Lenoctambule
|
24f1a8c9de
|
Merge pull request #4 from lenoctambule/dev
Labeling test in MNIST example + fixes
|
2026-04-18 19:34:36 +02:00 |
|
Lenoctambule
|
583fc796f6
|
refactor: code de-dup __str__ method
|
2026-04-17 19:53:58 +02:00 |
|
Lenoctambule
|
6eaaa43285
|
fix: use inverse func in p method in Label class
|
2026-04-17 04:52:23 +02:00 |
|
Lenoctambule
|
65c6d3bbee
|
fix: wrong axis typo in Label's observe method
|
2026-04-17 02:31:33 +02:00 |
|
Lenoctambule
|
251d66a625
|
feat: test label accuracy in mnist example
|
2026-04-15 18:13:24 +02:00 |
|
Lenoctambule
|
0f1c9f920b
|
build: update package version 1.2 => 1.3
|
2026-04-14 21:17:31 +02:00 |
|
Lenoctambule
|
bf55c8eb7a
|
Merge pull request #3 from lenoctambule/dev
Add simple post-training labeling + Noise layer
|
2026-04-14 21:11:11 +02:00 |
|
Lenoctambule
|
32a4a39ab9
|
fix: add history mem cap and mid point pruning
|
2026-04-14 21:03:04 +02:00 |
|
Lenoctambule
|
4cc349c22c
|
feat: simple distances instead of std+mean for labeling
|
2026-04-14 20:53:13 +02:00 |
|
Lenoctambule
|
b635bf0467
|
feat: add monte-carlo method and MSE to labeling method
|
2026-04-14 19:24:20 +02:00 |
|
Lenoctambule
|
b1dc34e699
|
feat: post training and online labeling VAE class
|
2026-04-14 14:12:22 +02:00 |
|
Lenoctambule
|
e6c1229a7e
|
fix: return encoding after sample layer
|
2026-04-13 23:07:11 +02:00 |
|
Lenoctambule
|
82acc6b4f3
|
feat: NoiseLayer class + keep loss across trainings
|
2026-04-12 20:31:31 +02:00 |
|
Lenoctambule
|
a4334568ec
|
refactor: separate gradient back and weight updates
|
2026-04-12 19:40:04 +02:00 |
|
Lenoctambule
|
d23f3a903a
|
build: update package version 1.1 => 1.2
|
2026-04-10 22:26:42 +02:00 |
|
Lenoctambule
|
475ca0cf6f
|
Merge pull request #2 from lenoctambule/dev
Refactor of plotting and kb interrupt logic
|
2026-04-10 22:24:10 +02:00 |
|
Lenoctambule
|
7a822782a5
|
refactor: move kb interrupt handling to autoencoder classes
|
2026-04-10 22:20:35 +02:00 |
|
Lenoctambule
|
5ff6cfe55e
|
feat(plotters.py): add VAEPlotter class + seperate training logic
|
2026-04-10 20:37:03 +02:00 |
|
Lenoctambule
|
849d988de5
|
feat(autoencoder.py): __str__ method for VariationalAutoencoder class
|
2026-04-10 15:00:04 +02:00 |
|
Lenoctambule
|
d4679712a6
|
docs: fix typo in README.md
|
2026-04-10 14:33:25 +02:00 |
|
Lenoctambule
|
ea8a4079ac
|
refactor: move plot logic to plotters.py
|
2026-04-09 22:47:22 +02:00 |
|
Lenoctambule
|
9d718a6bc8
|
docs: readme full url to media
|
2026-04-09 19:55:50 +02:00 |
|
Lenoctambule
|
5e83983d96
|
Merge pull request #1 from lenoctambule/dev
Bug fixes and README improvements
|
2026-04-09 19:25:21 +02:00 |
|
Lenoctambule
|
9b42e9db85
|
build: update package version 1.0 => 1.1
