Ecosyste.ms: OpenCollective

An open API service for software projects hosted on Open Collective.

github.com/HyperGAN/HyperGAN

Composable GAN framework with api and user interface
https://github.com/HyperGAN/HyperGAN

[doc] bump version

9a7dcc98c9ece732ae4900afffecb40b1774350f authored over 5 years ago by martyn <[email protected]>
[doc] bump bersion

a01b4205391cde03026f089766bd33238e731097 authored over 5 years ago by martyn <[email protected]>
[fix] adversarial robustness on standard gan

8b73e2e44be62a76d18f400949ddc57c4ad15e5e authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] dfgan config

24b54ae6b3ebe99e4365f3f7ee6e0830fe79683c authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] FitnessTrainer error on load

9d0c55c3184b891d3e6c2b8c31c4a9c9199c9f7c authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] Max Gp config

351b25088d65a640a90c3e36df3554b1e1882a85 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] progress compress config

f98d7405c9496fb6e6e204a24d3654a6990225c1 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] script to run all configurations

934cd5efffbe2e6fce1c758640ae08eeca7e4a4c authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] sgd config

a2602066429862cc5a9743b798f470f5d306e598 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] add hypergan_version and runtime to configs

00b9fae45777b2dba4bb9eb186b2f09bbfa30288 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] remove unstable wgan config

8dbd79f1358475e375bd40687b3088ede7856fb9 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] self-supervised

eefc2980016e03e56e38b4c2f8ffd45c973e1046 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] wgan-gp

d17488d8ec1f318a4da92c9d3172474cfaac2567 authored over 5 years ago by Mikkel Garcia <[email protected]>
Merge branch 'master' into develop

* master:
Update README.md
[fix] resize generator
[fix] resize generator
[fix] no layer_...

eb1678ef7c98ecd1dc01ea8356d8b40a13abc27a authored over 5 years ago by Mikkel Garcia <[email protected]>
Update README.md

3b6846be281a363cf9f0aad4a6df5122f8a4f4d5 authored over 5 years ago by mikkel <[email protected]>
[fix] resize generator

711151a64da262ad4cf78e034cdbe858b7f0bebe authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] resize generator

72eba76dcf21c75b456eb445515511a941e6700b authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] no layer_filter for dcgan_generator

94068f9b4e819cbb2f503c743ab4d88a4d310231 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] X and G sizes differ validation exception

aba2dbc15b6c7523c2f5c0c65f7d29d29ad22f98 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] resize generator

3b8b20a464a5af18470090e443c5f00c5d8f486b authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] resize generator

c2e92c28c1e6eec528efd4b309bd807c6545a2b6 authored over 5 years ago by Martyn Garcia <[email protected]>
Merge pull request #222 from HyperGAN/develop

0.10 alpha 23 WIP

c2a62ac4a2b1e32662130272b01f4d3da2208e75 authored over 5 years ago by mikkel <[email protected]>
Merge branch 'master' into develop

944c0e809d90c8be36357ebb302772b24695cc24 authored over 5 years ago by mikkel <[email protected]>
[fix] potential memory leak when using tk viewer

e18e25ea99de18716eb7e59bc93692cf249c9708 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] potential memory leak when using tk viewer

a8f8ba9b1d8ab6f456ef2a6e482f50ff911eceee authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] updated colorizer config

[refactor] layer_filter moved from 'base' components to 'configurable'

65fcf0f193123e7c630da5b5a12726fb9fbd4ffe authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] colorizer sampling

d2c2371cc9a6f3b1d7e7806898044ad59cef4617 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] resizable_generator "resize_conv" sizing

fbd17d8e2de3bc77b8d601f5eb7ae2751cce13e0 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] updated colorizer.json

d85a893e941b038841f1f749b45d47ec68d895ab authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] updated colorizer.json

055b1235cfc85662344ecd9f627a5832655bb908 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] rgba (alpha channel) support - set size to 4 channel ie. 256x256x4

c8db4e842972f4c19b75a935a3905e36ea175b0a authored over 5 years ago by Martyn Garcia <[email protected]>
[feature] resizable generator new options adaptive_instance_norm_layers and block resize_conv

e0fa8da3076377c51e4e3a0d3f92eecc11755dd6 authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] resizeable generator

98388dd58296cf56ad710db1061727e11903b7db authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] resizeable generator

34137e2f11188462590d89c47021a37f7d301729 authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] default.json and resizeablegenerator subpixel support

