github.com/HyperGAN/HyperGAN commits | Ecosyste.ms: OpenCollective

[fix] test command

1be4fec3eba2e0452a3cc75ef4b03940a1ddb12a authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] circle.yml

497ab0a0be5b99a69b18aa502a1a2145b54d9959 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] use different g/d if defined in alpha_trainer

61ea87a081bb1cd9a89b23f72ad690f19fe90b21 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] autoencode discriminator config argument

89f30a978be57532dfe96908c15f244c3ac34626 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] hc.io sample

5e8f25b92a0e74713a9f48cbad8fbbb1471c57c5 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] -c least-square-gan-improved

539fe96e8ff819e510f2af1ff77eec6508e303e1 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] no class loss

5470fd4ae92824498cfac4900bbb9e001700e952 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] alpha go is running! probably bugged

a23956d165852a80fec924c4d4f0f04ac13e6567 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] fully connected discriminator

f54fbe18ae39c164ea519adab071da3ab68836ec authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] alphagan (wip, completely broken)

c31c2f72f7861dc7a1ca70e1dfdb47725e5f6955 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] improved-gan-resizable

bf3aea029feb0ecbd88a65a80638baddbfde0dc5 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] make tests run again

e6a739388f7e6a828756c74460deb47fe3ca64e2 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] autoencode discriminator config argument

1bc761d63490fb975f865fd92c1ed3158e599447 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] remove old json configs

964a2b401d35a9eed3b98fae5595df04e0734a52 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] block deconv and fix for began --autoencode

917e9337c06c9d1094b46d5febb49551421ecb66 authored over 7 years ago by Martyn Garcia <[email protected]>

Merge pull request #92 from 255BITS/text-onehot

Character GAN

11b681d9fd273a748fefec10f88e7e10bbfb3542 authored over 7 years ago by martyn <[email protected]>

[fix] viewer randomly dieing because it needis a thread lock

3138d6696856b4ddc467e86f2daef764e8628849 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] -c began-optimized

93721b00102a5f313875b562fa6d218ec252c1a6 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] -c began-optimized

d97cfab66b83c325a79cfa8b270eb8f0e1dc1a0d authored over 7 years ago by Martyn Garcia <[email protected]>

[refactor] extract TextInput from each sampler(not working)

5b09573779559460617a1e57f34c64fd89610dfb authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] text one hot example to work with 0.9

fe6dfc261aa56428cd655d23d8a1c672a4943590 authored over 7 years ago by Martyn Garcia <[email protected]>

Merge branch 'develop' into text-onehot

* develop: (260 commits)
[refactor] remove dead files and prints
[refactor] hypergan.sampler...

fe0976b13b84b4d72f4251ea55598ddafe11a25f authored over 7 years ago by Martyn Garcia <[email protected]>

[refactor] remove dead files and prints

d2bdaf0094b737bc2790fc8af84c075317b012a8 authored over 7 years ago by Mikkel Garcia <[email protected]>

[refactor] hypergan.samplers.viewer -> hypergan.viewer

fffa174ae5d04072c076cd5a8758e9214a07585d authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] initial autoencoder test

20859cf5ba229c3289f63fafd0e88e2cd7103f57 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] random search generator/discriminator

d7aad59842a12e8d877b642cfcc282b565d50df7 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] memory leak in viewer

f94ef2414fdfb9d6114909f271c3adc8d395f65f authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] closer to replicating began paper

9a057a199c912e72b4adf53fad760c63cbcb7159 authored over 7 years ago by Mikkel Garcia <[email protected]>

Merge pull request #84 from 255BITS/refactor-ops

Core refactorings. Much much nicer API. Prepackaged configs added including improved-wgan(with...

