2021-10-28 16:21:23 +00:00
|
|
|
// modified from https://github.com/NVIDIA/apex/blob/master/csrc/multi_tensor_adam.cu
|
|
|
|
#include <ATen/ATen.h>
|
|
|
|
#include <ATen/AccumulateType.h>
|
|
|
|
#include <ATen/cuda/CUDAContext.h>
|
|
|
|
#include <ATen/cuda/Exceptions.h>
|
|
|
|
// Another possibility:
|
|
|
|
// #include <torch/all.h>
|
|
|
|
|
|
|
|
#include <assert.h>
|
|
|
|
|
|
|
|
#include "type_shim.h"
|
|
|
|
#include "multi_tensor_apply.cuh"
|
|
|
|
|
|
|
|
#define BLOCK_SIZE 512
|
|
|
|
#define ILP 4
|
|
|
|
|
|
|
|
typedef enum
|
|
|
|
{
|
|
|
|
ADAM_MODE_0 = 0, // L2 regularization mode
|
|
|
|
ADAM_MODE_1 = 1 // Decoupled weight decay mode(AdamW)
|
|
|
|
} adamMode_t;
|
|
|
|
|
|
|
|
using MATH_T = float;
|
|
|
|
|
2022-03-25 06:15:53 +00:00
|
|
|
template <typename T_g, typename T_p>
|
2021-10-28 16:21:23 +00:00
|
|
|
struct AdamFunctor
|
|
|
|
{
|
|
|
|
__device__ __forceinline__ void operator()(
|
|
|
|
int chunk_size,
|
|
|
|
volatile int *noop_gmem,
|
|
|
|
TensorListMetadata<4> &tl,
|
|
|
|
const float beta1,
|
|
|
|
const float beta2,
|
|
|
|
const float beta1_correction,
|
|
|
|
const float beta2_correction,
|
|
|
|
const float epsilon,
|
|
|
|
const float lr,
|
|
|
|
adamMode_t mode,
|
|
|
|
const float decay)
|
|
|
|
{
|
|
|
|
// I'd like this kernel to propagate infs/nans.
|
|
|
|
// if(*noop_gmem == 1)
|
|
|
|
// return;
|
|
|
|
|
|
|
|
int tensor_loc = tl.block_to_tensor[blockIdx.x];
|
|
|
|
|
|
|
|
// potentially use to pass in list of scalar
|
|
|
|
// int tensor_num = tl.start_tensor_this_launch + tensor_loc;
|
|
|
|
|
|
|
|
int chunk_idx = tl.block_to_chunk[blockIdx.x];
|
|
|
|
int n = tl.sizes[tensor_loc];
|
|
|
|
|
2022-03-25 06:15:53 +00:00
|
|
|
T_g *g = (T_g *)tl.addresses[0][tensor_loc];
|
2021-10-28 16:21:23 +00:00
|
|
|
g += chunk_idx * chunk_size;
|
|
|
|
|
2022-03-25 06:15:53 +00:00
|
|
|
T_p *p = (T_p *)tl.addresses[1][tensor_loc];
|
2021-10-28 16:21:23 +00:00
|
|
|
p += chunk_idx * chunk_size;
|
|
|
|
|
2022-03-25 06:15:53 +00:00
|
|
|
T_p *m = (T_p *)tl.addresses[2][tensor_loc];
|
2021-10-28 16:21:23 +00:00
|
|
|
m += chunk_idx * chunk_size;
|
|
|
|
|
2022-03-25 06:15:53 +00:00
|
|
|
T_p *v = (T_p *)tl.addresses[3][tensor_loc];
|
2021-10-28 16:21:23 +00:00
|
|
|
v += chunk_idx * chunk_size;
|
|
|
|
|
|
|
|
n -= chunk_idx * chunk_size;
|
|
|
|
|
|
|
|
// see note in multi_tensor_scale_kernel.cu
|
|
|
|
for (int i_start = 0;
|
|
|
|
i_start < n && i_start < chunk_size;
|
|
|
|
i_start += blockDim.x * ILP)
|
|
|
|
{
|
|
|
|
MATH_T r_g[ILP];
|
|
|
|
MATH_T r_p[ILP];
|
|
|
|
MATH_T r_m[ILP];
|
|
|
|
MATH_T r_v[ILP];
|
|
|
|
#pragma unroll
|
|
|
|
for (int ii = 0; ii < ILP; ii++)
|
|
|
|
{
|
|
|
|
int i = i_start + threadIdx.x + ii * blockDim.x;
|
|
|
|
if (i < n && i < chunk_size)
|
|
|
|
{
|
|
|
|
r_g[ii] = g[i];
|
|
|
|
r_p[ii] = p[i];
|
|
|
|
r_m[ii] = m[i];
|
|
|
|
r_v[ii] = v[i];
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
