uds-lsv/bert-stable-fine-tuning

On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines

43
/ 100
Emerging

Provides empirical analysis of fine-tuning instability across BERT, RoBERTa, and ALBERT on GLUE tasks, identifying vanishing gradients and generalization variance as root causes rather than catastrophic forgetting. Implements stable fine-tuning baselines built on Huggingface Transformers (v2.5.1) with reproducible Docker-based evaluation scripts. The codebase includes diagnostic tools for analyzing optimization dynamics and gradient behavior during training.

138 stars. No commits in the last 6 months.

Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 17 / 25

How are scores calculated?

Stars

138

Forks

21

Language

Python

License

Apache-2.0

Last pushed

Sep 06, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/nlp/uds-lsv/bert-stable-fine-tuning"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.