experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Make ApplyTanh3 function more readable. This function is different from the previous one in the input scale.

Browse Source 
PiperOrigin-RevId: 269900289
This commit is contained in:
Jian Li 2019-09-18 15:08:53 -07:00 committed by TensorFlower Gardener
parent 959487fd0e
commit bdcf4b6296
1 changed files with 9 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
tensorflow/lite/kernels/internal/tensor_utils.h
View File

@ -186,7 +186,7 @@ void ApplyLayerNorm(const int16_t* input, const int16_t* layer_norm_weights,
int32_t layer_norm_scale_b, int32_t variance_limit, int32_t layer_norm_scale_b, int32_t variance_limit,
int n_batch, int n_input, int16_t* output); int n_batch, int n_input, int16_t* output);
// Apply Sigmoid to a Quantized vector. // Apply Sigmoid to a quantized vector.
// Parameters: // Parameters:
// - input: batch vector of size n_batch * n_input; 16 bit. // - input: batch vector of size n_batch * n_input; 16 bit.
// - n_batch: the number of batch. // - n_batch: the number of batch.
@ -196,7 +196,7 @@ void ApplyLayerNorm(const int16_t* input, const int16_t* layer_norm_weights,
void ApplySigmoid(const int16_t* input, int32_t n_batch, int32_t n_input, void ApplySigmoid(const int16_t* input, int32_t n_batch, int32_t n_input,
int16_t* output); int16_t* output);
// Apply Tanh to a Quantized vector. // Apply Tanh to a quantized vector.
// Parameters: // Parameters:
// - input: batch vector of size n_batch * n_input; 16 bit. // - input: batch vector of size n_batch * n_input; 16 bit.
// - n_batch: the number of batch. // - n_batch: the number of batch.
@ -206,6 +206,13 @@ void ApplySigmoid(const int16_t* input, int32_t n_batch, int32_t n_input,
void ApplyTanh0(const int16_t* input, int32_t n_batch, int32_t n_input, void ApplyTanh0(const int16_t* input, int32_t n_batch, int32_t n_input,
int16_t* output); int16_t* output);
// Apply Tanh to a quantized vector.
// Parameters:
// - input: batch vector of size n_batch * n_input; 16 bit.
// - n_batch: the number of batch.
// - n_input: the size for input and output.
// - output: the 16 bit output
// The input is in Q3.12 format and the output is in Q0.15 format.
void ApplyTanh3(const int16_t* input, int32_t n_batch, int32_t n_input, void ApplyTanh3(const int16_t* input, int32_t n_batch, int32_t n_input,
int16_t* output); int16_t* output);