aiy.vision.inference¶

An inference engine that communicates with the Vision Bonnet from the Raspberry Pi side.

It can be used to load a model, analyze local image or image from camera shot. It automatically unload the model once the associated object is deleted. See image_classification.py and object_recognition.py as examples on how to use this API.

class aiy.vision.inference.CameraInference(descriptor, params=None, sparse_configs=None)¶

Bases: object

Helper class to run camera inference.

close()¶

count¶

engine¶

rate¶

run(count=None)¶

class aiy.vision.inference.FirmwareVersion(major, minor)¶

Bases: tuple

major¶: Alias for field number 0

minor¶: Alias for field number 1

exception aiy.vision.inference.FirmwareVersionException(*args, **kwargs)¶: Bases: Exception

class aiy.vision.inference.FromSparseTensorConfig(logical_shape, tensor_name, squeeze_dims)¶

Bases: tuple

logical_shape¶: Alias for field number 0

squeeze_dims¶: Alias for field number 2

tensor_name¶: Alias for field number 1

class aiy.vision.inference.ImageInference(descriptor)¶

Bases: object

Helper class to run image inference.

close()¶

engine¶

run(image, params=None, sparse_configs=None)¶

class aiy.vision.inference.InferenceEngine¶

Bases: object

Class to access InferenceEngine on VisionBonnet board.

Inference result has the following format:

message InferenceResult {
  string model_name;  // Name of the model to run inference on.
  int32 width;        // Input image/frame width.
  int32 height;       // Input image/frame height.
  Rectangle window;   // Window inside width x height image/frame.
  int32 duration_ms;  // Inference duration.
  map<string, FloatTensor> tensors;  // Output tensors.

  message Frame {
    int32 index;        // Frame number.
    int64 timestamp_us; // Frame timestamp.
  }

  Frame frame;          // Frame-specific inference data.
}

camera_inference()¶: Returns the latest inference result from VisionBonnet.

close()¶

get_camera_state()¶: Returns current camera state.

get_firmware_info()¶: Returns firmware version as (major, minor) tuple.

get_inference_state()¶: Returns inference state.

get_system_info()¶: Returns system information: uptime, memory usage, temperature.

image_inference(model_name, image, params=None, sparse_configs=None)¶

Runs inference on image using model identified by model_name.

Parameters:	model_name – string, unique identifier used to refer a model. image – PIL.Image, params – dict, additional parameters to run inference
Returns:	pb2.Response.InferenceResult

load_model(descriptor)¶

Loads model on VisionBonnet.

Parameters:	descriptor – ModelDescriptor, meta info that defines model name, where to get the model and etc.
Returns:	Model identifier.

reset()¶

start_camera_inference(model_name, params=None, sparse_configs=None)¶: Starts inference running on VisionBonnet.

stop_camera_inference()¶: Stops inference running on VisionBonnet.

unload_model(model_name)¶

Deletes model on VisionBonnet.

Parameters:	model_name – string, unique identifier used to refer a model.

exception aiy.vision.inference.InferenceException(*args, **kwargs)¶: Bases: Exception

class aiy.vision.inference.ModelDescriptor(name, input_shape, input_normalizer, compute_graph)¶

Bases: tuple

compute_graph¶: Alias for field number 3

input_normalizer¶: Alias for field number 2

input_shape¶: Alias for field number 1

name¶: Alias for field number 0

class aiy.vision.inference.ThresholdingConfig(logical_shape, threshold, top_k, to_ignore)¶

Bases: tuple

logical_shape¶: Alias for field number 0

threshold¶: Alias for field number 1

to_ignore¶: Alias for field number 3

top_k¶: Alias for field number 2