Bausager
be2c193341
Next steps
...
Next up is dropout layers. remember that it can be used to see uncerstaty in multiply output, which is not in the book
2026-05-16 21:36:19 +02:00
Bausager
eb2374eaf5
Backpass on reg
...
Made backpass on regulariaztion and made it easy to see in main.cpp
2026-05-16 21:34:34 +02:00
Bausager
d2fe8aa65c
Fittet new functions to everying in neural networks. Still need to optimise for uint64_t vs int64_t and vec vs mat in some places.
2026-05-16 20:37:05 +02:00
Bausager
48f329feef
Regulaization
...
Sync public mirror / sync (push) Failing after 27s
Started on regulaization in Loss.h. I need to refactor the matsum.h since I need a total sum over the matrix. Also matmul needs a elementwise matmul function, which is the next this in the ragulaization
2026-01-03 22:10:50 +01:00
Bausager
e5f8c91be4
Optimizers - Part 1
...
Sync public mirror / sync (push) Failing after 24s
Done with SGD and Adagrad, still need to optimize them but they work.
2026-01-01 19:23:48 +01:00
Bausager
bd2edea8ef
p.238 in NNFS
Sync public mirror / sync (push) Failing after 24s
2025-12-23 14:47:40 +01:00
Bausager
22d6ea5fad
Backwards Pass
...
Sync public mirror / sync (push) Successful in 26s
Activation Softmax with CategoricalCrossentrophy is almost done from page. 234
2025-10-20 14:23:36 +02:00
Bausager
7fe6be9ba7
Backwards - ReLU & Dense Lasyer
...
Sync public mirror / sync (push) Successful in 27s
Made the backwards for dense layers and ReLU, but I think I soon need to refactor some code to make it all steamlined.
2025-10-09 10:43:45 +02:00
Bausager
c164d790db
NNFS p.135
...
Sync public mirror / sync (push) Successful in 37s
Started on the iteration, not need to rewrite matadd.h and need to make matrandom.h.
2025-10-07 19:06:33 +02:00
Bausager
66b3a4ee6b
CatagoricalCrossentrophy
...
Sync public mirror / sync (push) Successful in 27s
Fixed the loss functions for categories and is ready for accuracy calculation on page 129
2025-10-07 13:09:25 +02:00
Bausager
ea359f3b09
Started Loss, done softmax, up to p.125
...
I've implemented alot of support functions that needs to be refactored, optimised and tested; mean.h, exponential.h, matdiv.h matsum.h matsubtract.h. Maybe we need to have a look at if matdiv/matmul should be in the same. Same with matadd/matsubtract and if some of it should be in matvec.h.
2025-10-05 19:47:56 +02:00
Bausager
88227a38fc
Neural Network
...
Started on implementing neural network from NNFS. I've done ReLU and stopped at p.104. Softmax is not ready.
2025-10-03 20:54:37 +02:00
Bausager