Struct Qwen4ExpWeightLoader 

pub struct Qwen4ExpWeightLoader;

Trait Implementations

impl ModelWeightLoader for Qwen4ExpWeightLoader

fn load_final_norm( &self, store: &WeightStore, config: &ModelConfig, gpu: &dyn GpuBackend, ) -> Result<DenseWeight>

This model has no final norm tensor.

There is no model.norm.weight anywhere in the checkpoint. The model-level hyper_connection_mixer — which collapses the hc_mult residual streams back to a single hidden state before lm_head — carries hc_norm [hc_mult*hidden], and that IS the final normalization. It is the wrong width to stand in here (10240 against 2560), and applying it as though it were a plain final norm would be inventing math.

A ones-filled buffer keeps the shape contract so the footprint can be measured at load. It is unreachable at inference because the mHC forward refuses first; if that ever stops being true, this is the first thing to fix.

fn supports_tp(&self) -> bool

Whether this loader’s weight slicing is TP-aware. No default — every loader MUST declare this explicitly so adding a new model architecture cannot accidentally inherit a false and silently regress users who pass --tp-size > 1.

fn load_layers( &self, store: &WeightStore, config: &ModelConfig, gpu: &dyn GpuBackend, layer_kv_dtypes: &[KvCacheDtype], ) -> Result<Vec<Box<dyn TransformerLayer>>>

Load all transformer layers from the weight store.

fn load_embedding( &self, store: &WeightStore, config: &ModelConfig, _gpu: &dyn GpuBackend, ) -> Result<DenseWeight>

fn load_lm_head( &self, store: &WeightStore, config: &ModelConfig, _gpu: &dyn GpuBackend, ) -> Result<DenseWeight>

fn load_vision_encoder( &self, store: &WeightStore, config: &ModelConfig, gpu: &dyn GpuBackend, ) -> Result<Option<VisionEncoder>>

Load vision encoder weights (returns None for text-only models).

fn load_mtp_weights( &self, _store: &WeightStore, _config: &ModelConfig, _gpu: &dyn GpuBackend, ) -> Result<Option<MtpWeights>>

Load MTP head weights (returns None if no MTP weights in store).

fn prune_after_load( &self, _store: &mut WeightStore, _config: &ModelConfig, _gpu: &dyn GpuBackend, ) -> Result<()>

Drop store tensors this loader has finished with, after every load_* reader has run and before the buffer arena / KV cache are sized.

fn precision_schedule(&self, _config: &ModelConfig) -> PrecisionSchedule

Per-(layer, role) weight precision schedule (C.3, 2026-04-25). Default impl returns the empty schedule (every lookup yields Dtype::Inherit), preserving the existing per-checkpoint dtype logic byte-for-byte. Loader-specific implementations can override to honour MODEL.toml’s [precision] block.

fn load_ngram_embedding( &self, _store: &WeightStore, _config: &ModelConfig, _gpu: &dyn GpuBackend, _max_tokens: usize, ) -> Result<Option<NgramEmbedding>>

Build the n-gram embedding, when this architecture fuses hashed n-gram lookups into the input embedding (LongCat / Qwen3.8-Flash-Next).

fn load_mtp_weights_multi( &self, store: &WeightStore, config: &ModelConfig, gpu: &dyn GpuBackend, ) -> Result<Vec<MtpWeights>>

Load MTP weights for multi-module MTP (DeepSeek-V3 / MiniMax-M2 style: N independent transformer modules, each with its own attention + MoE + KV cache). Returns an empty Vec when the checkpoint has no MTP modules, a 1-element Vec for single-module MTP (Qwen3.5 family), or N elements for multi-module.

fn kv_layer_dims(&self, _config: &ModelConfig) -> Vec<(usize, usize)>

Per-layer (num_kv_heads, head_dim) overrides for heterogeneous attention models (e.g. Gemma-4 with sliding 16×256 and full 4×512). Default empty — homogeneous models skip per-layer dims and the KV cache allocator uses the global (num_kv_heads, head_dim). Populated by loaders whose models have different attention geometries per layer. Indexed by attention layer index (same as layer_kv_dtypes).

fn load_dflash_weights( &self, _drafter_store: &WeightStore, _config: &ModelConfig, _gpu: &dyn GpuBackend, _tp_size: usize, ) -> Result<Option<DflashWeights>>

Load DFlash drafter weights from a separate WeightStore pointing at the drafter checkpoint (z-lab/Qwen3.6-{27B,35B-A3B}-DFlash). Default impl returns None so loaders that don’t yet support DFlash silently fall through to the existing MTP path. Override in loaders whose target models pair with a DFlash drafter (Qwen3.5/3.6 family). The same drafter format works across both 27B-dense and 35B-A3B-MoE targets — only the target_hidden_size validated against the drafter’s fc input dimension differs.

fn load_lora_adapters( &self, adapters: &[LoraAdapterInput<'_>], config: &ModelConfig, gpu: &dyn GpuBackend, max_loras: usize, max_lora_rank: usize, ) -> Result<Option<LoraWeights>>

Load one or more startup-static PEFT LoRA adapters from their own WeightStores (the adapter_model.safetensors tensors, already on-device BF16) into the fixed-address rank-padded pool (one slot each).

fn binds_vision_encoder(&self) -> bool

Will this loader ever bind a vision encoder for a multimodal checkpoint?