Skip to content

Latest commit

 

History

History
59 lines (47 loc) · 2.36 KB

FUNIT.md

File metadata and controls

59 lines (47 loc) · 2.36 KB
Raw

Train

1. Modify the configuration file ("cfgs/FUNIT/train.yaml")

You can set these values by giving command-line arguments like argparse, not modifying this configuration file directly. For the detailed description, please refer here.

  • use_ddp: Whether to use DataDistributedParallel. Set True to this to use multi-gpu.

  • port: The port for the DataDistributedParallel training.

  • trainer: (leave blank)

    • resume: Path to the checkpoint to resume from.
    • work_dir: Path to save the checkpoints, the validation images, and log.

  • dset: (leave blank)

    • train: (leave blank)
      • source_path : path to the source font or source directory to use for the validation.
      • source_ext: extension of the source data.
        • If you are using a ttf file, set this to "ttf".
        • If you are using image files, set this to their extension ("png", "jpg" ...).

2. Run training

python train_FUNIT.py cfgs/FUNIT/train.yaml cfgs/data/train/custom.yaml --work_dir(optional) path/to/save/outputs

-g, -n, -nr, -p are arguments for the DistributedDataParallel training. You do not need to give these arguments if you are using a single GPU.

  • arguments
    • path/to/config (first argument): path to configration file.
      • Multiple values are allowed but the first one should locate in cfgs/FUNIT.
    • -g : number of gpus to use for the training.
    • -n : number of nodes to use for the training.
    • -nr : the ranking of current node within the nodes.
    • -p : the port to use for the DistributedDataParallel training.
    • --work_dir : path to save outputs. The trainer.work_dir in the configuration file will be overwrited to this value.

Evaluate

Run evaluation

python inference.py cfgs/FUNIT/default.yaml cfgs/data/eval/chn_ttf.yaml \
--model FUNIT \
--weight weights/FUNIT_chn.pth \
--result_dir ./result/FUNIT
  • arguments
    • path/to/config (first argument, multiple values are allowed): path to configration file.
      • Multiple values are allowed but the first one should locate in cfgs/FUNIT.
    • --model : The model to evaluate. DM, LF, MX and FUNIT are available.
    • --weight: The weight to evaluate.
    • --result_dir: Path to save generated images.
    • --n_ref: The number of reference characters to use for the generation.