Jones Chung

out of memory on testing

2

I set the batch size to 4 during training and it works fine. However, when I tried to test, it resulted in "out of memory" CUDA errors. And I can't...

path with spaces error

1

``` 16:15:05-647695 ERROR The path in which this python code is executed contain one or many spaces. This is not supported for running kohya_ss GUI. 16:15:05-649695 ERROR Please move the...

Strange performance of Flux-dev

7

### Your question I'm testing generating a batch of images using original Flux-dev, kijai's Flux-dev-fp8, and Comfy-Org's Flux-dev-fp8 checkpoints. Use the workflows from https://comfyanonymous.github.io/ComfyUI_examples/flux/ Using models: https://huggingface.co/black-forest-labs/FLUX.1-dev https://huggingface.co/Kijai/flux-fp8 https://huggingface.co/Comfy-Org/flux1-dev Here...

### What happened? ``` Traceback (most recent call last): File "D:\AI\Stable_Diffusion\OneTrainer\modules\ui\TrainUI.py", line 544, in __training_thread_function trainer.train() File "D:\AI\Stable_Diffusion\OneTrainer\modules\trainer\GenericTrainer.py", line 514, in train self.data_loader.get_data_set().start_next_epoch() File "d:\ai\stable_diffusion\onetrainer\venv\src\mgds\src\mgds\MGDS.py", line 49, in start_next_epoch self.loading_pipeline.start_next_epoch()...

[Bug]: Huggingface error despite using SD3 model locally

1

### What happened? ``` Clearing cache directory D:/AI/Stable_Diffusion/Datasets/V6/V6/SD3/workspace-cache/run! You can disable this if you want to continue using the same cache. Traceback (most recent call last): File "D:\AI\Stable_Diffusion\OneTrainer\modules\modelLoader\stableDiffusion3\StableDiffusion3ModelLoader.py", line 232,...

Flux LoRA resumes wrong epochs

1

I am training a LoRA with 15 images. I saved a state on epoch 4. The `train_state.json` file inside `x-0000004-state` is `{"current_epoch": 4, "current_step": 60}` And then I resume `x-0000004-state`...

[SD3-Flux.1] "train_blocks=single" not work

1

I have checked the json in model folder. Line 163: `"train_blocks": "single",`. But on the same toml. Line 38: `network_args = []`. In previous versions, it would have been `network_args...

Add support for python 3.10.14 or higher

1

The closest environments that Anaconda can create are 3.9.19 and 3.10.14, but they are just beyond the support range of 3.10.1 ~ 3.10.12. When installing ml-agents-envs, an error is encountered...

The sidebar is full of chinese characters

5

### Expected Behavior My UI language is English ### Actual Behavior Too many Chinese characters ### Steps to Reproduce  ### Debug Logs ```powershell --- ``` ### Other _No response_