mirror of https://github.com/Jittor/Jittor
215 lines
9.5 KiB
Python
215 lines
9.5 KiB
Python
# ***************************************************************
|
|
# Copyright (c) 2021 Jittor. All Rights Reserved.
|
|
# Maintainers:
|
|
# Guowei Yang <471184555@qq.com>
|
|
# Wenyang Zhou <576825820@qq.com>
|
|
# Meng-Hao Guo <guomenghao1997@gmail.com>
|
|
# Dun Liang <randonlang@gmail.com>.
|
|
#
|
|
#
|
|
# This file is subject to the terms and conditions defined in
|
|
# file 'LICENSE.txt', which is part of this source code package.
|
|
# ***************************************************************
|
|
import jittor as jt
|
|
from jittor import init, Module
|
|
import numpy as np
|
|
import math
|
|
|
|
class Pool(Module):
|
|
def __init__(self, kernel_size, stride=None, padding=0, dilation=None, return_indices=None, ceil_mode=False, count_include_pad=True, op="maximum"):
|
|
assert dilation == None
|
|
assert return_indices == None
|
|
self.kernel_size = kernel_size
|
|
self.op = op
|
|
self.stride = stride if stride else kernel_size
|
|
self.padding = padding
|
|
self.ceil_mode = ceil_mode
|
|
self.count_include_pad = count_include_pad and padding != 0
|
|
|
|
def execute(self, x):
|
|
N,C,H,W = x.shape
|
|
if self.ceil_mode == False:
|
|
h = (H+self.padding*2-self.kernel_size)//self.stride+1
|
|
w = (W+self.padding*2-self.kernel_size)//self.stride+1
|
|
use_code_op = self.op in ['maximum', 'minimum']
|
|
# some second order avg_pool is require, so we don't use code op here
|
|
else:
|
|
h = (H+self.padding*2-self.kernel_size + self.stride - 1)//self.stride+1
|
|
w = (W+self.padding*2-self.kernel_size + self.stride - 1)//self.stride+1
|
|
use_code_op = self.op in ['maximum', 'minimum', 'mean']
|
|
|
|
if use_code_op:
|
|
if self.op == 'mean':
|
|
if self.count_include_pad:
|
|
count = f"int count = {self.kernel_size*self.kernel_size};"
|
|
else:
|
|
count = "int count = (k2_ - k2) * (k3_ - k3);"
|
|
count += "float32 rcount = 1.0f / count;"
|
|
else:
|
|
count = ""
|
|
forward_body = f'''{{
|
|
int k3 = i3*{self.stride}-{self.padding};
|
|
int k2 = i2*{self.stride}-{self.padding};
|
|
int k3_ = min(k3 + {self.kernel_size}, in0_shape3);
|
|
int k2_ = min(k2 + {self.kernel_size}, in0_shape2);
|
|
k3 = max(0, k3);
|
|
k2 = max(0, k2);
|
|
@out(i0, i1, i2, i3) = init_{self.op}(out_type);
|
|
{count}
|
|
for (int p = k2; p < k2_; ++p)
|
|
for (int q = k3; q < k3_; ++q)
|
|
@out(i0, i1, i2, i3) = {self.op}(out_type, @out(i0, i1, i2, i3), @in0(i0, i1, p, q));
|
|
}}'''
|
|
backward_body = f'''{{
|
|
int k3 = i3*{self.stride}-{self.padding};
|
|
int k2 = i2*{self.stride}-{self.padding};
|
|
int k3_ = min(k3 + {self.kernel_size}, in0_shape3);
|
|
int k2_ = min(k2 + {self.kernel_size}, in0_shape2);
|
|
k3 = max(0, k3);
|
|
k2 = max(0, k2);
|
|
{count}
|
|
int bo=1;
|
|
for (int p = k2; p < k2_ && bo; ++p)
|
|
for (int q = k3; q < k3_ && bo; ++q) {{
|
|
{"atomicAdd(&@out(i0,i1,p,q), @dout(i0,i1,i2,i3)/count);"
|
|
if self.op == "mean" else
|
|
f"""if (@pout(i0,i1,i2,i3) == @in0(i0,i1,p,q)) {{
|
|
atomicAdd(&@out(i0,i1,p,q), @dout(i0,i1,i2,i3)),
|
|
bo=0;
|
|
}}"""}
|
|
}}
|
|
}}'''
|
|
out = jt.code([N,C,h,w], x.dtype, [x],
|
|
cuda_header="""
|
|
#include <ops/binary_op_defs.h>
|
|
#include <misc/cuda_limits.h>
|
|
""",
|
|
cuda_src=f'''
|
|
__global__ static void kernel1(@ARGS_DEF) {{
|
|
@PRECALC
|
|
int p3 = threadIdx.x;
|
|
int s3 = blockDim.x;
|
|
int p2 = threadIdx.y + blockIdx.x * blockDim.y;
|
|
int s2 = blockDim.y * gridDim.x;
|
|
int i1 = blockIdx.y;
|
|
int i0 = blockIdx.z;
|
|
for (int i3 = p3; i3 < out_shape3; i3 += s3)
|
|
for (int i2 = p2; i2 < out_shape2; i2 += s2)
|
|
{forward_body}
|
|
}}
|
|
int tx = min(1024, out_shape3);
|
|
int ty = min(1024 / tx, out_shape2);
|
|
int bx = (out_shape2 - 1) / ty + 1;
|
|
int by = out_shape1;
|
|
int bz = out_shape0;
|
|
dim3 s1(bx, by, bz);
|
|
dim3 s2(tx, ty);
|
|
kernel1<<<s1, s2>>>(@ARGS);
|
|
