List of all items

Structs

• AttentionConfig
• ModelConfig
• ModelInfo
• RopeScaling
• common::families::AudioBuffer
• common::families::EncoderState
• common::families::SpeakerRef
• common::families::TranscriptSegment
• common::llm::LlmRuntimeConfig
• common::paged_pool::BlockAllocator
• common::paged_pool::PagedSeqState
• definition::ConfigManager
• definition::ModelDefinition
• deltanet_s1::W3DeltaNetS1Dump
• deltanet_s1::W3DeltaNetS1Shape
• executor::bert_executor::BertModelExecutor
• executor::clip_executor::ClipModelExecutor
• executor::common::GenericKvCacheHandle
• executor::llm_executor::LlmExecutor
• executor::stub_executor::StubModelExecutor
• executor::tts_executor::TtsModelExecutor
• executor::vnext_executor::VNextDeterminismExecutionSpec
• executor::vnext_executor::VNextDeterminismParticipantSpec
• executor::vnext_executor::VNextExecutorConfig
• executor::vnext_executor::VNextModelExecutor
• executor::whisper_executor::WhisperModelExecutor
• hf_download::HfDownloader
• image_processor::ClipImageProcessor
• loader::gptq_loader::GptqLayerWeights
• loader::gptq_loader::QuantizeConfig
• loader::safetensors_loader::SafeTensorsLoader
• lora::ActiveLoraAdapter
• lora::LoraAdapterConfig
• lora::LoraTensorPair
• lora::RuntimeLoraAdapter
• lora::RuntimeLoraLinear
• lora::StartupLoraAdapter
• lora::StartupLoraSpec
• models::llama_family::LlamaFamilyConfig
• models::llama_family::LlamaFamilyLayer
• models::llama_family::LlamaFamilyLayerStageConfig
• models::llama_family::LlamaFamilyModel
• models::llama_family::LlamaFamilyScratch
• models::llama_family::RopeCache
• models::llama_family_pipeline::LlamaFamilyPipelineModel
• models::llama_family_pipeline::LlamaPipelinePlacement
• models::llama_family_pipeline::LlamaPipelineStagePlacement
• models::qwen3_moe::Qwen3MoeLayerState
• models::qwen3_moe::Qwen3MoeModel
• models::qwen3_moe::Qwen3MoeScratch
• moe::dispatch::DeviceRouteScratch
• moe::dispatch::ExpertStack
• moe::dispatch::MoeBucketPlan
• moe::dispatch::MoeBucketProfileSnapshot
• moe::dispatch::MoeForwardBucketedParams
• moe::dispatch::MoeForwardParams
• moe::dispatch::MoeRouteScratch
• moe::layer::Qwen3MoeLayer
• moe::router::RouterOutput
• moe_config::Qwen3MoeConfig
• multimodal::bert::BertModelWrapper
• multimodal::clip::ClipModelWrapper
• multimodal::qwen3_tts::Qwen3TTSTalker
• multimodal::qwen3_tts::SubTalker
• multimodal::qwen3_tts::TalkerConfig
• multimodal::qwen3_tts_backbone::TalkerBackboneBackend
• multimodal::qwen3_tts_backend::Qwen3TtsSubTalker
• multimodal::qwen3_tts_backend::Qwen3TtsTalker
• multimodal::qwen3_tts_vocoder::Qwen3TTSVocoder
• multimodal::qwen3_tts_vocoder::VocoderConfig
• multimodal::speaker_encoder::SpeakerEncoder
• multimodal::speech_tokenizer_encoder::SpeechTokenizerEncoder
• multimodal::whisper::WhisperModelWrapper
• qwen35_config::Qwen35LayerPlan
• qwen35_config::Qwen35LayerWeightManifest
• qwen35_config::Qwen35LinearAttentionConfig
• qwen35_config::Qwen35MoeTextConfig
• qwen35_config::Qwen35QuantizationConfig
• qwen35_config::Qwen35RecurrentStateDtypes
• qwen35_config::Qwen35RopeParameters
• qwen35_config::Qwen35TextConfig
• qwen35_config::Qwen35WeightManifest
• qwen35_config::Qwen35WeightSpec
• qwen35_weights::Qwen35ResolvedLayerWeights
• qwen35_weights::Qwen35ResolvedWeightPlan
• qwen35_weights::Qwen35ResolvedWeightSpec
• qwen35_weights::Qwen35WeightInventory
• qwen35_weights::Qwen35WeightPlanLoader
• qwen35_weights::Qwen35WeightValidation
• registry::DefaultModelRegistry
• registry::ModelAlias
• registry::ModelDiscoveryEntry
• source::DefaultModelSourceResolver
• source::ModelSourceConfig
• source::ResolvedModelSource
• tensor_wrapper::CandleTensorWrapper
• tokenizer::PlaceholderTokenizer
• tokenizer::TokenizerFactory
• tokenizer::TokenizerHandle
• vnext::CausalLanguageModelDescriptor
• vnext::PreparedProductionModel
• vnext::ProductionModelFamilyRegistry
• vnext::RegisteredProductionModel
• vnext::qwen35::Qwen35FamilyConfig
• vnext::qwen35::Qwen35FamilyProvider
• vnext::qwen3_moe::Qwen3MoeFamilyConfig
• vnext::qwen3_moe::Qwen3MoeFamilyProvider
• vnext::source::HuggingFaceSnapshotIdentity
• vnext::source::ProductionModelSourceBundle

