Skip to content

VariableLayer, change default add_batch_axis to False #764

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Nov 24, 2021
Merged

VariableLayer, change default add_batch_axis to False #764

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
618b708
VariableLayer, change default add_batch_axis to False
albertz Nov 24, 2021
6be2a5c
test_ScatterNdLayer_RangeLayer_RangeInAxisLayer fix
albertz Nov 24, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions returnn/tf/layers/basic.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7246,7 +7246,7 @@ class VariableLayer(LayerBase):
"""
layer_class = "variable"

def __init__(self, shape, dtype="float32", add_batch_axis=True, add_time_axis=False, trainable=True,
def __init__(self, shape, dtype="float32", add_batch_axis=False, add_time_axis=False, trainable=True,
init=0,
**kwargs):
"""
Expand Down Expand Up @@ -7298,7 +7298,7 @@ def transform_config_dict(cls, d, network, get_layer):

@classmethod
def get_out_data_from_opts(cls, name, network,
shape, dtype="float32", add_batch_axis=True, add_time_axis=False, **kwargs):
shape, dtype="float32", add_batch_axis=False, add_time_axis=False, **kwargs):
"""
:param str name:
:param returnn.tf.network.TFNetwork network:
Expand Down
12 changes: 6 additions & 6 deletions tests/test_TFNetworkLayer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2473,8 +2473,8 @@ def test_TileLayer():
def test_ScatterNdLayer_RangeLayer_RangeInAxisLayer():
from returnn.tf.util.data import BatchDim, DimensionTag, ImplicitDynSizeDim
n_batch, n_time, n_ts, n_in, n_out = 2, 3, 6, 7, 11
time_dim = DimensionTag(kind=DimensionTag.Types.Spatial, description="T")
feat_dim = DimensionTag(kind=DimensionTag.Types.Feature, description="F", dimension=n_in)
time_dim = DimensionTag(kind=DimensionTag.Types.Spatial, description="time")
feat_dim = DimensionTag(kind=DimensionTag.Types.Feature, description="in-feature", dimension=n_in)
ts_dim = DimensionTag(kind=DimensionTag.Types.Spatial, description="ts", dimension=n_ts)
rnd = numpy.random.RandomState(42)
config = Config({
Expand All @@ -2489,9 +2489,9 @@ def test_ScatterNdLayer_RangeLayer_RangeInAxisLayer():
"add_t": {
"class": "combine", "kind": "add", "from": ["t", "range"],
"out_shape": {time_dim, ts_dim, ImplicitDynSizeDim(BatchDim)}}, # (T,Ts)
"t_rel_var": {"class": "variable", "shape": (n_ts, n_out), "init": "glorot_uniform"}, # (B,Ts,D)
"output": {"class": "scatter_nd", "from": "t_rel_var", "position": "add_t", "position_axis": -1,
"output_dim_via_time_from": "data", "filter_invalid_indices": True}
"t_rel_var": {"class": "variable", "shape": (ts_dim, n_out), "init": "glorot_uniform"}, # (Ts,D)
"output": {"class": "scatter_nd", "from": "t_rel_var", "position": "add_t", "position_axis": ts_dim,
"output_dim_via_time_from": "data", "filter_invalid_indices": True} # (T,T,D)
}
with make_scope() as session:
network = TFNetwork(config=config, train_flag=True)
Expand All @@ -2506,7 +2506,7 @@ def test_ScatterNdLayer_RangeLayer_RangeInAxisLayer():
out_layer = network.get_default_output_layer()
assert isinstance(out_layer, ScatterNdLayer)
assert out_layer.output.shape == (None, None, 11)
assert out_layer.output.feature_dim_axis_or_unspecified is NotSpecified and out_layer.output.feature_dim_axis == 3
assert out_layer.output.feature_dim_axis_or_unspecified is NotSpecified and out_layer.output.feature_dim_axis == 2
assert out_layer.output.time_dim_axis == 0

session.run(tf_compat.v1.variables_initializer(tf_compat.v1.global_variables() + [network.global_train_step]))
Expand Down