dim, 1, num_k_heads * n_seq_tokens * n_seqs); + ggml_tensor * k_reshaped = ggml_reshape_3d(ctx0, k_conv, head_k_dim, 1, num_k_heads * n_seq_tokens * n_seqs); + + ggml_tensor * q_repeated = + ggml_repeat_4d(ctx0, q_reshaped, head_k_dim, repeat_factor, num_k_heads * n_seq_tokens * n_seqs, 1); + ggml_tensor * k_repeated = + ggml_repeat_4d(ctx0, k_reshaped, head_k_dim, repeat_factor, num_k_heads * n_seq_tokens * n_seqs, 1); + + q_conv = ggml_reshape_4d(ctx0, q_repeated, head_k_dim, num_k_heads * repeat_factor, n_seq_tokens, n_seqs); + k_conv = ggml_reshape_4d(ctx0, k_repeated, head_k_dim, num_k_heads * repeat_factor, n_seq_tokens, n_seqs); + } + + cb(q_conv, "q_conv_predelta", il); + cb(k_conv, "k_conv_predelta", il); + cb(v_conv, "v_conv_predelta", il); + + std::pair attn_out = build_delta_net_unified(ctx0, q_conv, k_conv, v_conv, + gate, beta, state, causal_mask, identity, diag_mask, + il, CHUNK_SIZE, hparams.f_norm_rms_eps); + + ggml_tensor * output = attn_out.first; + ggml_tensor * new_state = attn_out.second; + cb(output, "attn_output", il); + cb(new_state, "new_state", il); + + ggml_build_forward_expand(gf, + ggml_cpy(ctx0, new_state, + ggml_view_1d(ctx0, ssm_states_all, hparams.n_embd_s() * n_seqs, + kv_head * hparams.n_embd_s() * ggml_element_size(ssm_states_all)))); + + ggml_tensor * attn_out_2d_final = ggml_reshape_2d(ctx0, output, head_v_dim, num_v_heads * n_seq_tokens * n_seqs); + + ggml_tensor * z_2d = ggml_reshape_2d(ctx0, z, head_v_dim, num_v_heads * n_seq_tokens * n_seqs); + + ggml_tensor * attn_out_norm = build_norm_gated(attn_out_2d_final, model.layers[il].ssm_norm, z_2d, il); + + ggml_tensor * final_output = ggml_reshape_3d(ctx0, attn_out_norm, head_v_dim * num_v_heads, n_seq_tokens, n_seqs); + cb(final_output, "final_output", il); + + cur = build_lora_mm(model.layers[il].ssm_out, final_output); + cb(cur, "linear_attn_out", il); + + cur = ggml_cont_2d(ctx0, cur, n_embd, n_seq_tokens * n_seqs); + return cur; +} + +ggml_tensor * llm_build_qwen3_5::build_layer_ffn(ggml_tensor * cur, const int il) { + // Qwen3.5 Dense always uses dense FFN + 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); + return cur; +} diff --git a/src/models/qwen3-5moe.cpp b/src/models/qwen3-5moe.cpp new file mode 100644 index 0000000000..a488443218 --- /dev/null +++ b/src/models/qwen3-5moe.cpp @@ -0,0 +1,52 @@ +#include "models.h" + +llm_build_qwen3_5_moe::llm_build_qwen3_5_moe(const llama_model & model, const llm_graph_params & params) : + llm_build_qwen3_5(model, params, defer_graph_build_t{}) { + build_graph(); +} + +ggml_tensor * llm_build_qwen3_5_moe::build_layer_ffn(ggml_tensor * cur, const int il) { + // Check if this is an MoE layer + if (model.layers[il].ffn_gate_inp != nullptr) { + // MoE branch + 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, + nullptr, + n_expert, n_expert_used, LLM_FFN_SILU, + true, false, 