Skip to content
goard
Go to anything
Open any GitHub user, repo or page
You're not signed in, so some data is limited. Sign in with GitHub for full access.

Intel® Nervana™ reference deep learning framework committed to best performance on all hardware

Python Apache-2.0 8.8 MB Created Apr 15, 2026
Stars 0 Forks 0 Watchers 0 Open issues 0 Open PRs 0
History

7 branches 35 tags

master Add legal discontinuation notice 8c3fb8a
Update requirements.txt ab2a46f
Update requirements.txt f0fc3b5
Merge pull request #466 from rsketine/patch-1 002c20f
Update vis_requirements.txt 9958e08
Update tox_requirements.txt 6ecfaed
Update copyrights from Nervana to Intel (#1135) a10f905
Added new tutorial for forecasting using synthetic timeseries data an… c855af1
Update the old link for 'aeon docs' with the current one on 'master' … 664d7fe
Updated mkl download url (#1130) f43cfa2
Zp/fix mklsumtensor (#1129) a287bf6
v2.6.0latest v2.6.0 release update (#1127) f9d771b
Relaxing tolerance of test_dialated_conv and xfail test_conv_layer (#… e496349
Picking up mklml 20171227 (#1126) 6968659
Replacing funcsigs (#1120) 9e858fb
Zp/ssd inference (#1124) f5b6129
switch to default mkl backend (#1122) 206693d
Adding aeon v1.3.0 dependencies to dockerfile (#1116) ba45cb7
v2.5.0 update for v2.5.0 (#1121) 6577b78
Bump Aeon version to 1.3.0 in Makefile (#1115) 9ce4020
Fix misleading message (#1118) ed263c6
update make clean and .gitignore (#1097) d1478b7
Fix inference performance of MKL batchnorm (#1109) e1d4019
SSD optimization: Dilated Convolution (#1101) d2b3e8f
Added neon dockerfile for IA with mkl and aeon built (#1107) ae4e9d5
Upgrade Sphinx to v1.6.5 to Fix the broken doc search (#1110) 2abcdca
Fixed doc built (#1103) 548f653
Enabling subset_pct for mnist dcgan (#1098) 45c3d66
Igor/batch prediction gpu fix (#1106) 1795a41
add --data_dir arg to pytest (#1085) 6e15f3c
v2.4.0 doc update for neon v2.4.0 release (#1089) 85253da
Clean cifar10 docs 74bceb6
Bump Aeon version to 1.2.0 in Makefile (#1094) 04d2a0f
python3 support (byte encoding) in get_google_word2vec_W (#1075) ac8bfde
convert .md to .rst for pypi (#1087) 597af6c
Add data_loading time measurement to benchmarking (#1079) 9e7dae5
Change resnet model back to original a054579
Update MKL version from 0920 to 1007 (#1076) c583333
fix flake8 E501 long lines af8522e
MKL build for Windows(msys64) 082da5d
Bump Aeon version to 1.1.0 in Makefile (#1073) c5b0b40
- Added requirements in setup.py, this enables python pkgs dependency… 22ac884
Enable pip install (#1074) e2e73f0
Fix flake8 style errors de14f99
Enable GAN networks benchmarking (#1065) 44e6b67
Update Alexnet weight file (fused conv+bias) used in Deep Dream (#1058) a879ae8
Delete an empty file uploaded in error 5ab7904
fix ssd bug (#1063) 455e392
fix issue #1055 (#1061) 5e036ac
xfail test_optimizer.py::test_shift_adamax[gpu] on maxwell and kepler… b35c522
Fix deepdream issue #1056 (#1059) 8c3aba7
Updated examples/deep_dream.py e77ac94
Baojun/fix style error (#1054) 9c7318e
Change misclassification callback (#1052) 55fe94a
v2.3.0 Baojun/neon2.3.0 release doc (#1050) 9eb09d7
Fixing faster-rcnn inference model loading issue (#1051) 49f8fa9
DeepSpeech2 IA Optimizations: v1:merge zp/profile v4 and fix fprop bu… 976f389
Fixed use of reset_cells hyperparameter (#940) 21be9da
Fuse Conv+Bias for all backends (#1029) 64bde1f
fix gan bug for mkl #1024 (#1047) 4d125a9
Zp/fix #issue911 (#1044) 09ad9b8
docker with MKL (#1018) 0c66fae
Updating neon_theme to more prettily render monospace code fonts corr… 0f158d0
rm softmax and fix bug for convert-tensor when input is list (#1016) 890a472
v2.2.0 neon 2.2 release notes (#1010) 5843e71
Zp/roll back conv bias (#1000) c59bf58
add notice that master will be pushed weekly with pre-release checkin… 199f215
Removing VGG temporarily 803ccad
Video3cD minor fixes for python 3 (inference and demo scripts) (#998) 697bb7b
Configurable batch size in dummy dataset (for benchmark scripts) (#997) c23c044
kbinias/vgg small improvements (#996) e178d7e
CHWD Support for Batch Normalization in mkl (#963) 3135ef9
Adding VGG that works for aeon v1.0.0 and neon 2.2. #808. #888. #925… 520a50e
Fix py3 and cache dir issues with SSD model (#994) 362d649
Update data.py (#995) cf21770
Peng Zhang/ Rong Zhang: modify neon ResNet-50 last layer to match Caf… 6d6a0df
Video c3d inference fix (#991) 042dc05
Bump Aeon version to v1.0.0 (#989) 231deb4
Enabling subset_pct for SSD (#992) 75164d9
Adding the missing arguments in function (#984) 18e3bd7
Removing xfail thanks to the new MKL. (#985) 52f5a14
Benchmark tool update - support Tree-like models (SSD) (#986) 7e1f0f8
information about new aeon manifest added to changelog (#983) 9985dc0
Benchmark Seq2Seq (#967) 44a3cb4
Remove deprecated loaders (#980) 4d0e696
Zp/deepspeech (#970) 065f05c
Add SSD (#982) 175bd3c
Aeon 1.0.0 in Neon (#957) e64a2a0
Zp/update mkl version to 20170908 (#976) 6bca818
updated numpy version to 1.13.1 to resolve issue #948 (#956) 38d9dcf
update tools.rst (#931) 912846e
add conv_bias layer (#888) b187b0b
fix skip node and test_mergesum_layer for issue #916 (#955) e1daa91
Fix shared outputs allocation for LookupTable (#969) 4d7aa62
Tyler/seq2seq testing (#849) 03e87c1
Serialize gmean and gvar attributes in BiBNRNN layers (#933) 312f2e1
v2.1.0 Update for v2.1.0 (#917) 5504320
Adding `numactl --interleave=all` Performance Best Practice Notes (#919) 268143b
add mkl backend for neon tests (#881) fea87bc
Adding GAN examples' prerequisites into README.md in example/gan fold… 3dd1692