mlx-bun library API
Documentation
Preparing search index...
@mlx-bun/inference
@mlx-bun/inference/contracts
Module @mlx-bun/inference/contracts
Interfaces
AdaptiveDraftGateSettings
AffineKvLayer
AssistantAttention
AssistantDonors
AssistantDrafterModel
AssistantRowsTarget
AttentionCache
AttentionRead
AudioEncoder
BatchableCache
BidirectionalAttentionCache
BlockAttentionCache
Cache
Cancellation
CheckpointAttachment
CommittedAttentionCache
CompiledDecodeCache
ComposedStorage
Composition
ContextKV
DecodeSlot
DecodeStepInputs
DecodeStepPlan
DecodeTrace
DeepspecContextAttention
DeepspecDrafterModel
DeepspecDraftRows
DenoisingGraph
DenoisingTraining
Dflash2ContextAttention
Dflash2DrafterModel
DisposableResource
DrafterOperationRange
DrafterRowsStep
DraftProjection
DraftProjectionRange
DraftRowCheckpoint
DsparkContextAttention
DsparkDrafterModel
DsparkDraftOptions
EncoderAttentionCache
ExecutionCapabilities
ExecutionContext
ExecutionFeatures
ExecutionGroup
ExecutionPlan
ExecutionPlanner
ExecutionRequirements
FastFilterConfig
GatedDeltaCache
GatedDeltaHeads
GatedDeltaParameters
GenerationOutput
GenerationSession
Glm52MlaWeightSource
Glm52WeightSource
GradCheckpointRun
GradCheckpointTraining
GraphCapabilities
GraphDescriptor
GraphFactory
HeadQuant
HiddenLayerTapsTarget
InferenceEngine
InferenceMethod
KvAttentionView
KvCodec
KvDonorAttention
KvDonorRows
KvMaintenance
LatentAttentionCache
LatentAttentionState
LoraLeaves
Mask
MediaEncoders
MediaSidecarProbes
MemoryPlan
MethodResult
MethodRun
MlxCompiledDecodeStep
MlxDeclaredGraph
MlxDecodeState
MlxDenoisingAppendRange
MlxDenoisingDecodeRange
MlxDenoisingOperations
MlxDenoisingPrefillRange
MlxEmbeddingsRange
MlxForwardStep
MlxHeadRange
MlxKnownRange
MlxKnownSpans
MlxMixedGroupRange
MlxMixedRange
MlxModelMemory
MlxPhaseGroup
MlxPositionedEmbeddingsRange
MlxPositionedRange
MlxPreparationWork
MlxPromptInput
MlxPromptRange
MlxSeparateMixed
MlxSequenceRange
MlxTokenPhases
MropeForwardState
MropeRequestState
NativeMtpHead
ObjectCache
PaddedPrefillCache
PagedKvComposition
PixelInput
PrefillPadding
PrefillPolicy
PrefillPosition
PrefillStep
PrefixCache
PrefixCacheHit
PrefixLayout
PrefixSharedTraining
PreparedStateChange
RecurrentMtpModule
RecurrentMtpTarget
ResidualBasis
ResolvedExecution
ResourceOwner
RowBatchCache
SampledToken
SamplingSession
SchedulingClock
SegmentedPass
SegmentedTraining
SpeculativeRowLayout
SpeculativeTransaction
StateView
TargetView
TextEmbeddingModel
Timer
TokenLogprobs
TrainableGraph
TrainingAttentionCache
VisionEncoder
WhisperDecoderCache
WhisperDims
WhisperFastCache
WhisperFastPath
WhisperGraph
Type Aliases
AffineKernelChoice
AssistantDraftStep
CancelReason
Dflash2DraftBlock
DraftDepth
DrafterContextKV
DrafterContextProjection
DrafterOperations
DraftHeadOperations
DraftProjectionKernel
DraftProjectionKernels
ExecutionReason
FlashTrainingAttention
GenerationEvent
GenerationOutcome
GraphDeclaration
GraphMediaInput
KvStorageScheme
MaskMode
MlxCacheLayer
MlxCapturingForward
MlxEmbeddingsForward
MlxForwardWork
MlxHeadForward
MlxKnownOperations
MlxLayerCapture
MlxMixedForward
MlxPhaseForward
MlxPositionedEmbeddingsForward
MlxPositionedForward
MlxRowTable
MlxSequenceForward
MlxVerifyForward
OnToken
PrefillPhase
PromptNativeWork
RecurrentMtpForward
RunControl
SessionState
SharedKv
TargetPort
Vision
Variables
BF16_KV
COMPOSITION_DEFAULTS
DRAFT_CHECKPOINT_SCHEMA
Functions
nextPrefillStep
prefillPhase
targetLacks
Settings
Member Visibility
Protected
Inherited
External
Theme
OS
Light
Dark
On This Page
Interfaces
Adaptive
Draft
Gate
Settings
Affine
Kv
Layer
