google-research / google-research/tabfm

Regression bug?

Open
#43 1 comment 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
2.6k
Forks
270
Avg merge
1d 7h
Merged PRs (30d)
1

Description

I tried the regression option with 4000+ rows but it returns multiple Y (predicted) columns and the output fails here:

/content/tabfm/tabfm/src/classifier_and_regressor.py in _predict_internal(self, X)
3285
3286 output = self._batch_forward(Xs_all, ys_all, cat_masks_all, ds=ds_all)
-> 3287 predictions = output.squeeze(-1)
3288 return predictions
3289

Because it expects a single Y output.

I ran a debug version to look at the returned data structure and this is how I discovered the multiple Y output issue.

Contributor guide

Open the contributing guide

Research direction

Reproduce the regression case with 4000+ rows, then inspect classifier_and_regressor.py at _predict_internal and the _batch_forward output it receives. Confirm the expected prediction shape for regression and that the failing case completes without the multiple-Y output error.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
55/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.