Göran Sandström

My feedback

  1. 1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Product Feedback » Other  ·  Flag idea as inappropriate…  ·  Admin →
    Göran Sandström shared this idea  · 
  2. 1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Product Feedback » Cluster management  ·  Flag idea as inappropriate…  ·  Admin →
    Göran Sandström shared this idea  · 
  3. 3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    Göran Sandström commented  · 

    update, there was some new dependency updated that broke IPython today, fix is:

    conda install -y -c anaconda wrapt=1.11.1
    conda install -y -c anaconda numpy=1.16.4
    add this:
    conda install -y -c anaconda ipython=7.6.0

    This will cause an update to python 3.7, don't know why but some wrapt dep forced this upgrade, and it didn't cause an automatic update to ipython so notebooks became unresponsive...

    Göran Sandström commented  · 

    I solved it!

    To the script you need to add two lines like so:

    conda install -y -c anaconda cudatoolkit=10.0
    ### BELOW THIS, ADD THESE TWO, remove this line

    conda install -y -c anaconda wrapt=1.11.1
    conda install -y -c anaconda numpy=1.16.4

    ## ABOVE THIS, again, don't include this comment

    DEBIAN_FRONTEND=noninteractive sudo apt-get install -yq --no-install-recommends \
    libcudnn7=7.5.0.56-1+cuda10.0 \
    libcudnn7-dev=7.5.0.56-1+cuda10.0

    explanation is that numpy installation got weird, an also it couldn't install wrapt properly through pip install (as a part of tensorflow install) because its installations were managed by conda

    Göran Sandström shared this idea  · 
  4. 3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Product Feedback » Cluster management  ·  Flag idea as inappropriate…  ·  Admin →
    Göran Sandström shared this idea  · 
  5. 87 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    3 comments  ·  Product Feedback » Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
    Göran Sandström supported this idea  · 
  6. 29 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Product Feedback » Notebooks  ·  Flag idea as inappropriate…  ·  Admin →
    Göran Sandström supported this idea  · 
  7. 202 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    13 comments  ·  Product Feedback  ·  Flag idea as inappropriate…  ·  Admin →
    Göran Sandström supported this idea  · 
  8. 1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    Göran Sandström commented  · 

    You could use TensorFlow probability, sadly, only TF 1.12.0 is available so you're stuck at TFP 0.5.0

  9. 10 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    Göran Sandström supported this idea  · 

Feedback and Knowledge Base