Ecosyste.ms: OpenCollective

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

github.com/deeplearning4j/deeplearning4j

Suite of tools for deploying and training deep learning models using the JVM. Highlights include model import for keras, tensorflow, and onnx/pytorch, a modular and tiny c++ library for running math code and a java based math library on top of the core c++ library. Also includes samediff: a pytorch/tensorflow like library for running deep learn...
https://github.com/deeplearning4j/deeplearning4j

applyPairwiseLambda now can handle scalar case for its' second argument. It's tested in the tanh_bp scalar case.

Note: if your first argument is scalar but the second argument and output argument have the same ...

72725e7859a67fde5c36b83c1d02b64411a2d0d3 authored about 3 years ago by AbdelRauf <[email protected]>
Merge pull request #9516 from eclipse/ag_test_fixes_2

Redoes https://github.com/eclipse/deeplearning4j/pull/9514

9bad8d579528988d086ce2072ebd3900c596b1d2 authored about 3 years ago by Adam Gibson <[email protected]>
Update Conditions.java

b91dfee7f08ecc649d7887b271c6b8ec6d1f2362 authored about 3 years ago by Adam Gibson <[email protected]>
Update TestOptimizers.java

db6d7b91c2142a1d4dd4089a5eba879f947325f9 authored about 3 years ago by Adam Gibson <[email protected]>
Update TestOptimizers.java

8b9dd25105d36f3a1f6e437afd0764b5739f440a authored about 3 years ago by Adam Gibson <[email protected]>
More conditions feedback

3a3f3f90f62359faca3254bfc0a48b2664afcf0d authored about 3 years ago by Adam Gibson <[email protected]>
Remove extra files, change ordinal to index

96ae054bc064a82ca25b8ea7ab83b8ea020b000d authored about 3 years ago by Adam Gibson <[email protected]>
Propagate more c++ fixes

29d5a8a3e07d2610c9d5d253f7a3c4ae76da7ec4 authored about 3 years ago by Adam Gibson <[email protected]>
Redoes

d400122bc3e3a53a338f516e42575a30eaf9adcd authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9513 from quickwritereader/master2

Refactor + Format:

2f8487eaf3e1e5e679b6bd1787ad5946ea9eaf1f authored about 3 years ago by Adam Gibson <[email protected]>
Refactor + Format:

- introduce system/common.h
- prefix with sd_ and SD_
- status is enum
- changes Nd4jLong and sis...

d77597bae3253e13a3384a168fda789f68309c1a authored about 3 years ago by AbdelRauf <[email protected]>
Merge pull request #9512 from eclipse/ag_compilation_fixes

Fix compilation errors cleanup from PR merges

bc7bfd8792529972599ae78b3db08fb976509265 authored about 3 years ago by Adam Gibson <[email protected]>
Incorporate resize change from rauf

854f56dbc9d63fd469cc0d816c2bd59f5ef290bd authored about 3 years ago by Adam Gibson <[email protected]>
Fix compilation errors cleanup from PR merges

05928c030b375132fcf67465360e1a15ba477dc3 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9507 from eclipse/ag_scalar_reduce_fix

Ensure gradient check passes with scalars

6b65ad84abe953492d72c1767ff6267fdf9817c9 authored about 3 years ago by Adam Gibson <[email protected]>
Restore eps and extra println

834d726246119ca7e553ca16f48c5cbde0b5160b authored about 3 years ago by Adam Gibson <[email protected]>
Remove out of date constructors and tests

273a5bc706b4712d3a74165151760d44e55b90d6 authored about 3 years ago by Adam Gibson <[email protected]>
Change norm max thresholds, add gradient checks for matrix multiply

1874bc062d4caef9d583069b86b61e6de45dd30b authored about 3 years ago by Adam Gibson <[email protected]>
Add matrix multiply gradient check for matrix case, add matrix case that handles proper reductions

a08757b0de70d29aa039d9990c818bf213bdfaf4 authored about 3 years ago by Adam Gibson <[email protected]>
Add new matmul bp fix

735d0498a05f1d77da19ffc29df4818e5ec6f3ec authored about 3 years ago by Adam Gibson <[email protected]>
Merge branch 'ag_scalar_reduce_fix' of https://github.com/eclipse/deeplearning4j into ag_scalar_reduce_fix

86bb7c0be0b4416991191f5c599a528a2daa8af1 authored about 3 years ago by Adam Gibson <[email protected]>
Revert mmul shape changes

a2c516a4029280b6500c7c93a695a252f2618b06 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9511 from quickwritereader/test_fixes

