aimet_tensorflow.adaround

Top-level API

aimet_tensorflow.keras.adaround_weight.Adaround.apply_adaround(model, params, path, filename_prefix, default_param_bw=4, default_quant_scheme=QuantScheme.post_training_tf_enhanced, config_file=None)

Returns model with optimized weight rounding of every op (Conv and Linear) and also saves the corresponding quantization encodings to a separate JSON-formatted file that can then be imported by QuantSim for inference or QAT

Parameters:

  • model (Model) – Model to adaround

  • params (AdaroundParameters) – Parameters for adaround

  • path (str) – path where to store parameter encodings

  • filename_prefix (str) – Prefix to use for filename of the encodings file

  • default_param_bw (int) – Default bitwidth (4-31) to use for quantizing layer parameters. Default 4

  • default_quant_scheme (QuantScheme) – Quantization scheme. Supported options are QuantScheme.post_training_tf or QuantScheme.post_training_tf_enhanced. Default QuantScheme.post_training_tf_enhanced

  • config_file (Optional[str]) – Configuration file for model quantizers

Return type:

Model

Returns:

Model with Adarounded weights

Adaround Parameters

class aimet_tensorflow.keras.adaround_weight.AdaroundParameters(data_set, num_batches, default_num_iterations=10000, default_reg_param=0.01, default_beta_range=(20, 2), default_warm_start=0.2)[source]

Configuration parameters for Adaround

Parameters:

  • data_set (DatasetV2) – TF Data set

  • num_batches (int) – Number of batches

  • default_num_iterations (int) – Number of iterations to adaround each layer. Default 10000

  • default_reg_param (float) – Regularization parameter, trading off between rounding loss vs reconstruction loss. Default 0.01

  • default_beta_range (Tuple) – Start and stop beta parameter for annealing of rounding loss (start_beta, end_beta). Default (20, 2)

  • default_warm_start (float) – warm up period, during which rounding loss has zero effect. Default 20% (0.2)