mirror of https://github.com/facefusion/facefusion-labs.git synced 2026-04-19 15:56:37 +02:00

Files

T

Henry Ruhs 2e6394565a Next (#93 )

* add sync batchnorm

* replace random.choice with hash

* fifty percent reduction

* fix discriminator input

* restore dataset.py

* Remove duplicates

* add discriminator_ratio to config

* fix onnx export bug: replace round() with int()

* Fix embedding naming

* Introduce ModelWithConfigCheckpoint callback (#86)

* Fix dist ini

* Style: Refactor typing and improve code clarity in training.py (#88)

* Add type casting for trainer params

* Add type casting for trainer params

* Add type casting for trainer params

* Remove inplace activations for torch.compile compatibility (#89)

* Fix README

* improvise with norm layers & weighted average

* add skip layer

* use gelu instead of leaky_relu

* cleanup

* cleanup

* Update dependencies

* Different defaults and enable validation

* Different defaults and enable validation

* Revert to higher batch size

* Just use copy over copy2

---------

Co-authored-by: harisreedhar <h4harisreedhar.s.s@gmail.com>
Co-authored-by: NeuroDonu <112660822+NeuroDonu@users.noreply.github.com>
Co-authored-by: Harisreedhar <46858047+harisreedhar@users.noreply.github.com>

2025-09-06 19:12:29 +02:00

1.4 KiB

Raw Permalink Blame History

CrossFace

Seamless face embedding across various models.

Preview

Installation

pip install -r requirements.txt

Setup

This config.ini utilizes the MegaFace dataset to train the CrossFace model for SimSwap.

[training.dataset]
file_pattern = .datasets/megaface/**/*.jpg

[training.loader]
batch_size = 128
num_workers = 8
split_ratio = 0.95

[training.model]
source_path = .models/arcface_w600k_r50.pt
target_path = .models/arcface_simswap.pt

[training.trainer]
max_epochs = 4096
strategy = auto
precision = 16-mixed

[training.optimizer]
learning_rate = 0.001

[training.logger]
logger_path = .logs
logger_name = crossface_simswap

[training.output]
directory_path = .outputs
file_pattern = crossface_simswap_{epoch}_{step}
resume_path = .outputs/last.ckpt

[exporting]
directory_path = .exports
source_path = .outputs/last.ckpt
target_path = .exports/crossface_simswap.onnx
ir_version = 10
opset_version = 15

Training

Train the model.

python train.py

Launch the TensorBoard to monitor the training.

tensorboard --logdir .logs

Exporting

Export the model to ONNX.

python export.py

1.4 KiB Raw Permalink Blame History