Test fixes

8003253c4c30e71dace3c5f4208fde7e20f0c724 authored about 3 years ago by Adam Gibson <[email protected]>
fixes test case failures introduced by typos

Signed-off-by: AbdelRauf <[email protected]>

90869e4303c511f6986d8c646ea139bea6391b1f authored about 3 years ago by AbdelRauf <[email protected]>
Merge pull request #2 from eclipse/ag_scalar_reduce_fix

9b5593d7719fd96b103b9c1e8381ef5e32493840 authored about 3 years ago by Abdelrauf <[email protected]>
Test updates

08b72965c15b9d12260f8b10337584fbc96da188 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9495 from quickwritereader/master

ImageResize improvements

d54c13dcf504e142c3fea2f71fcfe26a25abf608 authored about 3 years ago by Adam Gibson <[email protected]>
Remove debug statements

81ab3aadd78387f0a3b79672f283f5b1aa84f9fd authored about 3 years ago by Adam Gibson <[email protected]>
Ensure gradient check passes with scalars

d5ea7faae5600e07ac96f77a4163f34e038a133b authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9502 from eclipse/ag_onnx_pooling_generalization

Add onnx changes for global pooling to be more generalized

68fddcdd112f9d7cb6e97791dd4d037de16f6f2d authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9503 from eclipse/ag_splitv_codegen_updates

Codegen updates for splitv allowing variable inputs

ccefd04f845a2fb7801b08b0eaf0e341b30c02dd authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9501 from eclipse/ag_java_import_changes

Java side changes for model import and samediff uptraining

e7206a69f07dd3ac3c781bfc7018bf06603869eb authored about 3 years ago by Adam Gibson <[email protected]>
Merge branch 'master' into ag_java_import_changes

e6ad9ed58c44e9aa76ded86d931cb4c257f7ed6d authored about 3 years ago by Adam Gibson <[email protected]>
Fix up reduce operations/dimensions test failures , address github comments

2b6686713c8069ffa07e8f449d523d6d63d92c32 authored about 3 years ago by Adam Gibson <[email protected]>
Liberalize std/variance gradients

c7aa30e06faa0ac9d1404bb746b9d6e02e254dd6 authored about 3 years ago by Adam Gibson <[email protected]>
Add reduce bp constructors for dimensional variance, add bias corrected to true for variance gradient checks

dbac595a797a16fafd04b7029b03f1ee1d90fa19 authored about 3 years ago by Adam Gibson <[email protected]>
Add proper variables for dimensions similar to reduce forward

d6130d5b5f17780576cc6eb9d7611737fed4402f authored about 3 years ago by Adam Gibson <[email protected]>
Minor logsumexp fix for setting keepDims via reflection

0cb8f74c0f473a88d0ebdf140b755fdd06f133ac authored about 3 years ago by Adam Gibson <[email protected]>
Put pre processing in to single method

0c0162d5cfde4bf0eb268e6a553d6958aead9e42 authored about 3 years ago by Adam Gibson <[email protected]>
Update based on feedback

7f6e464622fedb6c8a49181172dbf347a5017bbb authored about 3 years ago by Adam Gibson <[email protected]>
C++ windows compilation fixes, onnx import updates, reshape bug fixes, training fixes, gcc mingw path fixes

cdc35ac173a3d0fbf27c9a0eeffbb46938eaad04 authored about 3 years ago by Adam Gibson <[email protected]>
Merge branch 'master' into master

1fba2dd76d5d25e55bb8f7e36b843cd15fdd7ad1 authored about 3 years ago by Abdelrauf <[email protected]>
- ImageResize methods can accept strided input NdArray

- Use instance fields instead of ordinals for ImageResizeMethod
- v1 version of the bicubic Metho...

e5d0e115e6ca7a840add73bb1e2f9af688bef6b0 authored about 3 years ago by AbdelRauf <[email protected]>
Update pooling, convolution to have proper error messages and consolidated configuration

ba2dcdfebbf88dd798aa58a5cb8688b33ace2e1c authored about 3 years ago by Adam Gibson <[email protected]>
Codegen updates

ece740bf6cb1841df4686a9edb0f99e058e2f646 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9500 from eclipse/ag_reduce_signature_update

Update reduction signatures to allow optional shapes

475c73df5041024b72694347eeea1ed31736273d authored about 3 years ago by Adam Gibson <[email protected]>
Add onnx changes for global pooling to be more generalized

