OneTrainer

OneTrainer copied to clipboard

**Description:** This pull request introduces modifications to the Text Encoder to support input token lengths exceeding 75 tokens, specifically up to 3 chunks. The implementation has been confirmed to work...

celll1

### Describe your use-case. The possibility to train from flux Schnell trough a training adapter. ### What would you like to see as a solution? Here's all the information https://huggingface.co/ostris/FLUX.1-schnell-training-adapter...

Mixomo

[Feat]: LoRA - Advanced Layer Filter

1

### Describe your use-case. Flux has layers named single_transformer_blocks.* and transformer_blocks.*. If I want to train only the **transformer_blocks.*** layers but exclude **single_transformer_blocks.***, the current filtering method won't suffice. ###...

bananasss00

[Bug]: Just installed but can't start it

1

### What happened? I just did an automatic install in Windows. Whenever I try to run start-ui.bat, I get the log below. Already searched for "ModuleNotFoundError: No module named 'torch'"...

Iagmon

[Bug]: ModuleNotFoundError: No module named 'util'

4

### What happened? I tried running OneTrainer using start-ui.bat and instead got an error regarding being unable to import 'util'. I've tried git clone, updating, and reinstalling several times, both...

InvaderX9

[Bug]: Can't train SD1.5 - no model

5

### What happened? Training with 1.5 doesn't work after new install. I can't find the model directory to download and put it on there manually. Repository Not Found for url:...

michaelLochlann

[Bug]: Setting "train" to false for TEs does not freeze TEs during SDXL finetuning

5

### What happened? Configuring TEs as follows: ``` "text_encoder": { "train": false, "learning_rate": 2e-8, "layer_skip": 0, "weight_dtype": "FLOAT_32", "stop_training_after": 5000, "stop_training_after_unit": "STEP" }, "text_encoder_2": { "train": false, "learning_rate": 1e-8, "layer_skip":...

orcinus

[Bug]: SDXL DoRA training fails to start when base model weight data type is set to float8

4

### What happened? When I activate Decomposed Weights (DoRA) training in "Lora" tab and have base SDXL model loaded as float8 in "model" tab the training process fails to start...

ikarsokolov

[Bug]: Repeated error during training after 1 epoch and 10% of the next.

3

[instagirl_config.json](https://github.com/user-attachments/files/16643076/instagirl_config.json) ### What happened? I've been experiencing the same error for the last few months every time I try to train a model (SDXL finetune) I get an error that...

velourlawsuits

### Describe your use-case. It's hard to find presets in the current list if the preset count grows in time:  ### What would you like to see as a...

aa956