3b84a7ba29de5e6b8ed39cd02bedc69f8923743e authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] rename file

49d767e0547343a541fc96e2b47cf3eb880e1e3f authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] second autoencode config (with skip connections)

35017d4a66c7ec0dc391bc79dacb40c9c215a8c8 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] custom adabound 256x256 config

18b65de99912b13b9f4fcead1c55d807c10dd674 authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] trainer errors

64f7c36c9b5d9396507d788b8e6089a9342d3a07 authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] set attributes on aligan

70ba4aac56cb057ca2a646631f0f4cdb28826e27 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] "layer x src=component" src argument added

47a6311250ee7c84a0b83912301770f33e15d589 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] aligan config [feature] depth_trainer freeze_latent [feature] depth_trainer heuristic measuring relative gradient magnitude

d99ca0df2aedad98eb6004c58ad766a63f4c42a7 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] call train hooks in batch_fitness_trainer

5763f24fcdd09783fb45070e6e1c381d7f0cc977 authored over 5 years ago by Mikkel Garcia <[email protected]>
[refactor] use depth_trainer in autoencode

d55ac322e5b6349b6ba2ae88c978d5475585fbad authored over 5 years ago by Mikkel Garcia <[email protected]>
[experiment] reset_optimizer before each stepcount

b87cfe1f0312da144b15e5ea5b44653495907faa authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] configuration: aligan autoencoder

6222adeda235621b9145ac9b57fb4af503ce6340 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] xa_hat and xb_hat

3b90ad07a881cc3294d26c99f6ea69671eac8ca5 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] terms

213d6dc93efb3c92962ec152b2a07a138f2c7964 authored over 5 years ago by Mikkel Garcia <[email protected]>
[experiment] aligan

39d77e726e7ea96bf114881b1f26419e977d91ff authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] config option

4a1b9df21963c363a9020e80da30fbf032f9e209 authored over 5 years ago by Martyn Garcia <[email protected]>
[experiment] new loss targets for match_support

8425d9a9291dbc039cd65896b33e3dcd9f5c6086 authored over 5 years ago by Mikkel Garcia <[email protected]>
[experiment] aligned_ali_gan8.mirror_joint

cd4ab271ed30168b9b540d7f1919d33e1a8b8e81 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] revert class name

40cfa19742342296e0b873681028451443c447bd authored over 5 years ago by Martyn Garcia <[email protected]>
[feature] skip_restore for input_fitness_train_hook

e63507a737c5f7f51efba782772b65ec9cef2d88 authored over 5 years ago by Martyn Garcia <[email protected]>
[experiment] alignment

16ac109595df1775f6448e86e9268952e423497f authored over 5 years ago by Martyn Garcia <[email protected]>
[experiment] alignment

412a75e6618f5ee73b8b182e552833d1f1e6172f authored over 5 years ago by Martyn Garcia <[email protected]>
[config] match_support example

42f52eb0a27efdc53657694a023cd5755a75a7d2 authored over 5 years ago by Mikkel Garcia <[email protected]>
[experiment] parameters for match_support train hook

a153f985b8d71486049fb0509d109012955f06fb authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] support search_steps = 0 for input fitness

33f273dc89130a455bf70d7966fbdd13977ef94b authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] negative momentum optimizer

2d06455879b175af06c23df274a65089d36e4adf authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] search heuristic for learn rate on match_support optimizer

dde64019ba3c48c68def58c77cf935669dc92b77 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] match_support parameterization and convergence end condition

77e2e8f781b41a02866ca9a32371e50dbbbc4a21 authored over 5 years ago by Mikkel Garcia <[email protected]>
[experiment] match_support trains X and G before each step

594b14c6a5bd7d78a38ea1110adddc02d75feb7d authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] zero_penalty_train_hook applied l2 loss to zero, used in previously committed config

c4a41d561a211633d2e8290d4271cc52f92a243d authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] add dfgan noise layer to debug sampler

99483e54b3076930498fcc211470f56dd84c44d5 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] distribution filtering gan https://arxiv.org/pdf/1906.04612v1.pdf

96d1c04a9a1d00659cc14c0aa3100b945cd1fb0e authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] l2nn configuration and l2nn changes for stability

a1d44e13820582072bcfa9f0654714c96b226bd3 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] depth trainer call train hooks

8bf401abd4a1aea4d2f4400cfd5483d79030853f authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] force equilibrium train hook