64bdd91076fbe74a37cec192600f6fee372fbdab authored about 3 years ago by Adam Gibson <[email protected]>
Update BaseLoss.java

b84bdc4e1db73bcb2471481505258541a9b8d72d authored about 3 years ago by Adam Gibson <[email protected]>
Update ImportClassMapping.java

cc802f7dad353e777fbe098d637d98f69bb15e0d authored about 3 years ago by Adam Gibson <[email protected]>
Java side changes for model import and samediff serialization

0d44d120d910ff924ee7d3cc51c0f010223b1b73 authored about 3 years ago by Adam Gibson <[email protected]>
Add updated signatures for reduce operations reflecting the ability to allow input shapes as optional extra inputs

0462411b08d4ffa66e3e402f708970aa6f180622 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9494 from eclipse/ag_lint_fix

Fix maven lint issues

3a89be17fb0e49ca75ad07d2ced97dc53bf80b4c authored about 3 years ago by Adam Gibson <[email protected]>
Fix aurora op executioner imports

b36ecbfd3a0cb01c8bafa5873c5231e42efc6b8e authored about 3 years ago by Adam Gibson <[email protected]>
Fix maven lint issues

7104521324c4e38d5cc962b955721f436c254d89 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9493 from eclipse/ag_fix_9481

Fixes https://github.com/eclipse/deeplearning4j/issues/9481

d665d5e8f7ec91515aaa8ecf3ff0f2516a7665f2 authored about 3 years ago by Adam Gibson <[email protected]>
Add new platform dependency

b674ba6021b13b76b0e4ba6331d0021631c133bc authored about 3 years ago by Adam Gibson <[email protected]>
Fixes https://github.com/eclipse/deeplearning4j/issues/9481

fb61e781309b9cd6a2308ed3e9de0e34d2b43631 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9492 from eclipse/ag_cast_fix

Fix misc issues with import (sub graph serialization, onnx definitions,..)

80876a1461aa009a585aea1684b7c518c6243c60 authored about 3 years ago by Adam Gibson <[email protected]>
Update TestOnnxProcessLoader.kt

dfdef2d63654f7f0d99439d078237f634830ff54 authored about 3 years ago by Adam Gibson <[email protected]>
Fix conversions, sub graph serialization

dc3394b68d980353b0ef2e33dc8ba8e1acaa4518 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9491 from eclipse/ag_ci_fixes

CI fixes

784d9f5ba1f1b4d97b4df927f4a441d14dbec129 authored about 3 years ago by Adam Gibson <[email protected]>
More removal of scopes

0f350bf227bab35cde52292b2a08ccc1fc3c5006 authored about 3 years ago by Adam Gibson <[email protected]>
Add native preset as default for mac

1a9e983d29070b6c074eb663855960be4c303909 authored about 3 years ago by Adam Gibson <[email protected]>
Remove scope declaration on implementations

1f0f6db4ab35027b0ec8608a551a8aac7f094311 authored about 3 years ago by Adam Gibson <[email protected]>
Add more explicit variable names

1e0c5f95f8d8dd801259dc72ad36a51fac9c8cd2 authored about 3 years ago by Adam Gibson <[email protected]>
Update activations.h

8b59f11e607bb3c8034714c21eed741ce181d292 authored about 3 years ago by Adam Gibson <[email protected]>
Fix scope declaration

0e162d3c7304f4f6fcd1d70d4ae2591c865132a6 authored about 3 years ago by Adam Gibson <[email protected]>
CI fixes

6ebe5ff9c0c15c6c3f785cb3c95fb6902fe4c421 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9489 from eclipse/ag_onnx_if

Add more onnx ops for import (if, etc)

a667b938a82741f15d41b042b7fd665b69e7df51 authored about 3 years ago by Adam Gibson <[email protected]>
Minor import fixes

61285168bb730238011a93697d55f71e172e00d3 authored about 3 years ago by Adam Gibson <[email protected]>
Update types, add new set shape op for variable set shape, convert transpose to pre processor

b9c050bce6d25af5f05fa72a44409e21edb6e7a9 authored about 3 years ago by Adam Gibson <[email protected]>
Update Cast.kt

72a9a0c9cfbc3b4500b44fa1503f5a67465c8ea5 authored about 3 years ago by Adam Gibson <[email protected]>
Update OnnxOpDeclarations.kt

d5372298faa0393cbf83ab8a47ce05ce8daa9053 authored about 3 years ago by Adam Gibson <[email protected]>
Update comments

1b439b00d30ec60696a2f383887889aa8f420d73 authored about 3 years ago by Adam Gibson <[email protected]>
remove temp files

affbdf26b6214ef553e26e6e854f0a47c71d0cfd authored about 3 years ago by Adam Gibson <[email protected]>
Add ROIAlign

f0fb6a3cf72e2e91214841be40dc836b705758ff authored about 3 years ago by Adam Gibson <[email protected]>
Add support for if statement

8a70275e19e7067d06f8b0013d2fd2c7ed0344fa authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9488 from eclipse/treo/fix_sbt_eca

Fix samediff-import-api pom file

db5566c88111f6e1bcd38a4b0eb8c0b6b65476ab authored about 3 years ago by Adam Gibson <[email protected]>
Fix samediff-import-api pom file

This removes the space character in the very first position of the pom.xml file.

