sassoftware / sassoftware/python-sasctl

Use PZMM to Save Tensorflow Model

Open
#108 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
52
Forks
45
Avg merge
23h 38m
Merged PRs (30d)
2

Description

I was working with a customer who was trying to use the pzmm.PickleModel.pickleTrainedModel() function on a TensorFlow model, but they were getting a 'TypeError: can't pickle weakref objects'. My ideal solution would be to for the PickleTrainedModel() function to have the ability to ingest Tensorflow models or have a similar function available for Tensorflow models. For now, I think the work-around would be to use Tensorflow's save() function to save the model, make that model available to Model Manager, and then edit the score code to use Tensorflow's load_model() function instead of un-pickling the model, but I am open to other work-around ideas.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the TypeError from pzmm.PickleModel.pickleTrainedModel() with a TensorFlow model, then compare the requested behavior with TensorFlow's save() and load_model() workflow. Done should mean an agreed, tested way for Model Manager to handle TensorFlow models, or a documented alternative if direct pickling is not supported.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, tensorflow
Domain
machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.