develop #5

Fabel · 2025-01-29T21:33:51Z

Fabel commented

2025-01-29 21:33:51 +00:00

first fully working version

Fabel added 59 commits 2025-01-29 21:33:51 +00:00

7287ba543f first code commi. First protoyp for AIIA Config

0ba2a3f23c added first dataloader conzept

34b4f35e51 added init File for the DataLoader

530f499efb added "denosing" and "rotation" as pretraining options in the DataLoader

6757718569 improved datahandling

b371d747fd models for training

cbacd5e03c imrpoved Dataloader to actually load the images

4c19838dab converted to cnn models

99c3ec38c7 updated config for cnn

106539f48a uupdated DataLoader for new AIIA CNN arch

b87ce68c82 updated config t add kwrags to support future changes to the config for different models

74973a325b updated models for improved config

26b701fd77 corrected activation function

ab58d352c4 updated saving and first implementation of new additonal parameter handling

6e6f4c4a21 updated models and config to improve parameter handling and adding a copy function to use the same base config for mutliple models

599b8c4835 working shared model (with way to few params)

8ac31c5bf1 improved shared model to have ~10% of params

6f70ebdf53 Merge pull request 'first_pretraining' (#1 ) from first_pretraining into develop

Reviewed-on: Fabel/AIIA#1

de3d58f6db improved cnn

d7d2095d4a Merge pull request 'improved cnn' (#2 ) from improved_shared_cnn into develop

Reviewed-on: Fabel/AIIA#2

32daaadddd updated pretrain method

e5a5618160 correct copying

59b2784e92 fixed spelling error

338ac5dee5 corrected imports

00168af32d corrected init

b7dc835a86 fixed dataloading

9124221346 correct loading files

b5da2e477d correct allogcation for indexes

51da6b5aa2 corrected loading

e0abdb9d39 fixed datasplitting

a8cd9b00e5 limit data loadng to 10k

7a1eb8bd30 updated loader

cae3fa7fb3 proper image transformation

3f6e6514a9 rgba conversion to rgb

7c4aef0978 updated dataloader to work with tupels

8a809269e5 fix inline error

2b55f02b50 eorr handling because we have a tensor misshaping

b501ae8317 new pretraining script

29f0d86ff7 kernel size as large as channel size

3749ba9c5f updated base models MaxPool2D

d205346741 downsized trainingdata from 20k to 1k

32526c3c30 5k image

8dad1d7150 downszied batchsize

b546f4ee27 furhter batchsizing downsizing

6c146f2767 added progressbar for batches

91a568731f removed placeholder collate function

13cf1897ae handle both pretraining methods

b6b63851ca addeed tasks for both denosing and rotation

fe4d6b5b22 corrected viewing and some prints

58baf0ad3c overall improvement

8d08bfc14c removed rint statemetns and added csv saving

102d952043 increased epochs

1e665c4604 added first pip install version 0.1

7de7eef081 updated pretraing to create a extra class for Pretraining

1e79a93a5e Merge pull request 'improved_shared_cnn' (#3 ) from improved_shared_cnn into develop

Reviewed-on: Fabel/AIIA#3

a369c49f15 improved pretraining

3631df7f0a updated pretrainer to handle multiple classes and configs.

3c0e9e8ac1 fixed savings

0d5e5688eb Merge pull request 'first_pip' (#4 ) from first_pip into develop

Reviewed-on: Fabel/AIIA#4

Fabel merged commit a2909002a7 into main

2025-01-29 21:33:57 +00:00

Fabel referenced this issue from a commit

2025-01-29 21:33:59 +00:00

Merge pull request 'develop' (#5) from develop into main

Sign in to join this conversation.

No reviewers

No Label

No Milestone

No project

No Assignees

1 Participants

Notifications

Due Date

The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: Machine-Learning/AIIA#5