COMPMID-1266 : Add support for FP16 in CLWinogradConvolutionLayer: 5x5 kernels
Introduced F32 accumulation for F16 winograd gemm and output transform
WinogradConvolution will be available for F16 only if fast math flag is enabled
Change-Id: I215593c205236a0f9669218437bb40b184ec6a4f
diff --git a/utils/Utils.h b/utils/Utils.h
index 130e1f7..92ab1a3 100644
--- a/utils/Utils.h
+++ b/utils/Utils.h
@@ -181,6 +181,8 @@
return endianness + "u" + support::cpp11::to_string(sizeof(uint64_t));
case DataType::S64:
return endianness + "i" + support::cpp11::to_string(sizeof(int64_t));
+ case DataType::F16:
+ return endianness + "f" + support::cpp11::to_string(sizeof(half));
case DataType::F32:
return endianness + "f" + support::cpp11::to_string(sizeof(float));
case DataType::F64:
@@ -275,6 +277,43 @@
}
#endif /* ARM_COMPUTE_GC */
+/** Specialized class to generate random non-zero FP16 values.
+ * uniform_real_distribution<half> generates values that get rounded off to zero, causing
+ * differences between ACL and reference implementation
+*/
+class uniform_real_distribution_fp16
+{
+ half min{ 0.0f }, max{ 0.0f };
+ std::uniform_real_distribution<float> neg{ min, -0.3f };
+ std::uniform_real_distribution<float> pos{ 0.3f, max };
+ std::uniform_int_distribution<uint8_t> sign_picker{ 0, 1 };
+
+public:
+ using result_type = half;
+ /** Constructor
+ *
+ * @param[in] a Minimum value of the distribution
+ * @param[in] b Maximum value of the distribution
+ */
+ explicit uniform_real_distribution_fp16(half a = half(0.0), half b = half(1.0))
+ : min(a), max(b)
+ {
+ }
+
+ /** () operator to generate next value
+ *
+ * @param[in] gen an uniform random bit generator object
+ */
+ half operator()(std::mt19937 &gen)
+ {
+ if(sign_picker(gen))
+ {
+ return (half)neg(gen);
+ }
+ return (half)pos(gen);
+ }
+};
+
/** Numpy data loader */
class NPYLoader
{
@@ -416,6 +455,7 @@
case arm_compute::DataType::QASYMM8:
case arm_compute::DataType::S32:
case arm_compute::DataType::F32:
+ case arm_compute::DataType::F16:
{
// Read data
if(!are_layouts_different && !_fortran_order && tensor.info()->padding().empty())
@@ -699,6 +739,18 @@
switch(tensor.info()->data_type())
{
+ case arm_compute::DataType::F16:
+ {
+ std::uniform_real_distribution<float> dist(lower_bound, upper_bound);
+
+ execute_window_loop(window, [&](const Coordinates & id)
+ {
+ *reinterpret_cast<half *>(it.ptr()) = (half)dist(gen);
+ },
+ it);
+
+ break;
+ }
case arm_compute::DataType::F32:
{
std::uniform_real_distribution<float> dist(lower_bound, upper_bound);