Pre-trained language models (PTLMs) have revolutionized the field of natural language processing (NLP), enabling significant advancements in tasks such as text generation and translation. Similar to software package management, which involves centralized version control and distributed consumption, PTLMs are trained using code and environment scripts hosted in an upstream repository (e.g., a GitHub (GH) repository), while the family of model variants trained from a given repository’s scripts are distributed using dedicated downstream distribution platforms like Hugging Face (HF). Despite these similarities, coordinating development activities between GH and HF presents several challenges to avoid misaligned release timelines, inconsistent versioning practices, and other obstacles to seamless reuse of PTLM variants. To understand how commit activities are coordinated between these two platforms, we conducted an in-depth mixed-method study of 325 PTLM families consisting of 904 HF PTLM variants. Our analysis reveals that GH contributors typically make changes related to specifying the version of the model, improving code quality, performance optimization, and dependency management within the training scripts, while HF contributors make changes related to improving model descriptions, data set handling, and setup required for model inference. Furthermore, to understand the synchronization aspects of commit activities between GH and HF, we examined three dimensions of these activities—lag (delay), type of synchronization, and intensity—which together yielded eight distinct synchronization patterns. The prevalence of partially synchronized patterns, such as \emph{Disperse synchronization} and \emph{Sparse synchronization}, reveals structural disconnects in current cross-platform release practices. These patterns often result in isolated changes—where improvements or fixes made on one platform are never replicated on the other—and in some cases, indicate an abandonment of one repository in favor of the other. Such fragmentation risks exposing end users to incomplete, outdated, or behaviorally inconsistent models. Hence, recognizing these synchronization patterns is critical for improving oversight and traceability in PTLM release workflows.