Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Added Hardtanh activation function #14

Merged
merged 1 commit into from
Mar 31, 2023
Merged

Added Hardtanh activation function #14

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 26 additions & 0 deletions dsp/dsp.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -205,6 +205,11 @@ inline float fast_tanh_(const float x) {
(2.44506634652299f + x2) * fabs(x + 0.814642734961073f * x * ax)));
}

inline float hard_tanh_(const float x) {
const float t = x < -1 ? -1 : x;
return t > 1 ? 1 : t;
}

void tanh_(Eigen::MatrixXf &x, const long i_start, const long i_end,
const long j_start, const long j_end) {
for (long j = j_start; j < j_end; j++)
Expand All @@ -227,6 +232,27 @@ void tanh_(Eigen::MatrixXf &x) {
}
}

void hard_tanh_(Eigen::MatrixXf& x, const long i_start, const long i_end,
const long j_start, const long j_end) {
for (long j = j_start; j < j_end; j++)
for (long i = i_start; i < i_end; i++)
x(i, j) = hard_tanh_(x(i, j));
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if it would perform better if you used unaryExpr instead? Like so:
x = x.unaryExpr([](float in) {return hard_tanh(in);});

Seems to do better according to: https://godbolt.org/z/zfePz8MTa

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Well, you don't want to update the entire matrix, so some x.middleCols() are needed

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I wonder if it would perform better if you used unaryExpr instead? Like so:
x = x.unaryExpr([](float in) {return hard_tanh(in);});

Agreed - I just copy/pasted a duplicate of the tanh function structure, but didn't worry about it for now as that code path isn't used by WaveNet - it just uses the full matrix overload and I've already optimized those (for both tanh and hardtanh).

My plan is to refactor the activation code in general soon, but I wanted to get the hardtanh in there asap so we can start think about targeting it with training models.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Btw, I tried using unaryExpr(), and if I recall it performed slightly worse than just rolling over the data directly.

That Compiler Explorer tool you linked to is pretty cool.

}

void hard_tanh_(Eigen::MatrixXf& x, const long j_start, const long j_end) {
hard_tanh_(x, 0, x.rows(), j_start, j_end);
}

void hard_tanh_(Eigen::MatrixXf& x) {
float* ptr = x.data();

long size = x.rows() * x.cols();

for (long pos = 0; pos < size; pos++) {
ptr[pos] = hard_tanh_(ptr[pos]);
}
}

void Conv1D::set_params_(std::vector<float>::iterator &params) {
if (this->_weight.size() > 0) {
const long out_channels = this->_weight[0].rows();
Expand Down
8 changes: 8 additions & 0 deletions dsp/dsp.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -153,6 +153,14 @@ void tanh_(Eigen::MatrixXf &x, const long i_start, const long i_end);

void tanh_(Eigen::MatrixXf &x);

// In-place Hardtanh on (N,M) array
void hard_tanh_(Eigen::MatrixXf& x, const long i_start, const long i_end,
const long j_start, const long j_end);
// Subset of the columns
void hard_tanh_(Eigen::MatrixXf& x, const long i_start, const long i_end);

void hard_tanh_(Eigen::MatrixXf& x);

class Conv1D {
public:
Conv1D() { this->_dilation = 1; };
Expand Down
4 changes: 3 additions & 1 deletion dsp/wavenet.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,9 @@ void wavenet::_Layer::process_(const Eigen::MatrixXf &input,
this->_conv.process_(input, this->_z, i_start, ncols, 0);
// Mix-in condition
this->_z += this->_input_mixin.process(condition);
if (this->_activation == "Tanh")
if (this->_activation == "Hardtanh")
hard_tanh_(this->_z);
else if (this->_activation == "Tanh")
tanh_(this->_z);
else if (this->_activation == "ReLU")
relu_(this->_z, 0, channels, 0, this->_z.cols());
Expand Down