Enums

• Activation
• ModelType
• NormType
• executor::vnext_executor::VNextDeterminismExecutionMode
• executor::vnext_executor::VNextDeterminismInitialState
• executor::vnext_executor::VNextDeterminismPhase
• executor::vnext_executor::VNextDeterminismWorkspacePoison
• gguf_runtime::GgufArch
• gguf_runtime::GgufRuntime
• models::llama_family::RopeScalingConfig
• models::llama_family_pipeline::LlamaPipelineMode
• models::llama_family_pipeline::LlamaPipelineStageBridge
• models::llama_family_pipeline::LlamaPipelineTransport
• qwen35_config::Qwen35LayerType
• qwen35_config::Qwen35MlpKind
• qwen35_weights::Qwen35ResolvedWeightSource
• registry::Architecture
• registry::ModelFormatType
• source::ModelFormat
• vnext::ProductionExecutionKind
• vnext::ProductionModelRegistration
• vnext::source::ProductionWeightArtifact
• weight_format::WeightFormat

Traits

• ModelExecutor
• common::families::EmbeddingModel
• common::families::EncoderDecoderLM
• common::families::MultimodalLLM
• common::families::Transcriber
• common::families::TtsModel
• common::llm::DecoderOnlyLLM
• multimodal::qwen3_tts_backbone::TalkerBackboneForward
• source::ModelSourceResolver

Functions

• audio_processor::chunk_pcm
• audio_processor::load_audio
• audio_processor::load_audio_at_rate
• audio_processor::load_audio_bytes
• common::decoder_unified::compute_cu_seqlens_q
• common::decoder_unified::compute_final_indices
• common::decoder_unified::compute_max_kv_len
• common::decoder_unified::compute_pos_offsets
• common::decoder_unified::concat_q_tokens
• common::decoder_unified::stack_block_tables
• common::decoder_unified::unified_attention_launch_key
• common::decoder_unified::unified_graph_key
• common::decoder_unified::unified_layers_only_graph_key
• common::decoder_unified::unified_lm_head_eager_graph_key
• common::paged_pool::block_hash
• common::paged_pool::block_hash_chain
• deltanet_s1::compute_w3_deltanet_s1_dump
• deltanet_s1::write_w3_deltanet_s1_dump
• executor::common::default_executor_status
• executor::common::tensor_to_tokens
• executor::common::tokens_to_tensor
• executor::common::wrap_tensor
• executor::tts_executor::sample_token
• gguf_config::gguf_num_layers
• gguf_engine_loader::auto_discover_tokenizer_path
• gguf_engine_loader::is_gguf_path
• gguf_engine_loader::load_gguf_decoder
• gguf_engine_loader::load_gguf_decoder_with_info
• legacy_capabilities::from_definition_with_weight_bytes
• legacy_capabilities::from_definition_with_weight_bytes_for_hardware
• legacy_capabilities::quantization_from_definition
• loader::gptq_loader::load_gptq_weights
• lora::default_lora_model_id
• lora::load_runtime_lora_adapter
• lora::load_startup_lora_adapter
• lora::load_startup_lora_adapters
• lora::render_lora_model_id
• mel::log_mel_spectrogram
• moe::dispatch::drain_moe_bucket_profile
• moe::dispatch::moe_forward
• moe::dispatch::moe_forward_bucketed
• moe::dispatch::moe_forward_cpu
• moe::router::route
• moe::router::route_into
• multimodal::speaker_encoder::mel_spectrogram_speaker_encoder
• vnext::gguf_architecture_requires_typed_product_sources
• vnext::open_registered_colocated_safetensors
• vnext::open_registered_product_sources
• vnext::qwen35::prepare_from_model_dir
• vnext::qwen3_moe::prepare_from_model_dir
• vnext::resolve_registered_model_from_dir
• vnext::resolve_registered_model_from_sources
• vnext::source::huggingface_snapshot_identity
• vnext::validate_registered_model_semantics

Type Aliases

• Result
• common::families::AudioTokens
• common::families::ImageBuffer
• common::families::PcmSamples
• common::families::VisualTokens
• common::paged_pool::BlockHash

Statics

• moe::dispatch::MOE_BUCKET_COMBINE_US
• moe::dispatch::MOE_BUCKET_D2H_US
• moe::dispatch::MOE_BUCKET_GATHER_US
• moe::dispatch::MOE_BUCKET_GEMM1_US
• moe::dispatch::MOE_BUCKET_GEMM3_US
• moe::dispatch::MOE_BUCKET_LAYER_CALLS
• moe::dispatch::MOE_BUCKET_PLAN_US
• moe::dispatch::MOE_BUCKET_ROUTE_US
• moe::dispatch::MOE_BUCKET_SILU_US
• moe::dispatch::MOE_BUCKET_SYNC_US
• moe::dispatch::MOE_COPY_CALLS
• moe::dispatch::MOE_COPY_US
• moe::dispatch::MOE_GEMV_DOWN_CALLS
• moe::dispatch::MOE_GEMV_DOWN_US
• moe::dispatch::MOE_GEMV_GATE_UP_CALLS
• moe::dispatch::MOE_GEMV_GATE_UP_US
• moe::dispatch::MOE_HOST_TOPK_CALLS
• moe::dispatch::MOE_HOST_TOPK_US
• moe::dispatch::MOE_SCALED_ADD_CALLS
• moe::dispatch::MOE_SCALED_ADD_US
• moe::dispatch::MOE_SILU_CALLS
• moe::dispatch::MOE_SILU_US
• moe::dispatch::MOE_SYNC_CALLS
• moe::dispatch::MOE_SYNC_US

Constants

• audio_processor::CHUNK_SAMPLES
• deltanet_s1::DUMP_MANIFEST_NAME
• executor::vnext_executor::MAX_VNEXT_DETERMINISM_PARTICIPANTS
• moe::dispatch::MOE_BLOCK_SIZE_MAX
• qwen35_config::QWEN35_CONV_STATE_NAME
• qwen35_config::QWEN35_DELTA_STATE_NAME
• vnext::qwen35::EXTERNAL_METADATA_ID
• vnext::qwen35::FAMILY_ID
• vnext::qwen35::MOE_EXTERNAL_METADATA_ID
• vnext::qwen3_moe::EXTERNAL_METADATA_ID
• vnext::qwen3_moe::FAMILY_ID