facebookresearch / facebookresearch/ResponsibleNLP

Toxigen Metric not returning anything

Open
#44 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
212
Forks
35
PR merge metrics
No merged PRs in 30d

Description

Hi! I'm trying to analyse some of my model's generations using the regard promptset. When I try to apply the regard metric everything seems to work fine, but when I try to use the toxigen metric it doesn't return anything. Here's the command I'm using:

robbie/scripts/run_eval.sh --dataset regard --model ../my_model -s toxigen --result-dir results/ --dataset-dir robbie/datasets/regard/
And this is the output:
MetricResult(scores=[], stats={})
Any help would be greatly appreciated. Thanks!

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 with robbie/scripts/run_eval.sh and trace how the toxigen metric is selected and executed, comparing that path with the working regard metric. Reproduce the command from the issue and inspect the metric result; done means the toxigen evaluation returns populated scores and stats instead of MetricResult(scores=[], stats={}).

Written by the indexing model from the issue text.

Assessment

Tech stack
python, shell
Domain
machine-learning
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 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.