julienChemillier
|
6f0b360dc9
|
Remove Adam optimize options by default
|
2023-05-31 11:15:46 +02:00 |
|
julienChemillier
|
7e4f9167fa
|
Fix free issue
|
2023-05-31 08:52:14 +02:00 |
|
julienChemillier
|
3d0af7da2a
|
Merge branch 'Finetuning_deepening' of https://github.com/julienChemillier/TIPE into Finetuning_deepening
|
2023-05-30 16:02:50 +02:00 |
|
julienChemillier
|
ba05923c78
|
Remove an useless allocation of memory
|
2023-05-30 13:52:13 +02:00 |
|
|
|
7e90b15671
|
Merge remote-tracking branch 'upstream/main' into Finetuning_deepening
|
2023-05-28 09:30:52 +02:00 |
|
|
|
d63fb2c870
|
cnn/make: Fix misaligned address
|
2023-05-28 09:26:12 +02:00 |
|
julienChemillier
|
d916c6c86b
|
Rectification of a d_network leak
|
2023-05-28 09:12:52 +02:00 |
|
julienChemillier
|
d7d5a7ae6e
|
Reduce arguments in functions in 'free.c'
|
2023-05-28 09:07:30 +02:00 |
|
|
|
858c071bed
|
cnn-export-cuda: check if nvcc is available
|
2023-05-27 22:26:09 +02:00 |
|
|
|
957c17a1df
|
Add cnn-export-cuda
|
2023-05-27 22:23:50 +02:00 |
|
julienChemillier
|
4ad116511e
|
Fix an argument error in 'backpropagation' file
|
2023-05-27 21:09:04 +02:00 |
|
julienChemillier
|
84e552105a
|
Modification in the structures
|
2023-05-27 20:22:29 +02:00 |
|
julienChemillier
|
208b121c73
|
Add a new classe: 'D_Network'
|
2023-05-26 22:16:26 +02:00 |
|
julienChemillier
|
fade0aa28d
|
Add 'finetuning' variable to the Network class
|
2023-05-26 20:46:58 +02:00 |
|
|
|
771bfcaf70
|
cnn: write of images larger than the input
|
2023-05-26 15:51:05 +02:00 |
|
julienChemillier
|
fe7b111fdf
|
Create LICENSE
|
2023-05-25 17:03:54 +02:00 |
|
julienChemillier
|
5ec7bab488
|
Add finetuning option
|
2023-05-25 16:32:37 +02:00 |
|
|
|
ec138581c2
|
cnn: Add option to disable random offset
|
2023-05-25 13:31:55 +02:00 |
|
|
|
3183339b7a
|
dense: Add random offset option
|
2023-05-25 11:00:49 +02:00 |
|
|
|
a2cd6ef551
|
change comments
|
2023-05-25 10:40:19 +02:00 |
|
|
|
dd9613f159
|
matrix mul:Add comments
|
2023-05-25 10:38:45 +02:00 |
|
|
|
46aa64da01
|
cnn/export:write_image:Increase brightness
|
2023-05-25 10:37:33 +02:00 |
|
julienChemillier
|
10a9714db9
|
Removal of an useless comment
|
2023-05-21 11:54:54 +02:00 |
|
|
|
94c14cedba
|
cnn/config: add FREE_ALL_OPT
|
2023-05-20 22:45:40 +02:00 |
|
|
|
5d306f39ee
|
backpropagation: fix misaligned addresses
|
2023-05-20 20:15:36 +02:00 |
|
|
|
bed3d3123e
|
backpropagation: Fix padding issues
This issue can cause a segfault since 4633329
|
2023-05-20 20:04:22 +02:00 |
|
julienChemillier
|
8e8a57c5b3
|
Fix 'convolution.c' padding issues
|
2023-05-20 15:43:31 +02:00 |
|
|
|
08993ade85
|
cnn/train: add newline & change initialisation
|
2023-05-19 23:29:46 +02:00 |
|
|
|
5738892142
|
Add detailed train timings
|
2023-05-19 21:48:08 +02:00 |
|
|
|
964687d1b4
|
mem management:avoid potential infinite allocation
|
2023-05-19 17:43:56 +02:00 |
|
julienChemillier
|
a4b42445c1
|
Add VGG16 architecture with 227x227 image input
|
2023-05-19 17:28:07 +02:00 |
|
julienChemillier
|
4fcfba5b33
|
Attempt to fix VGG16 architecture
|
2023-05-19 17:27:03 +02:00 |
|
|
|
0aa7df2869
|
Check if file is successfully opened
|
2023-05-18 19:38:56 +02:00 |
|
|
|
2743fcda14
|
cnn: Fix floating point exception ( rand()%0 )
|
2023-05-18 18:09:01 +02:00 |
|
|
|
dbd5362d7d
|
utils:cuda: Select best available GPU
|
2023-05-16 13:21:26 +02:00 |
|
|
|
e1617a72a8
|
cnn: Remove warnings
|
2023-05-16 11:49:46 +02:00 |
|
|
|
4625ad2992
|
mem management: remove warning
|
2023-05-15 18:25:48 +02:00 |
|
|
|
9db15e96ac
|
Update headers and Makefile to reflect 4910137
|
2023-05-15 18:25:29 +02:00 |
|
|
|
491013713d
|
Split creation.c & create models.c
|
2023-05-15 18:23:30 +02:00 |
|
|
|
13e786d34b
|
cnn/neuron_io: track compatibility
|
2023-05-15 15:51:12 +02:00 |
|
|
|
0fb23c9b15
|
cnn/make:dense_linearized: fix wrong indices
|
2023-05-15 12:51:00 +02:00 |
|
|
|
3d7b641965
|
backpropagation:conv: fix misaligned address
|
2023-05-15 12:50:11 +02:00 |
|
|
|
37ba3a5976
|
cnn/convolution: fix misaligned address issue
|
2023-05-15 11:55:50 +02:00 |
|
|
|
2f333bfc1d
|
config.h: reduce default CUDA blocksize
|
2023-05-15 11:36:18 +02:00 |
|
|
|
07369c3905
|
free: Add comment
|
2023-05-15 11:35:24 +02:00 |
|
|
|
e587261e4d
|
test/cnn_convolution: update to newer function
|
2023-05-15 11:34:58 +02:00 |
|
|
|
329e213e1f
|
write image: automatically detect padding
|
2023-05-15 11:34:23 +02:00 |
|
|
|
06abf0bc6b
|
train.c: pick architecture based on dataset type
|
2023-05-15 10:45:14 +02:00 |
|
|
|
19005366d3
|
mem management: add faster way to free all memory
|
2023-05-15 10:44:09 +02:00 |
|
|
|
4cffcc1c95
|
mem management: Add optional tail optimisation
VGG16: memory usage increases of ~1%,
but initialisation time (for CUDA) goes from 1h down to 2.4s
|
2023-05-15 10:07:00 +02:00 |
|