3ebb48285bd76292948897ce6f53dfa0e2dd4746 authored over 7 years ago by martyn <[email protected]>

[fix] default g/d

2f49f5d198bf4ff0216227617c2929bff04228e9 authored over 7 years ago by Martyn Garcia <[email protected]>

[refactor] remove unused examples

5568a30d121285ac7907a83b6eb911d0e0e22cc8 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] --save_samples defaults to False

d995f6b03d93e46d8339adeab6313f818bf26120 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] relation layer

3d39f3af84a27ce9410779ddb6892b354bde6e02 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] --aligned and --autoencode

a1333e84dfd6eb16922b52663060c328073039b2 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] alignment sampler

a104b7406e4b41688dcc1a2bb66ffda57251b7d9 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] alignment test

6a65e91b8303144eda83efca5f5a4d9806fba6b7 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] one too many layers on G

1e96eafa3b95556ebbd4bc532d21e7a01ab7f2c2 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] resize conv and some loss functions

dba63484a8dca89d56c9bb3782b7a44953a9e066 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] resize conv

11fdb08515c6582cc75d054fdb049a5f919d1c94 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] least squares uses dcgan

daa912cf5bb17a3b42103d0719607fcfabb43b2b authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] wgan / improved wgan / dcgan architecture. more

0927dff9fe1631564f4c750b81e42c9ea91a8bda authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] merge conflict

01ec8a5c29932ec55b82b0a362b2e86fb005ebe1 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] s/--viewer/--noviewer

1e6bf8c88edf0485f135f317da978c0f29312ee3 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] generator final_activation optional

e17e7845a0621717f52707afc75cdfdc584fe269 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] split_batch returns 2 dimensional tensor

16d9b2e4cdb8e8b06a018e99fb758be0346a29c3 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] share cramer discriminator ops

955da1a7f32ba8d24f2f246df6326f5987cac071 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] activations work

c587abd6a3aff881cf3012249b33d5f90fed5a14 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] initial cramer gan configuration(needs debugging)

2c8a2dfbf3fc6afc0e6c4c51dbbf7b5224ca6db2 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix refactor] cramer loss

ca44e43ef11631a4b05881ac88d3e2fdef2b2d32 authored over 7 years ago by Martyn Garcia <[email protected]>

Merge branch 'develop' into refactor-ops

* develop:
[refactor] dead code
[tweak] search space
[feature] cramer gan https://arxiv.or...

6ac96ef19bc7d87b680a07bffa3f706bca066031 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] sampler readded

da5c8d228db1eca806c2b28d2e3f3799d597b6aa authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] colorizer re-working

9ba9d15299d92c8ba74f83b5b01e4da13c2debdd authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] -c began

1c7a042561821c57022c9c3aa02c5aa6added437 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] add publication and description link to every gan configuration

2aba79f4ecb757d31382e3b117ddba95d436367b authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] -c improved-wgan

98003ce07d535e8744d9f57b8bcc47ed2132a474 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] discriminator add noise

0f04fcfd2501c54d1e4880518cb358b53463dec2 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] gradient penalty math

e67bb00bdff9afa218326dd9001236a922512e3c authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] relational_layer - experimental https://arxiv.org/abs/1706.01427

48a85c09417ddc535e30d41b8a49f7e2c049b983 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] show config name in window title

2d6a44f3ed6371c429c85f4f404efda638e3e514 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] lots of bugs in multicomponent

65adbd131bd27892297804efa505810f3167ebe4 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] multicomponent

9950f411a0ec5ae5c9b98959a84f25fd10f1c9ea authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] tests

10de27ad077921010230585999f4753397d7a2eb authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] dont readd weights if reusing

971ab74619bcf98385b4ab036260a9875d38626f authored over 7 years ago by Mikkel Garcia <[email protected]>

[refactor] bin/hypergan creates the GAN object

f3100abb657c3261609174f9e422d4abb392f5f4 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] resize conv generator / layer norm

b1aa3efde5d97925208d334c2c4fd6836386baee authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] more tests

5ac175b705ff7e3fbc9af048cc9f2e748b82429f authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] supervised loss

69c83ab19b3e596845b85171e1c8a4d1a8ecb76b authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] tests

