N2V comparison

[jdeschamps]

• For each method:
  • uv's pyproject.toml
  • Python script to run the benchmark
  • HPC script to run the python benchmark
• n2v:
  • BSD68
  • SEM
• ppn2v
  • BSD68
  • SEM
• CAREamics
  • BSD68
  • SEM

[mese79]

ppn2v needs to have train and validation images of the same size.
For the SEM dataset, it's not true. So, I resized the validation images.