|
|
|
r_g[ii] = MATH_T(0);
|
|
|
|
r_p[ii] = MATH_T(0);
|
|
|
|
r_m[ii] = MATH_T(0);
|
|
|
|
r_v[ii] = MATH_T(0);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#pragma unroll
|
|
|
|
for (int ii = 0; ii < ILP; ii++)
|
|
|
|
{
|
|
|
|
if (mode == ADAM_MODE_0)
|
|
|
|
{ // L2
|
|
|
|
r_g[ii] = r_g[ii] + (decay * r_p[ii]);
|
|
|
|
r_m[ii] = beta1 * r_m[ii] + (1 - beta1) * r_g[ii];
|
|
|
|
r_v[ii] = beta2 * r_v[ii] + (1 - beta2) * r_g[ii] * r_g[ii];
|
|
|
|
MATH_T next_m_unbiased = r_m[ii] / beta1_correction;
|
|
|
|
MATH_T next_v_unbiased = r_v[ii] / beta2_correction;
|
|
|
|
MATH_T denom = sqrtf(next_v_unbiased) + epsilon;
|
|
|
|
MATH_T update = next_m_unbiased / denom;
|
|
|
|
r_p[ii] = r_p[ii] - (lr * update);
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{ // weight decay
|
|
|
|
r_m[ii] = beta1 * r_m[ii] + (1 - beta1) * r_g[ii];
|
|
|
|
r_v[ii] = beta2 * r_v[ii] + (1 - beta2) * r_g[ii] * r_g[ii];
|
|
|
|
MATH_T next_m_unbiased = r_m[ii] / beta1_correction;
|
|
|
|
MATH_T next_v_unbiased = r_v[ii] / beta2_correction;
|
|
|
|
MATH_T denom = sqrtf(next_v_unbiased) + epsilon;
|
|
|
|
MATH_T update = (next_m_unbiased / denom) + (decay * r_p[ii]);
|
|
|
|
r_p[ii] = r_p[ii] - (lr * update);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#pragma unroll
|
|
|
|
for (int ii = 0; ii < ILP; ii++)
|
|
|
|
{
|
|
|
|
int i = i_start + threadIdx.x + ii * blockDim.x;
|
|
|
|
if (i < n && i < chunk_size)
|
|
|
|
{
|
|
|
|
p[i] = r_p[ii];
|
|
|
|
m[i] = r_m[ii];
|
|
|
|
v[i] = r_v[ii];
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
};
|
|
|
|
|
|
|
|
void multi_tensor_adam_cuda(
|
|
|
|
int chunk_size,
|
|
|
|
at::Tensor noop_flag,
|
|
|
|
std::vector<std::vector<at::Tensor>> tensor_lists,
|
|
|
|
const float lr,
|
|
|
|
const float beta1,
|
|
|
|
const float beta2,
|
|
|
|
const float epsilon,
|
|
|
|
const int step,
|
|
|
|
const int mode,
|
|
|
|
const int bias_correction,
|
|
|
|
const float weight_decay)
|
|
|
|
{
|
|
|
|
using namespace at;
|
|
|
|
|
|
|
|
// Handle bias correction mode
|
|
|
|
float bias_correction1 = 1.0f, bias_correction2 = 1.0f;
|
|
|
|
if (bias_correction == 1)
|
|
|
|
{
|
|
|
|
bias_correction1 = 1 - std::pow(beta1, step);
|
|
|
|
bias_correction2 = 1 - std::pow(beta2, step);
|
|
|
|
}
|
|
|
|
|
2022-03-25 06:15:53 +00:00
|
|
|
DISPATCH_FLOAT_AND_HALF_FOR_G_P(
|
|
|
|
tensor_lists[0][0].scalar_type(),
|
|
|
|
tensor_lists[1][0].scalar_type(), 0, "adam",
|
2021-10-28 16:21:23 +00:00
|
|
|
multi_tensor_apply<4>(
|
|
|
|
BLOCK_SIZE,
|
|
|
|
chunk_size,
|
|
|
|
noop_flag,
|
|
|
|
tensor_lists,
|
2022-03-25 06:15:53 +00:00
|
|
|
AdamFunctor<g_scalar_t_0, p_scalar_t_0>(),
|
2021-10-28 16:21:23 +00:00
|
|
|
beta1,
|
|
|
|
beta2,
|
|
|
|
bias_correction1,
|
|
|
|
bias_correction2,
|
|
|
|
epsilon,
|
|
|
|
lr,
|
|
|
|
(adamMode_t)mode,
|
|
|
|
weight_decay);)
|
|
|
|
|
|
|
|
AT_CUDA_CHECK(cudaGetLastError());
|
|
|
|
}
|