''',
|
|
cuda_grad_src=[f'''
|
|
__global__ static void kernel3(@ARGS_DEF) {{
|
|
@PRECALC
|
|
int p3 = threadIdx.x;
|
|
int s3 = blockDim.x;
|
|
int p2 = threadIdx.y + blockIdx.x * blockDim.y;
|
|
int s2 = blockDim.y * gridDim.x;
|
|
int i1 = blockIdx.y;
|
|
int i0 = blockIdx.z;
|
|
for (int i3 = p3; i3 < pout_shape3; i3 += s3)
|
|
for (int i2 = p2; i2 < pout_shape2; i2 += s2)
|
|
{backward_body}
|
|
}}
|
|
cudaMemsetAsync(out_p, 0, out->size);
|
|
int tx = min(1024, pout_shape3);
|
|
int ty = min(1024 / tx, pout_shape2);
|
|
int bx = (pout_shape2 - 1) / ty + 1;
|
|
int by = pout_shape1;
|
|
int bz = pout_shape0;
|
|
dim3 s1_(bx, by, bz);
|
|
dim3 s2_(tx, ty);
|
|
kernel3<<<s1_, s2_>>>(@ARGS);
|
|
'''],
|
|
cpu_header='#include <ops/binary_op_defs.h>',
|
|
cpu_src=f'''
|
|
using namespace std;
|
|
for (int i0=0; i0<out_shape0; i0++)
|
|
for (int i1=0; i1<out_shape1; i1++)
|
|
for (int i2=0; i2<out_shape2; i2++)
|
|
for (int i3=0; i3<out_shape3; i3++)
|
|
{forward_body}
|
|
''',
|
|
cpu_grad_src = [f'''
|
|
using namespace std;
|
|
std::memset(out_p, 0, out->size);
|
|
#define atomicAdd(a,b) (*a) += b
|
|
|
|
for (int i0=0; i0<pout_shape0; i0++)
|
|
for (int i1=0; i1<pout_shape1; i1++)
|
|
for (int i2=0; i2<pout_shape2; i2++)
|
|
for (int i3=0; i3<pout_shape3; i3++)
|
|
{backward_body}
|
|
'''])
|
|
return out
|
|
else:
|
|
# TODO: backward
|
|
xx = x.reindex([N,C,h,w,self.kernel_size,self.kernel_size], [
|
|
"i0", # Nid
|
|
"i1", # Cid
|
|
f"i2*{self.stride}-{self.padding}+i4", # Hid
|
|
f"i3*{self.stride}-{self.padding}+i5", # Wid
|
|
])
|
|
return xx.reduce(self.op, [4,5])
|
|
|
|
|
|
class AdaptiveAvgPool2d(Module):
|
|
def __init__(self, output_size):
|
|
self.output_size = output_size
|
|
|
|
def execute(self, x):
|
|
if isinstance(self.output_size, int):
|
|
oh = self.output_size
|
|
ow = self.output_size
|
|
elif isinstance(self.output_size, tuple) or isinstance(self.output_size, list):
|
|
oh = x.shape[2] if self.output_size[0] is None else self.output_size[0]
|
|
ow = x.shape[3] if self.output_size[1] is None else self.output_size[1]
|
|
else:
|
|
raise TypeError(f"AdaptiveAvgPool2d only support int, tuple or list input. Not support {type(self.output_size)} yet.")
|
|
if oh == 1 and ow == 1:
|
|
return x.reduce("mean", [2,3], keepdims=True)
|
|
N,C,H,W = x.shape
|
|
self.sh = math.floor(H / oh)
|
|
self.sw = math.floor(W / ow)
|
|
self.ksh = H - (oh - 1) * self.sh
|
|
self.ksw = W - (ow - 1) * self.sw
|
|
h = (H-self.ksh)//self.sh+1
|
|
w = (W-self.ksw)//self.sw+1
|
|
xx = x.reindex([N,C,h,w,self.ksh,self.ksw], [
|
|
"i0", # Nid
|
|
"i1", # Cid
|
|
f"i2*{self.sh}+i4", # Hid
|
|
f"i3*{self.sw}+i5", # Wid
|
|
])
|
|
return xx.reduce("mean", [4,5])
|
|
|
|
def pool(x, kernel_size, op, padding=0, stride=None):
|
|
return Pool(kernel_size, stride, padding, op=op)(x)
|
|
|
|
class AvgPool2d(Module):
|
|
def __init__(self, kernel_size, stride=None, padding=0, ceil_mode=False, count_include_pad=True):
|
|
self.layer = Pool(kernel_size=kernel_size, stride=stride, padding=padding, ceil_mode=ceil_mode, count_include_pad=count_include_pad, op="mean")
|
|
|
|
def execute(self, x):
|
|
return self.layer(x)
|
|
|
|
def avg_pool2d(x, kernel_size, stride=None, padding=0, ceil_mode=False, count_include_pad=True):
|
|
return AvgPool2d(kernel_size, stride, padding, ceil_mode, count_include_pad)(x)
|
|
|
|
class MaxPool2d(Module):
|
|
def __init__(self, kernel_size, stride=None, padding=0, dilation=None, return_indices=None, ceil_mode=False):
|
|
self.layer = Pool(kernel_size=kernel_size, stride=stride, padding=padding, dilation=dilation, return_indices=return_indices, ceil_mode=ceil_mode, op="maximum")
|
|
|
|
def execute(self, x):
|
|
return self.layer(x)
|
|
|
|
def max_pool2d(x, kernel_size, stride=None, padding=0, dilation=None, return_indices=None, ceil_mode=False):
|
|
return MaxPool2d(kernel_size, stride, padding, dilation, return_indices, ceil_mode)(x) |