def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
self.rnns = model_param['n_hidden']
self.film_dim = model_param['film_dim']
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.n_hs = model_param['n_hidden']
self.static_bias = compute_static_bias_initialization(model_param['activation_ratio'])
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.n_hs = model_param['n_hidden'] + [int(self.orch_dim)]
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.n_hidden = model_param['n_hidden']
self.n_visible = self.orch_dim
self.n_condition = self.orch_dim * (self.temporal_order-1)+ self.piano_dim
self.Gibbs_steps = model_param["Gibbs_steps"]
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.n_hs = model_param['n_hidden']
self.n_hs_piano = model_param['n_hidden_piano']
return
def __init__(self, model_param, dimensions): Model_lop.__init__(self, model_param, dimensions) # Hidden layers architecture self.layers = model_param['n_hidden'] # Number of different ordering when sampling self.num_ordering = model_param['num_ordering'] # Is it a keras model ? self.keras = True return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Stack conv
self.filters = model_param["num_filter_piano"]
self.kernels = model_param["kernel_size_piano"]
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.n_hs = model_param['n_hidden']
self.num_filter = model_param['num_filter']
self.filter_size = model_param['filter_size']
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.layers = [self.piano_dim] + list(model_param['n_hidden'])
# Is it a keras model ?
self.keras = False
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.MLP_piano_emb = model_param['MLP_piano_emb']
self.GRU_orch_emb = model_param['GRU_orch_emb']
self.last_MLP = model_param['last_MLP']
return
def __init__(self, model_param, dimensions): Model_lop.__init__(self, model_param, dimensions) # Hidden layers architecture self.layers = model_param['n_hidden'] # Is it a keras model ? self.keras = True return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Architecture
self.layers = model_param['n_hidden']
self.recurrent_layers = model_param['n_hidden']
# Is it a keras model ?
self.keras = True
# Will be computed later
self.context_embedding_size = None
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.layers = model_param['n_hidden']
# Number of different ordering when sampling
self.num_ordering = model_param['num_ordering']
# Is it a keras model ?
self.keras = True
# Will be computed later
self.context_embedding_size = None
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Architecture
self.mlp_piano_present = model_param['mlp_piano_present']
self.recurrent_layers = model_param['recurrent_layers']
self.mlp_orch_present = model_param['mlp_orch_present']
self.mlp_last_pred = model_param['mlp_last_pred']
# Is it a keras model ?
self.keras = True
# Will be computed later
self.context_embedding_size = None
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
self.num_filter_piano = model_param["num_filter_piano"]
self.kernel_size_piano = model_param["kernel_size_piano"]
self.num_filter_orch = model_param["num_filter_orch"]
self.kernel_size_orch = model_param["kernel_size_orch"]
self.embeddings_size = model_param["embeddings_size"]
# The last recurrent layer output a vector of dimension embedding size
self.gru_orch = list(model_param["gru_orch"])
self.gru_orch.append(self.embeddings_size)
self.mlp_pred = model_param["mlp_pred"]
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
self.num_filter_piano = model_param["num_filter_piano"]
self.kernel_size_piano = model_param["kernel_size_piano"]
self.mlp_piano = model_param["mlp_piano"]
self.mlp_pred = model_param["mlp_pred"]
self.gru_orch = model_param["gru_orch"]
# Is it a keras model ?
self.keras = True
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.n_h = model_param['n_hidden']
self.n_v = self.orch_dim
self.n_c = self.orch_dim * (self.temporal_order-1)
self.n_l = self.piano_dim
self.n_f = model_param["n_factor"]
self.n_fv = model_param["n_factor"]
self.n_fh = model_param["n_factor"]
self.Gibbs_steps = model_param["Gibbs_steps"]
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Architecture
self.mlp_piano_present = model_param['mlp_piano_present']
self.recurrent_layers = model_param['recurrent_layers']
self.mlp_orch_present = model_param['mlp_orch_present']
self.mlp_last_pred = model_param['mlp_last_pred']