|
2026-04-09 19:23:25 +02:00 |
|
Lenoctambule
|
0b46028702
|
docs: illustration of MNIST latent-space representation
|
2026-04-09 19:21:13 +02:00 |
|
Lenoctambule
|
b4963513cc
|
fix: unneeded func params
|
2026-04-09 19:19:59 +02:00 |
|
Lenoctambule
|
075382cfb0
|
fix(mnist_test.py): subplot invalid ncols
|
2026-04-09 18:43:27 +02:00 |
|
Lenoctambule
|
15865812d8
|
refactor(autoencoder.py): __init__ code de-dup
|
2026-04-09 18:31:32 +02:00 |
|
Lenoctambule
|
058b7a0f2a
|
docs: update readme.md w/ remote install
|
2026-04-08 19:08:16 +02:00 |
|
Lenoctambule
|
d048ddc6db
|
feat: packaging, project structure + updated README.md
|
2026-04-08 18:58:32 +02:00 |
|
Lenoctambule
|
e9c79f463f
|
fix: missing args in np.clip
|
2026-04-08 14:49:56 +02:00 |
|
Lenoctambule
|
510ad8720c
|
feat: plot 2d latent space + signal handling + fix SGD in Sampler
|
2026-04-07 22:25:39 +02:00 |
|
Lenoctambule
|
3440de851a
|
fix: dup load method
|
2026-04-05 08:41:17 +02:00 |
|
Lenoctambule
|
82d61dd10f
|
fix: bad type hint and typo for forward method return value
|
2026-04-05 01:44:10 +02:00 |
|
Lenoctambule
|
5a8fb2c48b
|
feat: working implementation of VAE
|
2026-04-05 01:17:51 +02:00 |
|
Lenoctambule
|
577e679425
|
feat: VariationalAutoencoder class + sampling nn layer
|
2026-04-01 22:32:35 +02:00 |
|
Lenoctambule
|
cc74b62afd
|
feat: identity activation func
|
2026-03-30 05:14:02 +02:00 |
|
Lenoctambule
|
a93bb0a692
|
feat: error handling and re-train in mnist_test
|
2026-03-29 21:12:19 +02:00 |
|
Lenoctambule
|
8a3d408b7a
|
docs(README.md): update Autoencoder class usage
|
2026-03-29 19:25:42 +02:00 |
|
Lenoctambule
|
53c7f73055
|
feat: ActivationFunc classes ReLU and LeakyReLU
|
2026-03-29 19:20:05 +02:00 |
|
Lenoctambule
|
44bf4c0286
|
feat: str methods for Autoencoder
|
2026-03-29 09:41:33 +02:00 |
|
Lenoctambule
|
7aabc5db48
|
feat: leaky relu func
|
2026-03-29 09:19:15 +02:00 |
|
Lenoctambule
|
09835e9afa
|
fix: missing activation func derivative + send error before update
|
2026-03-29 08:23:15 +02:00 |
|
Lenoctambule
|
efd328e530
|
feat: save and load methods for Autoencoder
|
2026-03-28 17:50:27 +01:00 |
|
Lenoctambule
|
6155649655
|
feat: loss method + mv data reshaping out of Autoencoder class
|
2026-03-28 02:40:10 +01:00 |
|
Lenoctambule
|
e5520bf050
|
feat: use DeepNNLayer in Autoencoder
|
2026-03-28 02:09:47 +01:00 |
|
Lenoctambule
|
a50a09b337
|
feat: DeepNNLayer class
|
2026-03-28 02:09:47 +01:00 |
|
Lenoctambule
|
6bd230c070
|
licensing: MIT
|
2026-03-27 23:52:05 +01:00 |
|
Lenoctambule
|
eee37c8496
|
refactor: code de-dup
|
2026-03-27 23:15:13 +01:00 |
|
Lenoctambule
|
c37d1c9c26
|
refactor: use tqdm instead of custom load bar
|
2026-03-27 22:16:17 +01:00 |
|