提交 0691fb01 authored 作者: Frederic Bastien's avatar Frederic Bastien 提交者: sentient07

Don't register ShapeOptimizer in fast_compile. local_shape_to_shape_i isn't…

Don't register ShapeOptimizer in fast_compile. local_shape_to_shape_i isn't registered, so it won't be used.
上级 3ff56c1d
...@@ -1477,8 +1477,10 @@ class UnShapeOptimizer(Optimizer): ...@@ -1477,8 +1477,10 @@ class UnShapeOptimizer(Optimizer):
# Register it after merge1 optimization at 0. We don't want to track # Register it after merge1 optimization at 0. We don't want to track
# the shape of merged node. # the shape of merged node.
# As local_shape_to_shape_i is not in fast_compile,
# dont register this opt in fast_compile.
theano.compile.mode.optdb.register('ShapeOpt', ShapeOptimizer(), theano.compile.mode.optdb.register('ShapeOpt', ShapeOptimizer(),
0.1, 'fast_run', 'fast_compile') 0.1, 'fast_run')
# Not enabled by default for now. Some crossentropy opt use the # Not enabled by default for now. Some crossentropy opt use the
# shape_feature. They are at step 2.01. uncanonicalize is at step # shape_feature. They are at step 2.01. uncanonicalize is at step
# 3. After it goes to 48.5 that move to the gpu. So 10 seem resonable. # 3. After it goes to 48.5 that move to the gpu. So 10 seem resonable.
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论