English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep Neural Networks (DNNs) are being adopted as components in softwaresystems. Creating and specializing DNNs from scratch has grown increasinglydifficult as state-of-the-art architectures grow more complex. Following thepath of traditional software engineering, machine learning engineers have begunto reuse large-scale pre-trained models (PTMs) and fine-tune these models fordownstream tasks. Prior works have studied reuse practices for traditionalsoftware packages to guide software engineers towards better packagemaintenance and dependency management. We lack a similar foundation ofknowledge to guide behaviors in pre-trained model ecosystems.  In this work, we present the first empirical investigation of PTM reuse. Weinterviewed 12 practitioners from the most popular PTM ecosystem, Hugging Face,to learn the practices and challenges of PTM reuse. From this data, we modelthe decision-making process for PTM reuse. Based on the identified practices,we describe useful attributes for model reuse, including provenance,reproducibility, and portability. Three challenges for PTM reuse are missingattributes, discrepancies between claimed and actual performance, and modelrisks. We substantiate these identified challenges with systematic measurementsin the Hugging Face ecosystem. Our work informs future directions on optimizingdeep learning ecosystems by automated measuring useful attributes and potentialattacks, and envision future research on infrastructure and standardization formodel registries.

An Empirical Study of Pre-Trained Model Reuse in the Hugging Face Deep Learning Model Registry