Skip to main content

Module model

Module model 

Source
Expand description

Hardware-independent checkpoint inspection and preparation requirements.

Adapters assign meaning to checkpoint names. Consumers use typed roles, encodings, and data references; display names never select execution behavior.

Modules§

format 🔒
index
Persistent model references and explicit ownership of prepared/source stores.
inspect 🔒
preparation 🔒
qwen 🔒
Checkpoint names are interpreted only at the architecture boundary.

Structs§

Checkpoint
A checkpoint description and the immutable objects backing its tensor views. Pass the packed directory explicitly to inspect prepared data.
CheckpointFormat
Container layout and the conventions used to interpret tensor names and encodings.
DataSpan
A byte range within one object; offsets are independent of tensor encoding.
MappedBytes
A byte range that keeps its backing mapping alive after the source is dropped. The backing file must remain unchanged for the lifetime of every clone.
MappedObjects
Read-only file mappings, numbered in the order passed to Self::open.
ModelDescription
Architecture-adapted metadata and storage views, independent of execution hardware.
NgramTable
A table’s hashing semantics are separate from its row encoding and placement.
ObjectId
Identifies an object within an opened checkpoint. It is not a filename or a content digest; a store can assign content identities while copying it.
ObjectInfo
An object’s identity and total size within a byte source.
StoredTensor
Byte strides also describe interleaved records without introducing a table-specific layout or treating packed 4-bit values as addressable bytes.
TableShard
A contiguous logical row range backed by a tensor.
Tensor
A logical tensor and the physical storage used to represent its values.
TensorId
Index of a tensor in its containing inventory or model description.
TensorType
Logical values are distinct from the physical types used to encode them.

Enums§

AffineOffset
Per-group offset used to reconstruct affine-quantized values.
Architecture
Model architecture recognized by the loader.
BitPacking
Placement of quantized codes within addressable storage words.
CheckpointConventions
Recognized naming and encoding conventions within a container.
CompatibilityIssue
A representation constraint that prevents preparation by the current engine.
ContainerFormat
On-disk container family, including Cherenkov’s prepared format.
Dtype
Scalar types supported by physical tensor views.
GgmlEncoding
Recognized GGML element and block encodings; unknown numeric codes are preserved.
NgramHash
Hashing scheme and parameters, separate from the table’s stored representation.
Preparation
Representation requirements for loading with the current Qwen4-exp engine.
PreparationStep
A transformation needed to produce the engine’s prepared representation.
TensorEncoding
Storage layout and quantization metadata, with explicit references to all data.
TensorRole
An architecture-assigned tensor role, independent of its checkpoint name.

Traits§

ByteSource
Object IDs belong to one source; they are neither paths nor content hashes. A copy/import operation must rebind them to its destination’s identities.