You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, I want to verify that my implementation of sru is correct, so I want to know more details.
Can you tell me how to printf in sru_cuda_kernel.cu, for example:
const auto u0 = *up;
const auto u1 = *(up + 1);
const auto u2 = *(up + 2);
const auto wc1 = *vp1;
const auto wc2 = *vp2;
const auto x_val = (skip_type) ? (*xp) : (scalar_t)0.f;
const auto g1 = sigmoidf(u1 + wc1*cur + bias1);
const auto g2 = sigmoidf(u2 + wc2*cur + bias2);
cur = (cur-u0)*g1 + u0;
const auto val = calc_activation(activation_type, cur);
*hp = skip_type ? ((val - x_val) * mask * g2 + x_val) : (val * mask * g2);
I want to know how to printf wc1, wc2 and g1? Can you give me some help?
The text was updated successfully, but these errors were encountered:
Hi, I want to verify that my implementation of sru is correct, so I want to know more details.
Can you tell me how to printf in sru_cuda_kernel.cu, for example:
I want to know how to printf wc1, wc2 and g1? Can you give me some help?
The text was updated successfully, but these errors were encountered: