augustin64/tipe
1
1
Fork 0
mirror of https://github.com/augustin64/projet-tipe synced 2025-01-23 15:16:26 +01:00
Code Issues Packages Projects Releases Wiki Activity

cnn/make:dense_linearized: fix wrong indices

Browse Source
This commit is contained in:
augustin64 2023-05-15 12:51:00 +02:00
parent 3d7b641965
commit 0fb23c9b15
2 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/cnn/make.c
View File

@ -257,7 +257,7 @@ __global__ void make_dense_linearized_kernel(float** weights, float* bias, float
for (int i=0; i < input_depth; i++) { for (int i=0; i < input_depth; i++) {
for (int j=0; j < input_width; j++) { for (int j=0; j < input_width; j++) {
for (int k=0; k < input_width; k++) { for (int k=0; k < input_width; k++) {
f += input[i][j][k]*weights[k + j*input_width + i*input_depth][idx]; f += input[i][j][k]*weights[k + (i*input_width+j)*input_width][idx];
} }
} }
} }
@ -285,7 +285,7 @@ void make_dense_linearized_cpu(Kernel_nn* kernel, float*** input, float* output,
for (int i=0; i < input_depth; i++) { for (int i=0; i < input_depth; i++) {
for (int j=0; j < input_width; j++) { for (int j=0; j < input_width; j++) {
for (int k=0; k < input_width; k++) { for (int k=0; k < input_width; k++) {
f += input[i][j][k]*kernel->weights[k + j*input_width + i*input_depth][l]; f += input[i][j][k]*kernel->weights[k + (i*input_width+j)*input_width][l];
} }
} }
} }

4
src/cnn/make.cu
View File

@ -257,7 +257,7 @@ __global__ void make_dense_linearized_kernel(float** weights, float* bias, float
for (int i=0; i < input_depth; i++) { for (int i=0; i < input_depth; i++) {
for (int j=0; j < input_width; j++) { for (int j=0; j < input_width; j++) {
for (int k=0; k < input_width; k++) { for (int k=0; k < input_width; k++) {
f += input[i][j][k]*weights[k + j*input_width + i*input_depth][idx]; f += input[i][j][k]*weights[k + (i*input_width+j)*input_width][idx];
} }
} }
} }
@ -285,7 +285,7 @@ void make_dense_linearized_cpu(Kernel_nn* kernel, float*** input, float* output,
for (int i=0; i < input_depth; i++) { for (int i=0; i < input_depth; i++) {
for (int j=0; j < input_width; j++) { for (int j=0; j < input_width; j++) {
for (int k=0; k < input_width; k++) { for (int k=0; k < input_width; k++) {
f += input[i][j][k]*kernel->weights[k + j*input_width + i*input_depth][l]; f += input[i][j][k]*kernel->weights[k + (i*input_width+j)*input_width][l];
} }
} }
} }