# Copyright (c) 2022 The BayesFlow Developers
# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:
# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.
# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.
from abc import ABC, abstractmethod
import tensorflow as tf
[docs]
class Setting(ABC):
"""Abstract base class for settings. It's here to potentially extend the setting functionality in future."""
[docs]
@abstractmethod
def __init__(self):
""""""
pass
DEFAULT_SETTING_INVARIANT_NET = MetaDictSetting(
meta_dict={
"num_dense_s1": 2,
"num_dense_s2": 2,
"num_dense_s3": 2,
"num_equiv": 2,
"pooling_fun": "mean",
"dense_s1_args": None,
"dense_s2_args": None,
"dense_s3_args": None,
"summary_dim": 10,
},
mandatory_fields=[],
)
DEFAULT_SETTING_MULTI_CONV = {
"layer_args": {"activation": "relu", "filters": 32, "strides": 1, "padding": "causal"},
"min_kernel_size": 1,
"max_kernel_size": 3,
}
DEFAULT_SETTING_DENSE_DEEP_SET = {"units": 64, "activation": "relu", "kernel_initializer": "glorot_uniform"}
DEFAULT_SETTING_DENSE_RECT = {"units": 256, "activation": "swish", "kernel_initializer": "glorot_uniform"}
DEFAULT_SETTING_DENSE_ATTENTION = {"units": 64, "activation": "relu", "kernel_initializer": "glorot_uniform"}
DEFAULT_SETTING_DENSE_EVIDENTIAL = {
"units": 64,
"kernel_initializer": "glorot_uniform",
"activation": "elu",
}
DEFAULT_SETTING_DENSE_PMP = {
"units": 64,
"kernel_initializer": "glorot_uniform",
"activation": "elu",
}
DEFAULT_SETTING_AFFINE_COUPLING = MetaDictSetting(
meta_dict={
"dense_args": dict(units=128, activation="relu", kernel_regularizer=tf.keras.regularizers.l2(1e-4)),
"num_dense": 2,
"spec_norm": False,
"mc_dropout": False,
"dropout": True,
"residual": False,
"dropout_prob": 0.05,
"soft_clamping": 1.9,
},
mandatory_fields=[],
)
DEFAULT_SETTING_SPLINE_COUPLING = MetaDictSetting(
meta_dict={
"dense_args": dict(units=128, activation="relu", kernel_regularizer=tf.keras.regularizers.l2(1e-4)),
"num_dense": 2,
"spec_norm": False,
"mc_dropout": False,
"dropout": True,
"residual": False,
"dropout_prob": 0.05,
"bins": 16,
"default_domain": (-5.0, 5.0, -5.0, 5.0),
},
mandatory_fields=[],
)
DEFAULT_SETTING_ATTENTION = {"key_dim": 32, "num_heads": 4, "dropout": 0.01}
DEFAULT_SETTING_INVERTIBLE_NET = MetaDictSetting(
meta_dict={
"num_coupling_layers": 5,
"coupling_net_settings": None,
"coupling_design": "affine",
"permutation": "fixed",
"use_act_norm": True,
"act_norm_init": None,
"use_soft_flow": False,
"soft_flow_bounds": (1e-3, 5e-2),
},
mandatory_fields=["num_params"],
)
DEFAULT_SETTING_EVIDENTIAL_NET = MetaDictSetting(
meta_dict={
"dense_args": dict(units=128, activation="relu"),
"num_dense": 3,
"output_activation": "softplus",
},
mandatory_fields=["num_models"],
)
DEFAULT_SETTING_PMP_NET = MetaDictSetting(
meta_dict={
"dense_args": dict(units=64, activation="relu"),
"num_dense": 3,
"output_activation": "softmax",
},
mandatory_fields=["num_models"],
)
OPTIMIZER_DEFAULTS = {"global_clipnorm": 1.0}
DEFAULT_KEYS = {
"prior_draws": "prior_draws",
"obs_data": "obs_data",
"sim_data": "sim_data",
"batchable_context": "batchable_context",
"non_batchable_context": "non_batchable_context",
"prior_batchable_context": "prior_batchable_context",
"prior_non_batchable_context": "prior_non_batchable_context",
"prior_context": "prior_context",
"hyper_prior_draws": "hyper_prior_draws",
"shared_prior_draws": "shared_prior_draws",
"local_prior_draws": "local_prior_draws",
"sim_batchable_context": "sim_batchable_context",
"sim_non_batchable_context": "sim_non_batchable_context",
"summary_conditions": "summary_conditions",
"direct_conditions": "direct_conditions",
"parameters": "parameters",
"hyper_parameters": "hyper_parameters",
"shared_parameters": "shared_parameters",
"local_parameters": "local_parameters",
"observables": "observables",
"targets": "targets",
"conditions": "conditions",
"posterior_inputs": "posterior_inputs",
"likelihood_inputs": "likelihood_inputs",
"model_outputs": "model_outputs",
"model_indices": "model_indices",
}
MMD_BANDWIDTH_LIST = [1e-6, 1e-5, 1e-4, 1e-3, 1e-2, 1e-1, 1, 5, 10, 15, 20, 25, 30, 35, 100, 1e3, 1e4, 1e5, 1e6]
# Minimum time interval between tqdm status updates to reduce
# load. Only respected when refresh=False in set_postfix
# and set_postfix_str
TQDM_MININTERVAL = 0.1 # in seconds