0.0, LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX, il); + cb(moe_out, "ffn_moe_out", il); + + // Add shared experts if present + if (model.layers[il].ffn_up_shexp != nullptr) { + 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); + + // Apply shared expert gating (sigmoid) + ggml_tensor * shared_gate = build_lora_mm(model.layers[il].ffn_gate_inp_shexp, cur); + cb(shared_gate, "shared_expert_gate", il); + + shared_gate = ggml_sigmoid(ctx0, shared_gate); + cb(shared_gate, "shared_expert_gate_sigmoid", il); + + ffn_shexp = ggml_mul(ctx0, ffn_shexp, shared_gate); + cb(ffn_shexp, "ffn_shexp_gated", il); + + cur = ggml_add(ctx0, moe_out, ffn_shexp); + cb(cur, "ffn_out", il); + } else { + cur = moe_out; + } + } else { + // Dense FFN branch (fallback) + cur = llm_build_qwen3_5::build_layer_ffn(cur, il); + } + return cur; +} diff --git a/src/models/qwen3next.cpp b/src/models/qwen3next.cpp index 99b1a76a48..0335f5ab76 100644 --- a/src/models/qwen3next.cpp +++ b/src/models/qwen3next.cpp @@ -1,10 +1,9 @@ -#include "ggml.h" #include "models.h" #define CHUNK_SIZE 64 llm_build_qwen3next::llm_build_qwen3next(const llama_model & model, const llm_graph_params & params) : - llm_graph_context_mamba(params), model(model) { + llm_graph_context_delta(params), model(model) { ggml_tensor * cur; ggml_tensor * inpL; @@ -86,362 +85,6 @@ llm_build_qwen3next::llm_build_qwen3next(const llama_model & model, const llm_gr ggml_build_forward_expand(gf, cur); } -// 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); -} - -std::pair llm_build_qwen3next::build_delta_net_chunking( - ggml_tensor * q, - ggml_tensor * k, - ggml_tensor * v, - ggml_tensor * g, - ggml_tensor * beta, - ggml_tensor * state, - ggml_tensor * causal_mask, - ggml_tensor * identity, - ggml_tensor * diag_mask, - 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(v->ne[2] == n_tokens); - GGML_ASSERT(k->ne[2] == n_tokens); - GGML_ASSERT(g->ne[0] == H_v && g->ne[1] == n_tokens && g->ne[2] == n_seqs); - GGML_ASSERT(beta->ne[0] == H_v && beta->ne[2] == n_tokens && beta->ne[3] == n_seqs); - GGML_ASSERT(state->ne[0] == S_v && state->ne[1] == S_v * H_v && state->ne[2] == 1 && state->ne[3] == n_seqs); - - 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(H_k == H_v); // we did a repeat to make sure this is the case - - const float eps_norm = hparams.f_norm_rms_eps; - - q = ggml_l2_norm(ctx0, q, eps_norm); - k = ggml_l2_norm(ctx0, k, eps_norm); - - const float scale = 1.0f / sqrtf(S_v); - - q = ggml_scale(ctx0, q, scale); - - beta = ggml_sigmoid(ctx0, beta); - - cb(q, "q_in", il); - cb(k, "k_in", il); - cb(v, "v_in", il); - cb(beta, "beta_in", il); - cb(g, "g_in", il); - - q = ggml_cont_4d(ctx0, ggml_permute(ctx0, q, 0, 2, 1, 3), S_v, n_tokens, H_v, n_seqs); - k = ggml_cont_4d(ctx0, ggml_permute(ctx0, k, 0, 2, 1, 3), S_v, n_tokens, H_v, n_seqs); - v = ggml_cont_4d(ctx0, ggml_permute(ctx0, v, 0, 2, 1, 3), S_v, n_tokens, H_v, n_seqs); - g = ggml_cont_4d(ctx0, ggml_permute(ctx0, g, 2, 0, 3, 1), n_tokens, 1, H_k, n_seqs); - - beta = ggml_cont(ctx0, ggml_permute(ctx0, beta, 2, 0, 1, 3)); - state = ggml_reshape_4d(ctx0, state, S_v, S_v, H_v, n_seqs); - - cb(q, "q_perm", il); - cb(k, "k_perm", il); - cb(v, "v_perm", il); - cb(beta, "beta_perm", il); - cb(g, "g_perm", il); - cb(state, "state_in", il); - - GGML_ASSERT(q->ne[1] == n_tokens && q->ne[0] == S_k && q->ne[2] == H_k && q->ne[3] == n_seqs); - GGML_ASSERT(k->ne[1] == n_tokens && k->ne[0] == S_k && k->ne[2] == H_k && k->ne[3] == n_seqs); - GGML_ASSERT(v->ne[1] == n_tokens && v->ne[0] == S_v && v->ne[2] == H_k && v->ne[3] == n_seqs); - GGML_ASSERT(beta->ne[1] == n_tokens && beta->ne[2] == H_k && beta->ne[0] == 1 && beta->ne[3] == n_seqs); - - // Do padding - const int64_t chunk_size = CHUNK_SIZE; - - const int64_t pad = (chunk_size - n_tokens % chunk_size) % chunk_size; - const int64_t n_chunks = (n_tokens + pad) / chunk_size; - - 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, pad, 0, 0, 0); - beta = ggml_pad(ctx0, beta, 0, pad, 0, 0); - - cb(q, "q_pad", il); - cb(k, "k_pad", il); - cb(v, "v_pad", il); - cb(beta, "beta_pad", il); - cb(g, "g_pad", il); - - ggml_tensor * v_beta = ggml_mul(ctx0, v, beta); - ggml_tensor * k_beta = ggml_mul(ctx0, k, beta); - - cb(v_beta, "v_beta", il); - cb(k_beta, "k_beta", il); - - q = ggml_reshape_4d(ctx0, q, S_k, chunk_size, n_chunks, H_k * n_seqs); - k = ggml_reshape_4d(ctx0, k, S_k, chunk_size, n_chunks, H_k * n_seqs); - k_beta = ggml_reshape_4d(ctx0, k_beta, S_k, chunk_size, n_chunks, H_k * n_seqs); - v = ggml_reshape_4d(ctx0, v, S_v, chunk_size, n_chunks, H_v * n_seqs); - v_beta = ggml_reshape_4d(ctx0, v_beta, S_v, chunk_size, n_chunks, H_v * n_seqs); - - g = ggml_reshape_4d(ctx0, g, chunk_size, 1, n_chunks, H_k * n_seqs); - beta = ggml_reshape_4d(ctx0, beta, 1, chunk_size, n_chunks, H_k * n_seqs); - - ggml_tensor * g_cumsum = ggml_cumsum(ctx0, g); - cb(g_cumsum, "g_cumsum", il); // shape: (chunk_size, 1, n_chunks, H_v * n_seqs) - - ggml_tensor * gcs_i = g_cumsum; // ggml_reshape_4d(ctx0, g_cumsum, chunk_size, 1, n_chunks, H_v * n_seqs); - ggml_tensor * gcs_j = ggml_reshape_4d(ctx0, g_cumsum, 1, chunk_size, n_chunks, H_v * n_seqs); - - ggml_tensor * gcs_j_broadcast = - ggml_repeat_4d(ctx0, gcs_j, chunk_size, chunk_size, n_chunks, H_v * n_seqs); - - ggml_tensor * decay_mask = ggml_sub(ctx0, gcs_j_broadcast, gcs_i); - cb(decay_mask, "decay_mask", il); // shape: (chunk_size, chunk_size, n_chunks, H_v * n_seqs) - - decay_mask = ggml_mul(ctx0, decay_mask, diag_mask); - decay_mask = ggml_exp(ctx0, decay_mask); - decay_mask = ggml_mul(ctx0, decay_mask, diag_mask); - - ggml_tensor * kmulkbeta = ggml_mul_mat(ctx0, k, k_beta); - - ggml_tensor * k_decay = ggml_mul(ctx0, kmulkbeta, decay_mask); - ggml_tensor * attn = ggml_neg(ctx0, ggml_mul(ctx0, k_decay, causal_mask)); - cb(attn, "attn_pre_solve", il); // shape: (chunk_size, chunk_size, n_chunks, H_v * n_seqs) - - ggml_tensor * attn_lower = ggml_mul(ctx0, attn, causal_mask); - ggml_tensor * lhs = ggml_sub(ctx0, ggml_repeat(ctx0, identity, attn_lower), attn_lower); - - ggml_tensor * lin_solve = ggml_solve_tri(ctx0, lhs, attn, true, true, false); - attn = ggml_mul(ctx0, lin_solve, causal_mask); - attn = ggml_add(ctx0, attn, identity); - cb(attn, "attn_solved", il); // shape: (chunk_size, chunk_size, n_chunks, H_v * n_seqs) - - v = ggml_mul_mat(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, v_beta)), attn); - - ggml_tensor * g_cumsum_t = ggml_cont(ctx0, ggml_transpose(ctx0, g_cumsum)); - ggml_tensor * gexp = ggml_exp(ctx0, g_cumsum_t); - - ggml_tensor * kbeta_gexp = ggml_mul(ctx0, k_beta, gexp); - cb(kbeta_gexp, "kbeta_gexp", il); // shape: (S_k, chunk_size, n_chunks, H_v * n_seqs) - - ggml_tensor * k_cumdecay = - ggml_cont(ctx0, ggml_transpose(ctx0, ggml_mul_mat(ctx0, attn, ggml_cont(ctx0, ggml_transpose(ctx0, kbeta_gexp))))); - cb(k_cumdecay, "k_cumdecay", il); // shape: (chunk_size, chunk_size, n_chunks, H_v * n_seqs) - - ggml_tensor * attn_kq = ggml_mul_mat(ctx0, k, q); - attn_kq = ggml_mul(ctx0, attn_kq, decay_mask); - attn_kq = ggml_mul(ctx0, attn_kq, diag_mask); - cb(attn_kq, "attn_kq", il); // shape: (chunk_size, chunk_size, n_chunks, H_v * n_seqs) - - - // 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 chunk_size dimension (ne0) - // example: [[x, y, z, ..., last], ...] -> [[last], ...] - ggml_tensor * g_last = ggml_view_4d(ctx0, g_cumsum, 1, 1, g_cumsum->ne[2], g_cumsum->ne[3], - g_cumsum->nb[1], g_cumsum->nb[2], g_cumsum->nb[3], - (g_cumsum->ne[0] - 1) * ggml_element_size(g_cumsum)); - g_last = ggml_cont(ctx0, g_last); - cb(g_last, "g_last", il); // shape: (1, 1, n_chunks, H_v * n_seqs) - - ggml_tensor * g_last_exp = ggml_exp(ctx0, g_last); - cb(g_last_exp, "g_last_exp", il); // shape: (1, 1, n_chunks, H_v * n_seqs) - - ggml_tensor * g_diff = ggml_neg(ctx0, ggml_sub(ctx0, g_cumsum, g_last)); - cb(g_diff, "g_diff", il); // shape: (chunk_size, 1, n_chunks, H_v * n_seqs) - - ggml_tensor * g_diff_exp = ggml_exp(ctx0, g_diff); - ggml_tensor * g_diff_exp_t = ggml_reshape_4d(ctx0, g_diff_exp, - 1, chunk_size, n_chunks, g_diff_exp->ne[3]); - - ggml_tensor * key_gdiff = ggml_mul(ctx0, k, g_diff_exp_t); - cb(key_gdiff, "key_gdiff", il); // shape: (S_k, chunk_size, n_chunks, H_v * n_seqs) - - ggml_tensor * key_gdiff_t = ggml_cont(ctx0, ggml_transpose(ctx0, key_gdiff)); - cb(key_gdiff_t, "key_gdiff_t", il); // shape: (chunk_size, S_k, n_chunks, H_v * n_seqs) - - - // state to be updated per chunk - ggml_tensor * new_state = state; // ggml_dup(ctx0, state); - cb(new_state, "new_state", il); // shape: (S_v, S_v, H_v, n_seqs) - - // shape after loop of chunks: (S_v, chunk_size, n_chunks, H_v * n_seqs) - ggml_tensor * core_attn_out = nullptr; - - for (int64_t chunk = 0; chunk < n_chunks; chunk++) { - // shape: (S_k, chunk_size, 1, H_k * n_seqs) - ggml_tensor * q_chunk = get_slice_2d(ctx0, q, chunk); // (no cont), next op: ggml_mul - - // shape: (S_v, chunk_size, 1, H_v * n_seqs) - ggml_tensor * v_chunk = get_slice_2d(ctx0, v, chunk); // (no cont), next op: ggml_repeat - - // shape: (chunk_size, 1, n_chunks, H_v * n_seqs) - ggml_tensor * gexp_chunk = get_slice_2d(ctx0, gexp, chunk); // (no cont), next op: ggml_mul - - // shape: (chunk_size, 1, H_v * n_seqs) - ggml_tensor * k_cumdecay_chunk = get_slice_2d(ctx0, k_cumdecay, chunk); // (no cont), next op: ggml_mul_mat - - // attn = (q_i @ k_i.transpose(-1, -2) * decay_mask[:, :, i]).masked_fill_(mask, 0) - // replaced by precomputed attn_kq - ggml_tensor * attn_chunk = get_slice_2d(ctx0, attn_kq, chunk); - cb(attn_chunk, "attn_chunk", il); - - ggml_tensor * state_t = ggml_cont_4d(ctx0, ggml_permute(ctx0, new_state, 1, 0, 2, 3), S_v, S_v, 1, H_v * n_seqs); - - // v_prime = (k_cumdecay[:, :, i]) @ last_recurrent_state - ggml_tensor * v_prime = ggml_mul_mat(ctx0, state_t, k_cumdecay_chunk); - cb(v_prime, "v_prime_chunk", il); // shape: (S_v, 1, H_v * n_seqs) - - // v_new = v_i - v_prime - ggml_tensor * v_new = ggml_sub(ctx0, ggml_repeat(ctx0, v_chunk, v_prime), v_prime); - ggml_tensor * v_new_t = ggml_cont(ctx0, ggml_transpose(ctx0, v_new)); - cb(v_new, "v_new_chunk", il); - - // attn_inter = (q_i * g[:, :, i, :, None].exp()) @ last_recurrent_state - ggml_tensor * q_g_exp = ggml_mul(ctx0, q_chunk, gexp_chunk); - ggml_tensor * attn_inter = ggml_mul_mat(ctx0, state_t, q_g_exp); - cb(attn_inter, "attn_inter_chunk", il); - - // core_attn_out[:, :, i] = attn_inter + attn @ v_new - ggml_tensor * v_attn = ggml_mul_mat(ctx0, v_new_t, attn_chunk); - cb(v_attn, "v_attn_chunk", il); - - ggml_tensor * core_attn_out_chunk = ggml_add(ctx0, attn_inter, v_attn); - cb(core_attn_out_chunk, "core_attn_out_chunk", il); // shape: (S_v, chunk_size, 1, H_v * n_seqs) - - core_attn_out = core_attn_out == nullptr - ? core_attn_out_chunk - : ggml_concat(ctx0, core_attn_out, core_attn_out_chunk, 2); - - // kgdmulvnew = (key_gdiff).transpose(-1, -2) @ v_new - ggml_tensor * k_gdiff_t = get_slice_2d(ctx0, key_gdiff_t, chunk); - //ggml_tensor * kgdmulvnew = ggml_mul_mat(ctx0, k_gdiff, v_new); // this is slower on metal, why? - ggml_tensor * kgdmulvnew = ggml_mul_mat(ctx0, v_new_t, k_gdiff_t); - - // last_recurrent_state = last_recurrent_state * g_last + kgdmulvnew - ggml_tensor * gexp_last_chunk = ggml_cont(ctx0, get_slice_2d(ctx0, g_last_exp, chunk)); - new_state = ggml_add(ctx0, - ggml_mul(ctx0, new_state, ggml_reshape_4d(ctx0, gexp_last_chunk, gexp_last_chunk->ne[0], gexp_last_chunk->ne[1], H_v, n_seqs)), - ggml_reshape_4d(ctx0, kgdmulvnew, kgdmulvnew->ne[0], kgdmulvnew->ne[1], H_v, n_seqs)); - } - - // truncate padded tokens - ggml_tensor * output_tokens = ggml_view_4d(ctx0, core_attn_out, - S_v, n_tokens, H_v, n_seqs, - ggml_row_size(core_attn_out->type, S_v), - ggml_row_size(core_attn_out->type, S_v * chunk_size * n_chunks), - ggml_row_size(core_attn_out->type, S_v * chunk_size * n_chunks * H_v), 0); - output_tokens = ggml_cont(ctx0, output_tokens); - cb(output_tokens, "output_tokens", il); - - // permute back to (S_v, H_v, n_tokens, n_seqs) - output_tokens = ggml_permute(ctx0, output_tokens, 0, 2, 1, 3); - output_tokens = ggml_cont(ctx0, output_tokens); - - return {output_tokens, new_state}; -} - -std::pair llm_build_qwen3next::build_delta_net_autoregressive( - ggml_tensor * q, - ggml_tensor * k, - ggml_tensor * v, - ggml_tensor * g, - ggml_tensor * beta, - ggml_tensor * 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); // This function is optimized for single token processing - GGML_ASSERT(v->ne[2] == n_tokens); - GGML_ASSERT(k->ne[2] == n_tokens); - GGML_ASSERT(g->ne[0] == H_v && g->ne[1] == n_tokens && g->ne[2] == n_seqs); - GGML_ASSERT(beta->ne[0] == H_v && beta->ne[2] == n_tokens && beta->ne[3] == n_seqs); - GGML_ASSERT(state->ne[0] == S_v && state->ne[1] == S_v * H_v && state->ne[2] == 1 && state->ne[3] == n_seqs); - - 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(H_k == H_v); // we did a repeat to make sure this is the case - - const float eps_norm = hparams.f_norm_rms_eps; - - q = ggml_l2_norm(ctx0, q, eps_norm); - k = ggml_l2_norm(ctx0, k, eps_norm); - - const float scale = 1.0f / sqrtf(S_v); - - q = ggml_scale(ctx0, q, scale); - beta = ggml_sigmoid(ctx0, beta); - - cb(q, "q_in", il); - cb(k, "k_in", il); - cb(v, "v_in", il); - cb(beta, "beta_in", il); - cb(g, "g_in", il); - - state = ggml_reshape_4d(ctx0, state, S_v, S_v, H_v, n_seqs); - - ggml_tensor * g_t = ggml_reshape_4d(ctx0, ggml_transpose(ctx0, g), 1, 1, H_k, n_seqs); - ggml_tensor * beta_t = ggml_reshape_4d(ctx0, ggml_transpose(ctx0, beta), 1, 1, H_k, n_seqs); - - // Apply exponential to g_t - g_t = ggml_exp(ctx0, g_t); - - // Apply the gated delta rule for the single timestep - // last_recurrent_state = last_recurrent_state * g_t - state = ggml_mul(ctx0, state, g_t); - - // kv_mem = (last_recurrent_state * k_t.unsqueeze(-1)).sum(dim=-2) - ggml_tensor * k_t_unsqueezed = ggml_reshape_4d(ctx0, k, 1, S_v, H_v, n_seqs); - ggml_tensor * kv_mem = ggml_mul(ctx0, state, k_t_unsqueezed); - // we need to sum over dim=-2, so we transpose, sum, then transpose again - kv_mem = ggml_transpose(ctx0, ggml_sum_rows(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, kv_mem)))); - - // v_t = v.unsqueeze(2) (we insert the singleton dimension after n_seqs and H_v) - ggml_tensor * v_t = ggml_reshape_4d(ctx0, v, S_v, 1, H_v, n_seqs); - // delta = (v_t - kv_mem) * beta_t - ggml_tensor * v_diff = ggml_sub(ctx0, v_t, kv_mem); // both should be [S_v, 1, H_v, n_seqs] - ggml_tensor * delta = ggml_mul(ctx0, v_diff, beta_t); - - // last_recurrent_state = last_recurrent_state + k_t.unsqueeze(-1) * delta - ggml_tensor * k_t_delta = ggml_mul(ctx0, ggml_repeat_4d(ctx0, k_t_unsqueezed, S_v, S_v, H_v, n_seqs), delta); - state = ggml_add(ctx0, state, k_t_delta); - - // Compute the attention output - // core_attn_out = (last_recurrent_state * q_t.unsqueeze(-1)).sum(dim=-2) - ggml_tensor * q_t_unsqueezed = ggml_reshape_4d(ctx0, q, 1, S_v, H_v, n_seqs); // unsqueeze q_t - ggml_tensor * state_q = ggml_mul(ctx0, state, q_t_unsqueezed); - // again, since it's over dim = -2, transpose, sum, transpose back - ggml_tensor * core_attn_out = - ggml_transpose(ctx0, ggml_sum_rows(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, state_q)))); - - // core_attn_out should be [S_v, 1, H_v, n_seqs] after this - cb(core_attn_out, "output_tokens", il); - cb(state, "new_state", il); - - return {core_attn_out, state}; -} - ggml_tensor * llm_build_qwen3next::build_norm_gated( ggml_tensor * input, ggml_tensor * weights, @@ -752,7 +395,7 @@ ggml_tensor * llm_build_qwen3next::build_layer_attn_linear( v_conv = ggml_cont_4d(ctx0, v_conv, head_v_dim, num_v_heads, n_seq_tokens, n_seqs); ggml_tensor * state = build_rs(inp, ssm_states_all, hparams.n_embd_s(), n_seqs); - state = ggml_reshape_4d(ctx0, state, head_v_dim, head_v_dim * num_v_heads, 1, n_seqs); + state = ggml_reshape_4d(ctx0, state, head_v_dim, head_v_dim, num_v_heads, n_seqs); cb(state, "state_predelta", il); // if head keys and value keys are different, repeat to force tensors into matching shapes @@ -781,13 +424,10 @@ ggml_tensor * llm_build_qwen3next::build_layer_attn_linear( cb(k_conv, "k_conv_predelta", il); cb(v_conv, "v_conv_predelta", il); - // Choose between build_delta_net_chunking, build_delta_net_recurrent, and build_delta_net_autoregressive based on n_tokens - std::pair attn_out; // pair of (output, new_state) - if (n_seq_tokens == 1) { - attn_out = build_delta_net_autoregressive(q_conv, k_conv, v_conv, gate, beta, state, il); - } else { - attn_out = build_delta_net_chunking(q_conv, k_conv, v_conv, gate, beta, state, causal_mask, identity, diag_mask, il); - } + std::pair attn_out = build_delta_net_unified(ctx0, q_conv, k_conv, v_conv, + gate, beta, state, causal_mask, identity, diag_mask, + il, CHUNK_SIZE, hparams.f_norm_rms_eps); + ggml_tensor * output = attn_out.first; ggml_tensor * new_state = attn_out.second; cb(output, "attn_output", il);