This model consists of an SMP-UNet that has been finetuned from the universal December 2023 model. The finetuning was for 11 epochs on a set of 20 gel images containing sharp gel bands (Stella dataset). The model sacrifices a tiny amount of general performance for superior results with clean, very sharp gel bands.
For more details on the configuration used for training, please visit https://huggingface.co/mattaq/GelGenie-Universal-FineTune-May-2024 and check the config.toml file. Our codebase is fully open-sourced and is available here: https://github.com/mattaq31/GelGenie.
Inference API (serverless) does not yet support pytorch models for this pipeline type.