search

NervanaSystems / neon

Intel® Nervana™ reference deep learning framework committed to best performance on all hardware
http://neon.nervanasys.com/docs/latest
Apache License 2.0
3.87k stars 811 forks source link

Update tox_requirements.txt #466

Closed rsketine closed 5 years ago

rsketine commented 5 years ago

https://pyyaml.org/download/pyyaml/ (SDL recommendation is to upgrade to 4.2 or higher but here the next version published is 5.1 so upgrading to that.)

raramer01 commented 5 years ago

Changes have been validated:

ran a modified version of the NEON_Merge_Request.sh script:

python 3.5 unit test results: ======================================================= 851 passed, 20 skipped, 654 deselected, 24 xfailed in 188.74 seconds =======================================================

python 2.7 unit test results: ======================================================= 851 passed, 20 skipped, 654 deselected, 24 xfailed in 175.63 seconds =======================================================

make all RC = 0 make style rc = 0

also ran mnist mlp example: (.venv2) [rramer@nervana-skx110 neon]$ neon examples/mnist_mlp.yaml 2019-04-17 10:37:16,462 - neon - DISPLAY - Downloading file: ./mnist.pkl.gz Download Progress |██████████████████████████████████████████████████| 2019-04-17 10:37:27,879 - neon - DISPLAY - Download Complete Epoch 0 [Train |████████████████████| 469/469 batches, 0.75 cost, 1.20s] Epoch 1 [Train |████████████████████| 469/469 batches, 1.33 cost, 1.59s] Epoch 2 [Train |████████████████████| 469/469 batches, 0.57 cost, 2.26s] Epoch 3 [Train |████████████████████| 468/468 batches, 0.14 cost, 2.18s] Epoch 4 [Train |████████████████████| 468/468 batches, 0.13 cost, 1.59s] Epoch 5 [Train |████████████████████| 468/468 batches, 0.12 cost, 1.87s] Epoch 6 [Train |████████████████████| 468/468 batches, 0.11 cost, 1.87s] Epoch 7 [Train |████████████████████| 468/468 batches, 0.10 cost, 1.28s] Epoch 8 [Train |████████████████████| 468/468 batches, 0.09 cost, 1.28s] Epoch 9 [Train |████████████████████| 468/468 batches, 0.08 cost, 1.28s]

Please merge...