From 706a78973c5401645a924b1c435a661da2ad3708 Mon Sep 17 00:00:00 2001
From: zhangbo9674 <82555433+zhangbo9674@users.noreply.github.com>
Date: Thu, 18 Nov 2021 15:28:10 +0800
Subject: [PATCH] Fix Layer.to() of device bug (#37156)

---
 python/paddle/fluid/dygraph/layers.py | 34 ++++++++++-----------------
 1 file changed, 12 insertions(+), 22 deletions(-)

diff --git a/python/paddle/fluid/dygraph/layers.py b/python/paddle/fluid/dygraph/layers.py
index 8bf8300c8a..8ff960a90e 100644
--- a/python/paddle/fluid/dygraph/layers.py
+++ b/python/paddle/fluid/dygraph/layers.py
@@ -1556,19 +1556,18 @@ class Layer(core.Layer):
             if dtype is None:
                 dtype = t.dtype
 
+            if type(dtype) is str:
+                dtype = convert_np_dtype_to_dtype_(dtype)
+
             # 1. gpu place need to determine whether the memory is sufficient for allocation:
             if t.place.is_gpu_place():
-                gpu_memory_available = core.gpu_memory_available()
                 # for gpu, minimum memory allocation unit is 256 bytes.
-                if type(dtype) is str:
-                    size_dtype = core.size_of_dtype(
-                        convert_np_dtype_to_dtype_(dtype))
-                else:
-                    size_dtype = core.size_of_dtype(dtype)
+                size_dtype = core.size_of_dtype(dtype)
                 # Note(zhangbo): Paddle GPU minimum memory allocation unit is 256 bytes, waiting_alloc_memory will comput ‘t’ occupied memory space.
                 # Coefficient 1.2 is used to avoid OOM that may occur in this critical state when the memory is just enough.
                 waiting_alloc_memory = (
-                    (t.numel().numpy()[0] * size_dtype) / 256 + 1) * 256 * 1.2
+                    (np.prod(t.shape) * size_dtype) / 256 + 1) * 256 * 1.2
+                gpu_memory_available = core.gpu_memory_available()
                 if gpu_memory_available < waiting_alloc_memory:
                     # Copy param / Tensor to cpu
                     t_used = t._copy_to(paddle.CPUPlace(),
@@ -1582,26 +1581,17 @@ class Layer(core.Layer):
 
             # 2. cast param / Tensor to dtype
             if dtype is not None and dtype != t_used.dtype:
-                if isinstance(t_used, framework.ParamBase):
-                    from paddle.fluid.layer_helper import LayerHelper
-                    helper = LayerHelper("cast", **locals())
-                    t_casted = helper.create_variable_for_type_inference(
-                        dtype=dtype)
-                    framework._dygraph_tracer().trace_op(
-                        type='cast',
-                        inputs={'X': t_used},
-                        outputs={'Out': t_casted},
-                        attrs={
-                            'in_dtype': t_used.dtype,
-                            'out_dtype': convert_np_dtype_to_dtype_(dtype)
-                        })
-                else:
+                with paddle.fluid.framework._dygraph_place_guard(
+                        place=t_used.place):
                     t_casted = t_used.cast(dtype=dtype)
             else:
                 t_casted = t_used
 
             # 3. Copy casted cpu param / Tensor to device
-            new_t = t_casted._copy_to(device, blocking)
+            if device is not None and not t_casted.place._equals(device):
+                new_t = t_casted._copy_to(device, blocking)
+            else:
+                new_t = t_casted
 
             # 4. share Tensor to origin param / Tensor
             dst_tensor = t.value().get_tensor()
-- 
GitLab