9b5636f3738237e7310ac7f8e0c41be6fc7851c7 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] turn on optimization for building (size = 1/4)

e1b3e93e95a7e559d3db25ae8bc366499f83664b authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] support faster fitness search with heuristic = zero

b49b79476ba14a2cc400bc69240a5820984688f1 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] heuristic integer lets the fitness search until no new best fit entries are found

21608c70f94278f4799cc3bd01c8908e58c24615 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] sorting by d_real

641ddf7d2c3ea2e1ff953d0d0e38d5f917853c04 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] slice layer for slicing z

6a02779b9d8e0827f7a46576565066e4b668520f authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] sorting method for input fitness train hook

be8f13e0e5ec11fbd74b992d0ab4514ccc753106 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] configurable component layer `split`

403256d6ea70a22079d5fb78665e89508908ed90 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] adabound

c17d78fee79285a86f588d0310324a667fd5a756 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] Sample->refresh does not save files

d50faa05fc61ad80b472544fb20de747f9c4c26e authored over 5 years ago by Mikkel Garcia <[email protected]>
[refactor] sort layers

f737b4b867db50e9cf74c4475ad665447bddc136 authored over 5 years ago by Martyn Garcia <[email protected]>
[fix] ali/ali8 experiments

a9b36562c13158380e62244b82d04b8524fe08a6 authored over 5 years ago by Martyn Garcia <[email protected]>
[feature] some configurable component layers

7910d16e58e29e0cffc46a50088a6db4bc018f2d authored over 5 years ago by Martyn Garcia <[email protected]>
[feature] repeat config keyword. example: decay(steps=100 range=0:1 repeat)

8d4800c16232282c9074dde0b3925745d287d417 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] use gan.configurable_params in adabound

1d7e980c0cb59dffca3fca5c1fd0b51a21092fbd authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] gan loss_type for adversarial_robust_train_hook

8cbf5ab1e789886e23214572ab314aae3b704faa authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] lambdas on adversarial examples

dcc9a315696f7753a1fa3bc3ed92142c8185de99 authored over 5 years ago by Mikkel Garcia <[email protected]>
[experiment] adversarial robust train hook https://openreview.net/pdf\?id\=HJE6X305Fm

e25589a2604c8233d5d0dc2278c77de7b5156921 authored over 5 years ago by Mikkel Garcia <[email protected]>
[fix] add_n on ewc

ed260786857a0826fc4c28e1a256ccf9d3c04594 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] dont save if nan in d_fake

291430b618269f41cd037b2eeaf66a5dcdb5df1c authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] AdaBound from https://github.com/taki0112/AdaBound-Tensorflow

3c7e26ce18b2dc06428f4d2c1fdc5d43c31cc9b4 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] sgd example. no gradient penalty

17473f528bbd4120e7d6d37e7edcc5c6ef0a7be0 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] batch_fitness_trainer and fixes for input_fitness_trainer

c4721816152afa736d1273b9ffcfca13ed0c054c authored over 5 years ago by Mikkel Garcia <[email protected]>
[refactor] config.fixed_input

10eb92902c5f2fbe716b11c5d543b7c3afdfe1ef authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] fixed_input standardgan option

548258bbe31ab59b4228dcabc9dabd3e1cc82220 authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] select input by fitness

cfdf068dbc80bf9412b3a9ce456e146dbca11cc5 authored over 5 years ago by Mikkel Garcia <[email protected]>
[experiment] u_to_z instead of reencode_u_to_z on ali manifold_guided

a0b3fce8e786c658aeee13c322b0cdf7fd7f9c9f authored over 5 years ago by Mikkel Garcia <[email protected]>
[feature] different types of minibatch gans

e6b42dbc60c416f578da6aa444e6ae0421db828f authored almost 6 years ago by Mikkel Garcia <[email protected]>
[feature] minibatch train hook - gan type

b4a65e51b6a71bebdb8e1cac7dd54519281b4277 authored almost 6 years ago by Mikkel Garcia <[email protected]>
[fix] minibatch train hook

645d090278a0b2637696f03707bc1a62f4fcda1c authored almost 6 years ago by Mikkel Garcia <[email protected]>
[refactor] gp_sn_train_hook becomes max_gp_train_hook and paper link added.

61618e0acd35312eef2e940529fed7c9827391af authored almost 6 years ago by Mikkel Garcia <[email protected]>
[feature] stylegan-custom1 config

6a6caa6f40d3448a3fc2ce9f3974a39243d39d1e authored almost 6 years ago by Mikkel Garcia <[email protected]>