List of all items
Structs
- capability::ArchProfile
- chat_template::ChatTemplate
- chat_template::RenderOptions
- config::KdaConfig
- config::KimiHybridAttention
- config::MlaConfig
- config::MlaRopeConfig
- config::ModelConfig
- decoder::AttnWeights
- decoder::Decoder
- decoder::GptOssLayer
- decoder::GptOssWeights
- decoder::LayerWeights
- decoder::MoeWeights
- deepseek_v4_decoder::DeepseekV4AttnWeights
- deepseek_v4_decoder::DeepseekV4CsaWeights
- deepseek_v4_decoder::DeepseekV4DecodeState
- deepseek_v4_decoder::DeepseekV4DecoderConfig
- deepseek_v4_decoder::DeepseekV4DecoderLayerWeights
- deepseek_v4_decoder::DeepseekV4DecoderWeights
- deepseek_v4_decoder::DeepseekV4LayerState
- deepseek_v4_decoder::DeepseekV4MoeFfnWeights
- device_budget::DeviceBudget
- engine::DeepseekV4Engine
- engine::Glm52Engine
- engine::KimiEngine
- engine::MlaDecodeState
- engine::MlaDenseFfn
- engine::MlaEngine
- engine::MlaLayerWeights
- engine::MlaMoeFfn
- engine::MlaMoeRuntime
- execution_plan::ExecutionPlan
- execution_plan::FusedOpCaps
- execution_plan::MemoryPlan
- execution_plan::PlanGeometry
- gdn::GdnConfig
- gdn::GdnState
- gdn::GdnWeights
- gemma4_engine::Gemma4AttnWeights
- gemma4_engine::Gemma4DecodeState
- gemma4_engine::Gemma4Engine
- gemma4_engine::Gemma4Hparams
- gemma4_engine::Gemma4LayerWeights
- gemma4_engine::Gemma4Weights
- glm52_decoder::Glm52DecodeState
- glm52_decoder::Glm52DecoderConfig
- glm52_decoder::Glm52DecoderLayerWeights
- glm52_decoder::Glm52DecoderWeights
- glm52_decoder::Glm52DenseFfnWeights
- glm52_decoder::Glm52MoeFfnWeights
- glm52_gguf_loader::Glm52DenseFfnWeights
- glm52_gguf_loader::Glm52FileMeta
- glm52_gguf_loader::Glm52GgufHparams
- glm52_gguf_loader::Glm52MoeFfnWeights
- glm_dsa::Glm52AttnState
- glm_dsa::Glm52AttnWeights
- glm_dsa::Glm52MlaConfig
- glm_dsa::IndexerConfig
- glm_dsa::IndexerWeights
- hybrid_engine::HybridEngine
- hybrid_engine::HybridUnavailable
- hybrid_gguf_loader::HybridHparams
- hyper_connections::HyperConnectionHeadWeights
- hyper_connections::HyperConnectionPreWeights
- kda::KdaAttnWeights
- kda::KdaState
- kimi_decoder::DenseMlpWeights
- kimi_decoder::KimiDecodeState
- kimi_decoder::KimiDecoderConfig
- kimi_decoder::KimiDecoderLayerWeights
- kimi_decoder::KimiDecoderWeights
- kimi_gguf_loader::KimiGgufHparams
- kimi_loader::BlockResidualWeights
- kimi_loader::KimiExpertSource
- kimi_loader::KimiRealHparams
- kimi_loader::KimiStoredExpertLayout
- kimi_tokenizer::KimiTokenizer
- kimi_validate::KimiCheckpointReport
- kv_budget::ContextFit
- kv_budget::KvBudget
- kv_budget::KvBudgetError
- kv_budget::KvShape
- kv_budget::SlidingWindow
- latent_moe::KimiExpertWeights
- latent_moe::KimiLatentMoeWeights
- latent_moe::KimiMoeConfig
- llama4_engine::Llama4Engine
- llama4_engine::Llama4Unavailable
- loader::GgufExpertSource
- loader::StoredExpertLayout
- loader::StoredMatrixSpec
- minimax_engine::MinimaxEngine
- minimax_engine::MinimaxUnavailable
- mla::MlaAttnWeights
- mla_gguf_loader::Deepseek2Hparams
- prefix_cache::PrefixCache
- prefix_cache::PrefixCacheStats
- prefix_cache::PrefixMatch
- recurrent_engine::RecurrentEngine
- recurrent_engine::RecurrentUnavailable
- residency_report::ResidencyAssumptions
- residency_report::ResidencyLine
- residency_report::ResidencyReport
- sampling::RecommendedSampling
- sampling::RequestedSampling
- sampling::Sampler
- sampling::SamplingParams
- speculative::DraftBlock
- speculative::DraftDist
- speculative::PromptLookupSpeculator
- speculative::SpeculativeDecodeResult
- speculative::SpeculativeOptions
- t5_engine::T5Engine
- t5_engine::T5Unavailable
- tokenizer::ByteTokenizer
- tokenizer::GgufBpeTokenizer
- tokenizer::GgufSpmTokenizer
- tokenizer::GgufUnigramTokenizer
- tokenizer::StopTokens
- vision::VisionConfig
- vision::VisionEncoderLayerWeights
- vision::VisionEncoderWeights
- vision::VisionMergerWeights
- vl_engine::VlEngine
- vl_engine::VlProjectorPair