Assistant
Attention
Assistant
Donors
Assistant
Drafter
Model
Assistant
Rows
Target
Attention
Cache
Attention
Read
Audio
Encoder
Batchable
Cache
Bidirectional
Attention
Cache
Block
Attention
Cache
Cache
Cancellation
Checkpoint
Attachment
Committed
Attention
Cache
Compiled
Decode
Cache
Composed
Storage
Composition
Context
KV
Decode
Slot
Decode
Step
Inputs
Decode
Step
Plan
Decode
Trace
Deepspec
Context
Attention
Deepspec
Drafter
Model
Deepspec
Draft
Rows
Denoising
Graph
Denoising
Training
Dflash2
Context
Attention
Dflash2
Drafter
Model
Disposable
Resource
Drafter
Operation
Range
Drafter
Rows
Step
Draft
Projection
Draft
Projection
Range
Draft
Row
Checkpoint
Dspark
Context
Attention
Dspark
Drafter
Model
Dspark
Draft
Options
Encoder
Attention
Cache
Execution
Capabilities
Execution
Context
Execution
Features
Execution
Group
Execution
Plan
Execution
Planner
Execution
Requirements
Fast
Filter
Config
Gated
Delta
Cache
Gated
Delta
Heads
Gated
Delta
Parameters
Generation
Output
Generation
Session
Glm52
Mla
Weight
Source
Glm52
Weight
Source
Grad
Checkpoint
Run
Grad
Checkpoint
Training
Graph
Capabilities
Graph
Descriptor
Graph
Factory
Head
Quant
Hidden
Layer
Taps
Target
Inference
Engine
Inference
Method
Kv
Attention
View
Kv
Codec
Kv
Donor
Attention
Kv
Donor
Rows
Kv
Maintenance
Latent
Attention
Cache
Latent
Attention
State
Lora
Leaves
Mask
Media
Encoders
Media
Sidecar
Probes
Memory
Plan
Method
Result
Method
Run
Mlx
Compiled
Decode
Step
Mlx
Declared
Graph
Mlx
Decode
State
Mlx
Denoising
Append
Range
Mlx
Denoising
Decode
Range
Mlx
Denoising
Operations
Mlx
Denoising
Prefill
Range
Mlx
Embeddings
Range
Mlx
Forward
Step
Mlx
Head
Range
Mlx
Known
Range
Mlx
Known
Spans
Mlx
Mixed
Group
Range
Mlx
Mixed
Range
Mlx
Model
Memory
Mlx
Phase
Group
Mlx
Positioned
Embeddings
Range
Mlx
Positioned
Range
Mlx
Preparation
Work
Mlx
Prompt
Input
Mlx
Prompt
Range
Mlx
Separate
Mixed
Mlx
Sequence
Range
Mlx
Token
Phases
Mrope
Forward
State
Mrope
Request
State
Native
Mtp
Head
Object
Cache
Padded
Prefill
Cache
Paged
Kv
Composition
Pixel
Input
Prefill
Padding
Prefill
Policy
Prefill
Position
Prefill
Step
Prefix
Cache
Prefix
Cache
Hit
Prefix
Layout
Prefix
Shared
Training
Prepared
State
Change
Recurrent
Mtp
Module
Recurrent
Mtp
Target
Residual
Basis
Resolved
Execution
Resource
Owner
Row
Batch
Cache
Sampled
Token
Sampling
Session
Scheduling
Clock
Segmented
Pass
Segmented
Training
Speculative
Row
Layout
Speculative
Transaction
State
View
Target
View
Text
Embedding
Model
Timer
Token
Logprobs
Trainable
Graph
Training
Attention
Cache
Vision
Encoder
Whisper
Decoder
Cache
Whisper
Dims
Whisper
Fast
Cache
Whisper
Fast
Path
Whisper
Graph
Type Aliases
Affine
Kernel
Choice
Assistant
Draft
Step
Cancel
Reason
Dflash2
Draft
Block
Draft
Depth
Drafter
Context
KV
Drafter
Context
Projection
Drafter
Operations
Draft
Head
Operations
Draft
Projection
Kernel
Draft
Projection
Kernels
Execution
Reason
Flash
Training
Attention
Generation
Event
Generation
Outcome
Graph
Declaration
Graph
Media
Input
Kv
Storage
Scheme
Mask
Mode
Mlx
Cache
Layer
Mlx
Capturing
Forward
Mlx
Embeddings
Forward
Mlx
Forward
Work
Mlx
Head
Forward
Mlx
Known
Operations
Mlx
Layer
Capture
Mlx
Mixed
Forward
Mlx
Phase
Forward
Mlx
Positioned
Embeddings
Forward
Mlx
Positioned
Forward
Mlx
Row
Table
Mlx
Sequence
Forward
Mlx
Verify
Forward
On
Token
Prefill
Phase
Prompt
Native
Work
Recurrent
Mtp
Forward
Run
Control
Session
State
Shared
Kv
Target
Port
Vision
Variables
BF16_
KV
COMPOSITION_
DEFAULTS
DRAFT_
CHECKPOINT_
SCHEMA
Functions
next
Prefill
Step
prefill
Phase
target
Lacks
Documentation
mlx-bun library API
Loading...