892ce17376d945a0800595642dcee4aae0252417 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] -c wgan

60ff04cce63d12343cf2b77760b74b9985a5a22b authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] initial began configuration(broken)

d501e05a2e2c939003f981824f76e92c6725ef89 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] disable gradient penalty for now in random search

d679cc80af13b61c79419cfab7ce52aabb31688d authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] crash

1109950e6eef31f84a2f37e2d61a9b5d37440775 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] configuration lookup

845660746190c03611398294d4772fffc04f68eb authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] some tests

16be55ed98340f9494b2eadb565dc18d6cf47d9c authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] -c softmax-gan

5f97196a9acf0ad7bdd70e0523d021158a9b92e8 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] discriminator

a27c5c3fef63f63e8ce5b0b4bb7236b93c2d007f authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] add search through gradient penalty and minibatch settings

ce83e533ae0cddac5a9a22b81ffc828db132e22e authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] least squares loss and reference configuration

2ad6247e199a2373919e9b4c66cac95351fd8c07 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] sample when starting hg

3971e1de2a76b0ff55d36b8f1c33635b65e2cba4 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] layer_regularizer interface

011d4384d8d6aa3e8fd36bde88837a97c43e0790 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] -c improved-gan

84fa485c27d8f68649dc31cf171fa635eddf6519 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] -c dcgan

d3872eb28ebf5f188c80294d1b9c1ac25fd0a15e authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] autoencoder test

6301b1e6e0b3bd988e46cbaf97e57acb10e0a6da authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] more random search options

6635b009bf55f1d86f5d769e7d01cf7e37a2d021 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] deconv

3fcc53d69e8974a5292ecce95bc61007a3a15671 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] use tf contrib batch norm

36a8b4de071291766ebb8376a1184b3557d3cf8f authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] standard loss working

1ea7369de4b3e4caf481b31520b9cbeb31b78d36 authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] generator block built incorrectly

47ee4646b5e4a341f1623f44a5c167ba10bdb607 authored over 7 years ago by Mikkel Garcia <[email protected]>

[refactor] reuse() in base gan_component

4462ae67f7d4d23b6bc982d437209f261f31f366 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] support other base configs

1c8d0bdbb2c66d4e68b31caad482fe2cb93c9303 authored over 7 years ago by Mikkel Garcia <[email protected]>

[feature] default to default.json config if unspecified.

377d5389150563da248e74f1cbac52851572005a authored over 7 years ago by Mikkel Garcia <[email protected]>

[fix] autoencoder - allocate weights on the discriminator

d0ac6704e43b5660c43858b7169f54cd98824a21 authored over 7 years ago by Martyn Garcia <[email protected]>

[optimize] autoencoder discriminator

25e57a1d525fffe215b6226f9206d1f51cbc5e58 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] package

cc962f4fdadf14c4bea74695065e9be7dd1d6141 authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] add autoencoder discriminator(for began)

32992655c76566c4be1f586701adb015cac63fbc authored over 7 years ago by Martyn Garcia <[email protected]>

[feature] AutoencoderDiscriminator

54e7e978f4c3ab3c100c8a8f4dfcecf86274f513 authored over 7 years ago by Martyn Garcia <[email protected]>

[fix] load config

c890487ae56ec0ba60e258fc073ae6a34e078d76 authored over 7 years ago by Martyn Garcia <[email protected]>

Merge branch 'develop' into refactor-ops

* develop:
[feature] softmax loss search
[fix] lsgan loss
[tweak] began learning rate

a8ad7b43b72702cdeb4a95eca03d8862021eb80a authored over 7 years ago by Martyn Garcia <[email protected]>

Merge pull request #93 from 255BITS/cramer-loss

Cramér GAN

a1c5d50807c9ccf7cba20fadf045632af53741fd authored over 7 years ago by martyn <[email protected]>

[fix] hypergan working

5835b5d552468f40bc953847857dc0a7d28f57d6 authored over 7 years ago by Martyn Garcia <[email protected]>