Accelerator¶
Represents an accelerator, i.e., an NPU, used for executing models.
Methods¶
| Name | Description |
|---|---|
__init__ |
Creates an Accelerator object for a specific device number. |
allocate_pinned_memory |
Allocates an NPU-accessible pinned memory buffer. |
__init__¶
Creates an Accelerator object for a specific device number.
The dev_no parameter represents the device number. For example, on Linux, if an ARIES NPU is attached as /dev/aries0, the device number is 0.
Parameters
dev_no— The device number to associate with the Accelerator.
allocate_pinned_memory¶
Allocates an NPU-accessible pinned memory buffer.
The returned buffer can be used as input or output for zero-copy inference via Model.infer_pinned_memory.
Note
This is an experimental API and is only supported on Regulus hardware.
Parameters
shape— The shape of the buffer to allocate.dtype— The element type of the buffer. Must benumpy.float32ornumpy.uint8.
Returns
A PinnedMemory wrapping the allocated buffer.