edgetpu.classification.engine
An inference engine that performs image classification.

class
edgetpu.classification.engine.
ClassificationEngine
(model_path, device_path=None)¶ Extends
BasicEngine
to perform image classification with a given model.This API assumes the given model is trained for image classification.
Parameters:  model_path (str) – Path to a TensorFlow Lite (
.tflite
) file. This model must be compiled for the Edge TPU; otherwise, it simply executes on the host CPU.  device_path (str) – The device path for the Edge TPU this engine should use. This argument is needed only when you have multiple Edge TPUs and more inference engines than available Edge TPUs. For details, read how to use multiple Edge TPUs.
Raises: ValueError
– If the model’s output tensor size is not 1.
classify_with_image
(img, threshold=0.1, top_k=3, resample=0)¶ Performs classification with an image.
Parameters:  img (
PIL.Image
) – The image you want to classify.  threshold (float) – Minimum confidence threshold for returned classifications. For example,
use
0.5
to receive only classifications with a confidence equalto or higherthan 0.5.  top_k (int) – The maximum number of classifications to return.
 resample (int) – A resampling filter for image resizing.
This can be one of
PIL.Image.NEAREST
,PIL.Image.BOX
,PIL.Image.BILINEAR
,PIL.Image.HAMMING
,PIL.Image.BICUBIC
, orPIL.Image.LANCZOS
. Default isPIL.Image.NEAREST
. See Pillow filters. (Note: A complex filter such asPIL.Image.BICUBIC
may create slightly better accuracy but it also causes higher latency.)
Returns: A
list
of classifications, each of which is a list [int, float] that represents the label id (int) and the confidence score (float).Raises: RuntimeError
– If the model’s input tensor shape doesn’t match the shape expected for an object detection model, which is [1, height, width, channel].ValueError
– If the input tensor channel is not 1 (grayscale) or 3 (RGB)ValueError
– If argument values are invalid.
 img (

classify_with_input_tensor
(input_tensor, threshold=0.0, top_k=3)¶ Performs classification with a raw input tensor.
This requires you to process the input data (the image) and convert it to the appropriately formatted input tensor for your model.
Parameters:  input_tensor (
numpy.ndarray
) – A 1D array as the input tensor.  threshold (float) – Minimum confidence threshold for returned classifications. For example,
use
0.5
to receive only classifications with a confidence equalto or higherthan 0.5.  top_k (int) – The maximum number of classifications to return.
Returns: A
list
of classifications, each of which is a list [int, float] that represents the label id (int) and the confidence score (float).Raises: ValueError
– If argument values are invalid. input_tensor (

device_path
()¶ Gets the path for the Edge TPU that’s associated with this inference engine.
See how to run multiple models with multiple Edge TPUs.
Returns: A string representing this engine’s Edge TPU device path.

get_all_output_tensors_sizes
()¶ Gets the size of each output tensor.
A model may output several tensors, but the return from
run_inference()
andget_raw_output()
concatenates them together into a 1D array. So this function provides the size for each original output tensor, allowing you to calculate the offset for each tensor within the concatenated array.Returns: An array ( numpy.ndarray
) with the length of each output tensor (this assumes that all output tensors are 1D).

get_inference_time
()¶ Gets the latency of the most recent inference.
This can be used by higher level engines for debugging.
Returns: A float representing the inference latency (in milliseconds).

get_input_tensor_shape
()¶ Gets the shape required for the input tensor.
For models trained for image classification / detection, the shape is always [1, height, width, channels]. To be used as input for
run_inference()
, this tensor shape must be flattened into a 1D array with sizeheight * width * channels
. To instead get that 1D array size, userequired_input_array_size()
.Returns: A 1D array ( numpy.ndarray
) representing the required input tensor shape.

get_num_of_output_tensors
()¶ Gets the number of output tensors.
Returns: An integer representing number of output tensors.

get_output_tensor_size
(index)¶ Gets the size of a specific output tensor.
Parameters: index (int) – The index position of the output tensor. Returns: An integer representing the size of the output tensor.

get_raw_output
()¶ Gets the output of the most recent inference.
This can be used by higher level engines for debugging.
Returns: A 1D array ( numpy.ndarray
) representing the output tensor. If there are multiple output tensors, they are compressed into a single 1D array. (Same as what’s returned byrun_inference()
.)

model_path
()¶ Gets the file path for model loaded by this inference engine.
Returns: A string representing the model file’s path.

required_input_array_size
()¶ Gets the required size for the
input_tensor
given torun_inference()
.This is the total size of the 1D array, once the tensor shape is flattened.
Returns: An integer representing the required input tensor size.

run_inference
(input, size=None)¶ Performs inference with a raw input tensor.
Parameters:  input – (
numpy.ndarray
): A 1D array as the input tensor. You can query the required size for this array withrequired_input_array_size()
.  size (int) – input buffer size. When size is not None, it will throw exception if size does not match the expected input size, denoted by n. When size is None, it will throw exception when total input buffer size is smaller than n, and only use the first n bytes of the input buffer to set the input tensor, ignoring the remaining bytes if any in the buffer. (This behavior allows callers to use input buffers with padding bytes at the end, and have extra sanity check that the input matches the model’s expectation.)
Returns: A 2tuple with the inference latency in milliseconds (float) and a 1D array (
numpy.ndarray
) representing the output tensor. If there are multiple output tensors, they are compressed into a single 1D array. For example, if the model outputs 2 tensors with values [1, 2, 3] and [0.1, 0.4, 0.9], the returned 1D array is [1, 2, 3, 0.1, 0.4, 0.9]. You can calculate the size and offset for each tensor usingget_all_output_tensors_sizes()
,get_num_of_output_tensors()
, andget_output_tensor_size()
. Note that the inference result array is a reference, which needs to be deep copied if it needs to be preserved before next inference call. input – (

total_output_array_size
()¶ Gets the expected size of the 1D output array returned by
run_inference()
andget_raw_output()
.Returns: An integer representing the output tensor size.
 model_path (str) – Path to a TensorFlow Lite (
Is this content helpful?