# Is it a keras model ?
self.keras = True
# Will be computed later
self.context_embedding_size = None
# Static bias
self.static_bias = compute_static_bias_initialization(
model_param['activation_ratio'])
return
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {'n_hidden': list_log_hopt(500, 2000, 10, 1, 2, "n_hidden")}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {
'num_filter_piano':
list_hopt_fixedSized([(20, 30, 1), (10, 20, 1)],
'num_filter_piano'),
'kernel_size_piano':
list_hopt_fixedSized([(12, 24, 1), (12, 24, 1)],
"kernel_size_piano"),
'num_filter_orch':
list_hopt_fixedSized([(30, 50, 1), (10, 20, 1)],
'num_filter_orch'),
'kernel_size_orch':
list_hopt_fixedSized([(12, 24, 1), (12, 24, 1)],
"kernel_size_orch"),
'embeddings_size':
qloguniform_int("embeddings_size", log(500), log(2000), 10),
'mlp_pred':
list_log_hopt(500, 2000, 10, 1, 3, "mlp_pred"),
'gru_orch':
list_log_hopt(500, 2000, 10, 0, 2, "gru_orch"),
}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {
'n_hidden':
hp.choice('n_hidden', [
[
hopt_wrapper.qloguniform_int('n_hidden_1_' + str(i),
log(100), log(5000), 10)
for i in range(1)
],
[
hopt_wrapper.qloguniform_int('n_hidden_2_' + str(i),
log(100), log(5000), 10)
for i in range(2)
],
[
hopt_wrapper.qloguniform_int('n_hidden_3_' + str(i),
log(100), log(5000), 10)
for i in range(3)
],
]),
}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {
'n_hidden': list_log_hopt(500, 2000, 10, 1, 2, "n_hidden"),
'num_ordering': quniform_int('num_ordering', 5, 5, 1)
}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {
'n_hidden': hopt_wrapper.qloguniform_int('n_hidden', log(100), log(5000), 10),
}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {
'filter_0': quniform_int('filter_0', 20, 50, 1),
'kernel_0': quniform_int('kernel_0', 8, 16, 1),
'filter_1': quniform_int('filter_1', 20, 50, 1),
'kernel_1': quniform_int('kernel_1', 8, 16, 1),
}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {
'num_filter_piano': quniform_int('num_filter_piano', 20, 50, 1),
'kernel_size_piano': quniform_int('kernel_size_piano', 8, 16, 1),
'mlp_piano': list_log_hopt(500, 2000, 10, 1, 3, "mlp_piano"),
'mlp_pred': list_log_hopt(500, 2000, 10, 1, 3, "mlp_pred"),
'gru_orch': list_log_hopt(500, 2000, 10, 1, 3, "gru_orch"),
}
space.update(super_space)
return space
def get_hp_space():
super_space = Model_lop.get_hp_space()
space = {
'n_hidden_embedding': hp.choice('n_hidden_embedding', [
[hopt_wrapper.qloguniform_int('n_hidden_embedding_'+str(i), log(1500), log(3000), 10) for i in range(1)],
[hopt_wrapper.qloguniform_int('n_hidden_embedding_'+str(i), log(1500), log(3000), 10) for i in range(2)],
[hopt_wrapper.qloguniform_int('n_hidden_embedding_'+str(i), log(1500), log(3000), 10) for i in range(3)],
]),
'n_hidden_NADE': hp.choice('n_hidden_NADE', [
[hopt_wrapper.qloguniform_int('n_hidden_NADE_'+str(i), log(1500), log(3000), 10) for i in range(1)],
[hopt_wrapper.qloguniform_int('n_hidden_NADE_'+str(i), log(1500), log(3000), 10) for i in range(2)],
[hopt_wrapper.qloguniform_int('n_hidden_NADE_'+str(i), log(1500), log(3000), 10) for i in range(3)],
]),
'num_ordering': quniform_int('num_ordering', 5, 10, 1)
}
space.update(super_space)
return space
def get_hp_space():
space = Model_lop.get_hp_space()
return space
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
return
def __init__(self, model_param, dimensions):
Model_lop.__init__(self, model_param, dimensions)
# Hidden layers architecture
self.n_hidden = model_param['n_hidden']
self.Gibbs_steps = model_param["Gibbs_steps"]
return
