returnn 1.20250425.85727__tar.gz → 1.20250430.145858__tar.gz

{returnn-1.20250425.85727/returnn.egg-info → returnn-1.20250430.145858}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: returnn
-Version: 1.20250425.85727
+Version: 1.20250430.145858
 Summary: The RWTH extensible training framework for universal recurrent neural networks
 Home-page: https://github.com/rwth-i6/returnn/
 Author: Albert Zeyer

returnn-1.20250430.145858/_setup_info_generated.py

@@ -0,0 +1,2 @@
+version = '1.20250430.145858'
+long_version = '1.20250430.145858+git.6447e67'

{returnn-1.20250425.85727 → returnn-1.20250430.145858}/returnn/datasets/hdf.py

@@ -1073,6 +1073,8 @@ class SimpleHDFWriter:
     which can be read later by :class:`HDFDataset`.
 
     Note that we dump to a temp file first, and only at :func:`close` we move it over to the real destination.
+
+    Can be used as a context manager, i.e. with the `with` statement.
     """
 
     def __init__(
@@ -1413,6 +1415,12 @@ class SimpleHDFWriter:
             os.remove(self.tmp_filename)
             self.tmp_filename = None
 
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        self.close()
+
 
 class HDFDatasetWriter:
     """

{returnn-1.20250425.85727 → returnn-1.20250430.145858}/returnn/frontend/array_.py

@@ -39,6 +39,7 @@ __all__ = [
     "pad_packed",
     "gather",
     "scatter",
+    "scatter_mean",
     "scatter_argmax",
     "scatter_logsumexp",
     "scatter_logmeanexp",
@@ -807,8 +808,8 @@ def scatter(
     :param source: [batch_dims..., indices_dim(s)..., feature_dims...]
     :param indices: [batch_dims..., indices_dim(s)...] -> out_dim
     :param indices_dim:
-    :param mode: "sum", "max", "min", "logsumexp", "logmeanexp", "argmax".
-        (Note: If you ever need mean, argmin, etc, please open an issue/PR.)
+    :param mode: "sum", "max", "min", "mean", "logsumexp", "logmeanexp", "argmax".
+        (Note: If you ever need another mode, please open an issue/PR.)
     :param fill_value:
     :param out_dim: The indices target dim.
         If not given, will be automatically determined as the sparse_dim from indices.
@@ -817,6 +818,8 @@ def scatter(
     :param use_mask:
     :return: [batch_dims..., out_dim(s)..., feature_dims...]
     """
+    if mode == "mean":
+        return scatter_mean(source, indices=indices, indices_dim=indices_dim, fill_value=fill_value, out_dim=out_dim)
     if mode == "logsumexp":
         return scatter_logsumexp(
             source, indices=indices, indices_dim=indices_dim, fill_value=fill_value, out_dim=out_dim
@@ -863,6 +866,35 @@ def scatter(
     return out
 
 
+def scatter_mean(
+    source: Tensor,
+    *,
+    indices: Tensor,
+    indices_dim: Union[Dim, Sequence[Dim]],
+    fill_value: Optional[Union[int, float]] = None,
+    out_dim: Optional[Union[Dim, Sequence[Dim]]] = None,
+) -> Tensor:
+    """
+    Scatters into new zero-tensor.
+    If entries in indices are duplicated, the corresponding values in source will be mean'ed together.
+    This is like :func:`scatter` with ``mode="mean"``.
+
+    :param source: [batch_dims..., indices_dim(s)..., feature_dims...]
+    :param indices: [batch_dims..., indices_dim(s)...] -> out_dim
+    :param indices_dim:
+    :param fill_value:
+    :param out_dim: The indices target dim.
+        If not given, will be automatically determined as the sparse_dim from indices.
+        If multiple out dims, use indices into the merged out dims,
+        and then we use :func:`rf.split_dims` afterwards.
+    :return: [batch_dims..., out_dim(s)..., feature_dims...]
+    """
+    ones = rf.ones(dims=indices.dims, dtype=source.dtype, device=source.device)
+    counts = rf.scatter(ones, indices=indices, indices_dim=indices_dim, fill_value=1, out_dim=out_dim)
+    y = scatter(source, indices=indices, indices_dim=indices_dim, fill_value=fill_value, out_dim=out_dim)
+    return y / counts
+
+
 def scatter_argmax(
     source: Tensor,
     *,

{returnn-1.20250425.85727 → returnn-1.20250430.145858/returnn.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: returnn
-Version: 1.20250425.85727
+Version: 1.20250430.145858
 Summary: The RWTH extensible training framework for universal recurrent neural networks
 Home-page: https://github.com/rwth-i6/returnn/
 Author: Albert Zeyer

returnn-1.20250425.85727/_setup_info_generated.py

@@ -1,2 +0,0 @@
-version = '1.20250425.085727'
-long_version = '1.20250425.085727+git.547e726'