tensorflow / tensorflow/metadata

Add a default value field for schema.feature

Open
#17 0 comments 2 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
110
Forks
62
PR merge metrics
No merged PRs in 30d

Description

We use TFMD schemas to construct feature_specs to parse TFRecords. Some records may have missing entries for certain features, and in such cases we may want to allow a default value. This functionality is already supported in tf.io.FixedLenFeature. It would be nice to add some fields to the Feature message to mirror this:

message Feature {
  ...
  oneof default_value_type {
    int64 int_default = ...;
    ...
  }
}

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 locating the Feature message and the code that constructs feature_specs from TFMD schemas, then compare its supported fields with tf.io.FixedLenFeature. Done means the schema can represent default values for missing features and those values are available when constructing the resulting feature_spec.

Written by the indexing model from the issue text.

Assessment

Tech stack
tensorflow
Domain
data, machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.