nlinear~australian_electricity_demand_dataset~0a6d26a4-e0e0-43e9-8c01-313777d56f0c~2.ini

[general]
run_name = 0a6d26a4-e0e0-43e9-8c01-313777d56f0c
run_folder = runs/nlinear_re_oos
dataset_folder = /fs-home/data/monash_forecasting_repo/
dataset = australian_electricity_demand_dataset.tsf
seed = 102
external_forecast_horizon = 336
validation_strategy = re_oos
value_col_name = series_value
time_col_name = start_timestamp
replace_missing_vals_with = NaN
integer_conversion = False
model = nlinear
ctx = cpu
context_length = 24

[model_trainer]
epochs = 50
learning_rate = 0.01
num_batches_per_epoch = 50
patience = 10
distr_output = StudentTOutput
batch_size = 64
decay = 0

[dtypes]
int_dtypes = prediction_length,epochs,num_batches_per_epoch,external_forecast_horizon,seed,patience,batch_size,distribution_hidden_state,context_length,
date_dtypes = ,
bool_dtypes = integer_conversion,
float_dtypes = learning_rate,decay,learning_rate,learning_rate,
str_dtypes = ,distr_output,activation,
list_int_dtypes = ,nodes,
list_str_dtypes = ,
list_bool_dtypes = ,

[model_nlinear]
distribution_hidden_state = 2
nodes = 0,
activation = elu