Commits

Branch selector

 ptr

User selector

All users

Datepicker

All time

Commit History

Commits on Jul 4, 2025

• Initialization of network-wide optimizer no longer needed now that we switched to per-layer optimizer instances

  milancurcic

  committed
  dc55df0

  Copy full SHA for dc55df0

Commits on Jun 20, 2025

• Define optimizer instance per layer to preserve memory across layers

  milancurcic

  committed
  0e11f10

  Copy full SHA for 0e11f10

• get_params_ptr and get_gradients_ptr for conv1d, conv2d, and locally_connected1d

  milancurcic

  committed
  2160f97

  Copy full SHA for 2160f97

Commits on May 30, 2025

• Get weights and weight gradients as 1d

  milancurcic

  committed
  9d68828

  Copy full SHA for 9d68828

Commits on May 27, 2025

• Send the data to optimizer without a copy works for dense layers

  milancurcic

  committed
  21c5707

  Copy full SHA for 21c5707

• WIP optimizer optimization

  milancurcic

  committed
  38896cc

  Copy full SHA for 38896cc

Commits on May 23, 2025

• WIP optimizer refactor w/ pointers

  milancurcic

  committed
  bf1478e

  Copy full SHA for bf1478e

Commits on May 2, 2025

• Generic conv & maxpool (#220)

  Show description for 402b84a
  

  milancurcic

  and

  jvdp1

  authored
  402b84a

  Copy full SHA for 402b84a

Commits on Mar 26, 2025

• Multihead Attention Fixes (#209)

  Show description for 2ed7b6a
  OneAdder

  authored
  2ed7b6a

  Copy full SHA for 2ed7b6a

Commits on Mar 14, 2025

• Generic reshape (#213)

  Show description for 1db0258
  milancurcic

  authored
  1db0258

  Copy full SHA for 1db0258

• Locally connected layer (#201)

  Show description for 5cf14eb
  

  3 peopleauthored
  5cf14eb

  Copy full SHA for 5cf14eb

Commits on Mar 5, 2025

• Embedding Layer (#205)

  Show description for e628d1e
  

  OneAdder

  and

  milancurcic

  authored
  e628d1e

  Copy full SHA for e628d1e

Commits on Feb 25, 2025

• Layernorm (#203)

  Show description for e68e6c2
  

  OneAdder

  and

  milancurcic

  authored
  e68e6c2

  Copy full SHA for e68e6c2

Commits on Feb 21, 2025

• Multihead attention (#199)

  Show description for ed8b340
  

  3 peopleauthored
  ed8b340

  Copy full SHA for ed8b340

• Dropout layer (#194)

  Show description for 039638d
  

  milancurcic

  and

  Vandenplas, Jeremie

  authored
  039638d

  Copy full SHA for 039638d

Commits on Feb 17, 2025

• Linear2d layer (#197)

  Show description for c316ee1
  

  OneAdder

  and

  milancurcic

  authored
  c316ee1

  Copy full SHA for c316ee1

Commits on Feb 16, 2025

• Generic flatten (2d and 3d) (#202)

  Show description for 4ad75bc
  

  milancurcic

  and

  OneAdder

  authored
  4ad75bc

  Copy full SHA for 4ad75bc

• Add Input2d layer (#198)

  Show description for a28a9be
  

  milancurcic

  and

  OneAdder

  authored
  a28a9be

  Copy full SHA for a28a9be

Commits on Sep 13, 2024

• Update for cmake use of neural-fortran (#192)

  Show description for d516437
  

  mathomp4

  and

  milancurcic

  authored
  d516437

  Copy full SHA for d516437

Commits on Sep 3, 2024

• Standard conformance fixes (#190)

  Show description for 9a40117
  milancurcic

  authored
  9a40117

  Copy full SHA for 9a40117

Commits on Aug 31, 2024

• Remove Keras HDF5 (#188)

  Show description for 3872202
  milancurcic

  authored
  3872202

  Copy full SHA for 3872202

Commits on Jun 14, 2024

• Intrinsic `pack` replaced by pointers in `get_params` and `get_gradients` (#183)

  Show description for 118f795
  

  3 peopleauthored
  118f795

  Copy full SHA for 118f795

• Remove some unused variables and a broadcast (#180)

  Show description for a843c83
  

  jvdp1

  and

  Vandenplas, Jeremie

  authored
  a843c83

  Copy full SHA for a843c83

• Proposition of API for the method `network % evaluate` (#182)

  Show description for e82d565
  

  3 peopleauthored
  e82d565

  Copy full SHA for e82d565

Commits on May 9, 2024

• Replacement of a matmul + use of merge (#181)

  Show description for 6dfaed0
  

  3 peopleauthored
  6dfaed0

  Copy full SHA for 6dfaed0

Commits on Apr 19, 2024

• Addidtion of Findneural-fortran.cmake (#178)

  Show description for c3924b5
  jvdp1

  authored
  c3924b5

  Copy full SHA for c3924b5

• Addition of the Loss derived type and of the MSE loss function (#175)

  Show description for f7b6006
  

  3 peopleauthored
  f7b6006

  Copy full SHA for f7b6006

Commits on Apr 18, 2024

• Add instructions using Conda (#172)

  certik

  authored
  cf47114

  Copy full SHA for cf47114

• More conv2d tests (#174)

  Show description for ca219cd
  milancurcic

  authored
  ca219cd

  Copy full SHA for ca219cd

Commits on Aug 29, 2023

• safegaurd Box-Muller normal random number generation against u=0.0 (#158)

  Show description for b1b2cac
  

  dacarnazzola

  and

  milancurcic

  authored
  b1b2cac

  Copy full SHA for b1b2cac

Commits on Aug 6, 2023

• Adagrad Optimizer Implementation (#154)

  Show description for b119194
  

  Spnetic-5

  and

  milancurcic

  authored
  b119194

  Copy full SHA for b119194

Commits on Jul 26, 2023

• Added Adam optimizer implementation (#150)

  Show description for 6adc1c2
  

  Spnetic-5

  and

  milancurcic

  authored
  6adc1c2

  Copy full SHA for 6adc1c2

Commits on Jul 14, 2023

• Added Momentum and Nesterov modifications (#148)

  Show description for e9bfbd6
  

  Spnetic-5

  and

  milancurcic

  authored
  e9bfbd6

  Copy full SHA for e9bfbd6

Commits on Jun 22, 2023

• Connect `flatten`, `conv2d`, and `maxpool2d` layers in backward pass (#142)

  Show description for 6bbc28d
  milancurcic

  authored
  6bbc28d

  Copy full SHA for 6bbc28d

• SGD optimizer stub (#139)

  Show description for 31fc061
  milancurcic

  authored
  31fc061

  Copy full SHA for 31fc061

Pagination

Previous

Next