convert npu roll op into paddle roll (#3138)

* convert npu roll op into paddle roll

* convert npu roll op into paddle roll

ppcls/arch/backbone/legendary_models/swin_transformer.py

@@ -42,79 +42,11 @@ MODEL_URLS = {
 
 __all__ = list(MODEL_URLS.keys())
 
-# The following re-implementation of roll is inspired by
-# https://gitee.com/ascend/pytorch/blob/master/torch_npu/contrib/function/roll.py
-
-
 def masked_fill(x, mask, value):
     y = paddle.full(x.shape, value, x.dtype)
     return paddle.where(mask, y, x)
 
 
-class RollWithIndexSelect(paddle.autograd.PyLayer):
-    @staticmethod
-    def forward(ctx, input1, index_fp, index_bp):
-        N, H, W, C = input1.shape
-        ctx.input1 = input1
-        ctx.index_bp = index_bp
-        result = input1.reshape([N, H * W, C]).index_select(
-            index_fp, 1).reshape([N, H, W, C])
-        return result
-
-    @staticmethod
-    def backward(ctx, grad):
-        input1 = ctx.input1
-        N, H, W, C = input1.shape
-        index_bp = ctx.index_bp
-        grad_input = grad.reshape([N, H * W, C]).index_select(
-            index_bp, 1).reshape([N, H, W, C])
-        return grad_input, None, None
-
-
-def get_roll_index(H, W, shifts, place):
-    index = np.arange(0, H * W, dtype=np.int64).reshape([H, W])
-    index_fp = np.roll(index, shift=shifts, axis=(0, 1)).reshape([-1])
-    index_bp = {i: idx for idx, i in enumerate(index_fp.tolist())}
-    index_bp = [index_bp[i] for i in range(H * W)]
-    index_fp = paddle.to_tensor(index_fp, place=place)
-    index_bp = paddle.to_tensor(index_fp, dtype='int64', place=place)
-    return [index_fp, index_bp]
-
-
-class NpuRollWithIndexSelect():
-    def __init__(self):
-        self.index_dict = {}
-        self.roll_with_index_select = RollWithIndexSelect.apply
-
-    def __call__(self, x, shifts, axis):
-        assert x.dim() == 4
-        assert len(shifts) == 2
-        assert len(axis) == 2
-        N, H, W, C = x.shape
-        key = (H, W, shifts, axis)
-        if key not in self.index_dict:
-            self.index_dict[key] = get_roll_index(H, W, shifts, x.place)
-        index_fp, index_bp = self.index_dict[key]
-        return self.roll_with_index_select(x, index_fp, index_bp)
-
-
-class RollWrapper(object):
-    _roll = None
-
-    @staticmethod
-    def roll(x, shifts, axis):
-        return RollWrapper._roll(x, shifts, axis)
-
-
-# NOTE(xiongkun): paddle.SOT can't analysis this builtin function, which will cause subgraph break in sot.
-# we do this here will not effect sot translate.
-paddle_custom_device_types = paddle.device.get_all_custom_device_type()
-
-if RollWrapper._roll is None:
-    RollWrapper._roll = NpuRollWithIndexSelect(
-    ) if 'npu' in paddle_custom_device_types else paddle.roll
-
-
 class Mlp(nn.Layer):
     def __init__(self,
                  in_features,
@@ -457,7 +389,7 @@ class SwinTransformerBlock(nn.Layer):
             5] > 0  # change variable name
         # cyclic shift
         if self.shift_size > 0:
-            shifted_x = RollWrapper.roll(
+            shifted_x = paddle.roll(
                 x, shifts=(-self.shift_size, -self.shift_size), axis=(1, 2))
         else:
             shifted_x = x
@@ -484,7 +416,7 @@ class SwinTransformerBlock(nn.Layer):
 
         # reverse cyclic shift
         if self.shift_size > 0:
-            x = RollWrapper.roll(
+            x = paddle.roll(
                 shifted_x,
                 shifts=(self.shift_size, self.shift_size),
                 axis=(1, 2))