- vl_engine::VlUnavailable
Enums
- capability::ArchPath
- capability::ArchScope
- capability::DecoderFamily
- capability::MemoryKind
- capability::QkNormStyle
- chat_template::BuiltinTemplate
- chat_template::TemplateError
- config::AttentionKind
- config::FfnActivation
- config::LayerAttentionKind
- config::RopeLayout
- decoder::ExpertBacking
- device_budget::BudgetBackend
- engine::MlaLayerFfn
- engine_factory::EngineSelectError
- engine_factory::SelectedEngineKind
- engine_factory::ServedEngine
- glm52_decoder::Glm52LayerFfn
- hybrid_gguf_loader::HybridLayerKind
- kimi_decoder::KimiLayerAttention
- kimi_decoder::KimiLayerFfn
- kimi_decoder::KimiLayerState
- kimi_loader::KimiLoadError
- kimi_tokenizer::KimiTokenizerError
- kimi_validate::KimiValidateError
- kv_budget::Ceiling
- kv_budget::ContextCap
- kv_budget::KvElem
- kv_budget::KvLayout
- latent_moe::KimiExpertBacking
- loader::LoadError
- tensor_role::TensorRole
- tokenizer::TokenizerLoadError
Traits
Functions
- block_residual::apply_attn_res
- block_residual::apply_attn_res_prescored
- capability::architecture_catalog
- capability::coverage_report_markdown
- capability::default_swa_pattern
- capability::resolve_architecture
- capability::resolve_profile
- capability::swa_rope_base_follows_model
- capability::unsupported_feature_keys
- capability::unsupported_scaling_keys
- config::deepseek_v4_pro
- config::glm_5_2
- config::kimi_k3
- config::test_dense_fixture
- config::test_mixed_fixture
- config::test_moe_fixture
- deepseek_v4_decoder::deepseek_v4_forward_token
- engine_factory::ensure_generic_decoder
- engine_factory::load_gemma4_engine_from_path
- engine_factory::load_glm52_engine_from_path
- engine_factory::load_mla_engine_from_path
- engine_factory::select_engine_kind
- gdn::gdn_forward_token
- gemma4_gguf_loader::load_gemma4_engine
- gemma4_gguf_loader::read_gemma4_hparams
- glm52_decoder::glm52_forward_token
- glm52_gguf_loader::glm52_indexer_config
- glm52_gguf_loader::glm52_mla_config
- glm52_gguf_loader::load_glm52_attn
- glm52_gguf_loader::load_glm52_dense_ffn
- glm52_gguf_loader::load_glm52_engine
- glm52_gguf_loader::load_glm52_moe_ffn
- glm52_gguf_loader::read_glm52_hparams
- glm_dsa::glm52_attn_forward_token
- hf_pull::looks_like_hf_repo
- hf_pull::pull_hf_gguf
- hf_pull::resolve_model_path
- hybrid_gguf_loader::classify_layers
- hybrid_gguf_loader::detect_layer_kind
- hybrid_gguf_loader::load_gdn_layer_weights
- hybrid_gguf_loader::read_hybrid_hparams
- hybrid_gguf_loader::try_load
- hyper_connections::head
- hyper_connections::post
- hyper_connections::pre
- hyper_connections::sinkhorn
- hyper_connections::weighted_sum
- kda::kda_forward_token
- kimi_decoder::kimi_forward_token
- kimi_generate::kimi_generate
- kimi_gguf_loader::load_kimi_gguf_checkpoint
- kimi_gguf_loader::load_kimi_gguf_layer
- kimi_loader::load_block_residual
- kimi_loader::load_dense_mlp
- kimi_loader::load_f32_vec
- kimi_loader::load_kda_attn
- kimi_loader::load_kimi_checkpoint
- kimi_loader::load_kimi_checkpoint_with_expert_cache
- kimi_loader::load_kimi_expert
- kimi_loader::load_kimi_layer
- kimi_loader::load_latent_moe
- kimi_loader::load_mla_attn
- kimi_tokenizer::parse_special_tokens
- kimi_tokenizer::parse_tiktoken_vocab
- kimi_validate::validate_kimi_checkpoint_dir
- latent_moe::kimi_latent_moe_forward
- loader::assert_every_tensor_consumed
- mla::mla_forward_token
- mla_gguf_loader::load_mla_engine
- mla_gguf_loader::read_deepseek2_hparams
- mmproj::eprint_mmproj_if_present
- mmproj::find_mmproj_beside
- mmproj::warn_mmproj_if_present
- output_projection::grouped_output_projection
- sampling::sampling_distribution
- speculative::accept_or_resample
- speculative::speculative_decode
- speculative::speculative_decode_with
- tokenizer::eog_token_ids
- tokenizer::prepend_bos
- tokenizer::should_add_bos_token
- vision::embed_patches
- vision::encoder_forward
- vision::patch_merge
- vision::project_merged_patches