With it pre...

358227ec1729728332bffcbd3c90b81927508db7 authored about 3 years ago by Paul Dubs <[email protected]>
Merge pull request #9484 from CodecLondon/bert-preprocess-more-punctuations

BertWordPieceTokenizer supports more punctuation marks

2036438b1a83e25515fda53635c3519a310c9be8 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9483 from eclipse/ag_libnd4j_gen_bash

Add main class and script for running libnd4j gen

0e89b3de58ed9f9702c73f3ce6caf689d747a8a5 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9478 from eclipse/ag_fix_9476

Add proper min,max op, update op specs

77ad9664a0771b032bd28da1aa15244dce171639 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9480 from eclipse/ag_spark_build_fix

Fix spark script exec

7c4ebf9704f80a643d0ecabb3c2126b9e9a00337 authored about 3 years ago by Adam Gibson <[email protected]>
Update

4c66cdd48f382c60afb6586efe8bee34c844386e authored about 3 years ago by Adam Gibson <[email protected]>
Support more punctuations not in unicode punctuation class

Adapted from https://github.com/huggingface/transformers/blob/v4.11.3/src/transformers/tokenizat...

f70f06619f7a628b90bad70ae5c36a4e18bc6eb8 authored about 3 years ago by Ahmed Mahran <[email protected]>
Support unicode punctuations for BertWordPieceTokenizer

6b1c6a33ab43a8f4cd6a9d060bc507eee40a9911 authored about 3 years ago by Ahmed Mahran <[email protected]>
Add main class and script for running libnd4j gen

2a125a0cdb89b1f6e4513bfb4768e943efdb9e17 authored about 3 years ago by Adam Gibson <[email protected]>
Merge pull request #9482 from quickwritereader/master

fix is_callable for cuda as well. replaced with working code

6652db887f38039513a5c7d44b3ced9a55a6c075 authored about 3 years ago by Adam Gibson <[email protected]>
Remove spark nlp java 8

dcb539fbec93773c81f0f501b9e19f5c7ea21c1b authored about 3 years ago by Adam Gibson <[email protected]>
Remove java 7 nlp module

c98595654c5cd0c30dea572554fa432335c93316 authored about 3 years ago by Adam Gibson <[email protected]>
Update pom.xml

c540b1249a24a517862a3035a94afe5ff7d21997 authored about 3 years ago by Adam Gibson <[email protected]>
Add

30cdae74c0846926e94c98e62064bee2170cfcac authored about 3 years ago by Adam Gibson <[email protected]>
Update clip by norm, value and associated gradients to accept ndarray inputs

2e510b22b70c48a2a9b1ae41acb64e55ff696628 authored about 3 years ago by Adam Gibson <[email protected]>
Allow clip by value to accept data types

a2d295638197a3adcb5353d94f9dc6c5fbca6429 authored about 3 years ago by Adam Gibson <[email protected]>
fix is_callable for cuda as well. replaced with working code from stackoverflow

Signed-off-by: AbdelRauf <[email protected]>

08c97990843e26b416044f2b340162cab773255f authored about 3 years ago by AbdelRauf <[email protected]>
Update all spark versions to 2.12

d8d6a03ac2fc7baadc4e320866a77ae82d91888a authored about 3 years ago by Adam Gibson <[email protected]>
Update build-deploy-cross-platform.yml

837fadeb72861149fa01707d00e85a80741e2ac9 authored about 3 years ago by Adam Gibson <[email protected]>
Fix spark executable build

e931fa5e7dcc38977bcca79cf4bcc721ff117976 authored about 3 years ago by Adam Gibson <[email protected]>
update current usages of min to previous assumptions for backwards compatibility

1799449d63010bc77bf491834598878186cb41f6 authored about 3 years ago by Adam Gibson <[email protected]>
Update Clip.kt

abd44ec5ad1db796afdc5d545f66dd8e72300c23 authored about 3 years ago by Adam Gibson <[email protected]>