Globally, lakes are losing ice cover, but our understanding of the rates and variability of change are biased towards few lakes with long-term records. Recent works have used remote sensing and predictive modeling to supplement the observational record, but these approaches produce large errors when estimating ice phenology for individual lakes. Our associated manuscript explores machine learning (ML) approaches for hindcasting ice phenology using daily weather drivers and lake ice records from 1980-2018 across 625 Minnesota lakes, covering 4359 lake-years of record. Most notably, this model application provides LSTM predictions of lake ice cover time series for 881 National Hydrography Dataset High Resolution (NHDHR) lakes in Minnesota from 1980 to 2018. Along with the best available predictions, we provide the trained model weights and feature scaling factors associated with the best model developed in PyTorch, an open-source deep learning library. This model application uses inputs available from a prior data release (https://doi.org/10.5066/P9PPHJE2) which are retrieved via code and provides an earlier version of Minnesota lake ice phenology data published at https://doi.org/10.13020/110f-j487, whose original source is the Minnesota Department of Natural Resources State Climatology Office. For more in-depth information on the methodology and a detailed exploration of the model(s), please refer to Diaz et al. 2025 (in review). This work was initially funded by the Integrated Information and Dissemination Division, then by the Predictive Understanding of Multiscale Processes project - both under the USGS Water Mission Area (WMA) to explore the potential of machine learning-based lake ice prediction in addition to attention-based transformers, very large models, and explainable AI (XAI). This project made use of the USGS Tallgrass (https://doi.org/10.5066/P9XE7ROJ) High Performance Computing environment for GPU acceleration of large neural network models. The methods and results from this modeling effort are described in: Jeremy Diaz, Samantha Oliver, Simon Topp, et al. Predicting Minnesota lake ice phenology with deep learning, explainable methods, and a physically based benchmark, 1980-2018. ESS Open Archive . September 05, 2025. DOI: 10.22541/essoar.175710698.87505589/v1. This is a USGS peer reviewed and approved preprint. This manuscript is also under review at Water Resources Research.