ComfyUI_FL-Trainer

★ 175

图像 LoRAsd1.5SDXLComfyUI 集成

在 ComfyUI 中用于训练图像 LoRA，兼容 sd1.5 与 SDXL；自动克隆训练组件并弹出第二终端执行训练，节点内实时显示推理样例，便于跟踪和快速上手。

💡 在 ComfyUI 内一键训练并实时预览 Image LoRA 结果，快速迭代模型。

🍴 21 Forks💻 Python🔄 2024-10-18

🔗 GitHub 原文

📦

网盘下载

复制链接后前往夸克网盘下载

https://pan.quark.cn/s/e00a65475347

📦 requirements.txt

opencv-python

📄 README

ComfyUI_FL-Trainer

Train Image Loras on both sd1.5 and SDXL. This repo git clones the pieces needed to train. It pops open a second terminal window do do the training.

It will also display the inference samples in the node itself so you can track the results.

Easy Trainer

Stripped-down version of the trainer for people who want settings that just work.

recommended settings in the screenshot! Load in your images and captions and let it rip.

Things I want to add.

More simplified nodes are similar to the easy train, with some focus on style and characters. Super easy 1 click trainers for people who just want a good result and dont wanna mess with anything or waste time.

Chainable training. If you notice on the main Kohya nodes, you can start training from an epoch or even a random Lora. With this in mind im going to build out chainable instances of training that allow you to efficiently modulate the training settings through the training process with different settings, learning rates, and even training data!