xmos-ai-tools 0.2.1

XMOS AI Tools

XMOS AI Tools

Usage

xformer

from xmos_ai_tools import xformer as xf

xf.convert("source model path", "converted model path", params=None)

where params is a dictionary of compiler flags and paramters and their values.

For example:

from xmos_ai_tools import xformer as xf

xf.convert("example_int8_model.tflite", "xcore_optimised_int8_model.tflite", {
    "mlir-disable-threading": None,
    "xcore-reduce-memory": None,
})

To see all available parameters, call

from xmos_ai_tools import xformer as xf

xf.print_help()

This will print all options available to pass to xformer. To see hidden options, run print_help(show_hidden=True)

To create a parameters file and a tflite model suitable for loading to flash, use the "xcore-flash-image-file" option.

xf.convert("example_int8_model.tflite", "xcore_optimised_int8_flash_model.tflite", {
    "xcore-flash-image-file ": "./xcore_params.params",
})

To use combine these files created by the code above into a .out file use the generate_flash() function

xf.generate_flash("xcore_optimised_int8_flash_model.tflite",  "xcore_params.params", "xcore_flash_binary.out")

xinterpreters

Host Interpreter

from xmos_ai_tools.xinterpreters import xcore_tflm_host_interpreter

ie = xcore_tflm_host_interpreter()
ie.set_model(model_path=xcore_model)
ie.set_tensor(ie.get_input_details()[0]['index'], value=input)
ie.invoke()

xformer_outputs = []
for i in range(num_of_outputs):
    xformer_outputs.append(ie.get_tensor(ie.get_output_details()[i]['index']))

Device Interpreter (USB)

from xmos_ai_tools.xinterpreters import xcore_tflm_usb_interpreter
from xmos_ai_tools.xinterpreters import xcore_tflm_spi_interpreter

ie = xcore_tflm_usb_interpreter()
ie.set_model(model_path=xcore_model, secondary_memory=False, flash=False)
ie.set_tensor(ie.get_input_details()[0]['index'], value=input)
ie.invoke()

xformer_outputs = []
for i in range(num_of_outputs):
    xformer_outputs.append(ie.get_tensor(ie.get_output_details()[i]['index']))