Function gelu_forward

experimental/kernels/ops.cpp:311–324

}

void gelu_forward(Context& ctx, float* out, float* inp, int n) {
  unsigned long N = static_cast<unsigned long>(n);
  setLogLevel(kError);
  Tensor input = createTensor(ctx, Shape{N}, kf32, inp);
  Tensor output = createTensor(ctx, Shape{N}, kf32);
  std::promise<void> promise;
  std::future<void> future = promise.get_future();
  Kernel op = createKernel(ctx, {kShaderGelu, 256, kf32},
                           Bindings{input, output},
                           /* nWorkgroups */ {cdiv(N, 256), 1, 1});
  dispatchKernel(ctx, op, promise);
  wait(ctx, future);
  toCPU(ctx, output, out, N * sizeof(float));
}

void gelu_backward(Context& ctx, float* dinp, float* inp, float* dout, int N){
  unsigned long n = static_cast<unsigned long>(N);