kohya-ss sd-scripts issues

kohya-ss / sd-scripts

Apache License 2.0

5.31k stars 880 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Fix CNet training for sd1 (and maybe 2)

#1646 elthariel closed 1 month ago
2
Dataset maker sd1.5 - wd-tagger module stopped working on colab

#1645 nevo993 opened 2 months ago
0
Please add AdEMAMix to the scripts.

#1644 DarkAlchy closed 2 months ago
2
Regarding the Multi-Level Caption

#1643 sdbds opened 2 months ago
3
about flux image normalization

#1642 Jimzhou82sub closed 2 months ago
2
Kohya_ss native trainer colab broken yesterday. Captured before/ after install log. Jax, Key array.

#1641 SarahPeterson2854 opened 2 months ago
7
New optimizer:AdEMAMix8bit and PagedAdEMAMix8bit

#1640 sdbds closed 2 months ago
9
Was "Validation Loss" ever successfully implemented?

#1638 mostlyhuman opened 2 months ago
0
Update train_network_README-zh.md

#1637 a67793581 opened 2 months ago
0
Specify blocks to Train w/ Finetuning?

#1636 setothegreat opened 2 months ago
3
Drop text encoder weights

#1635 oovm closed 2 months ago
1
FLUX LoRA Training - Text Encoder active training indicated when it should be disabled with --network_train_unet_only

#1634 Enyakk opened 2 months ago
2
FLUX: Bleeding?

#1633 DarkAlchy opened 2 months ago
0
fix import and make npz for alpha_mask

#1632 Maru-mee closed 1 month ago
17
Recommended settings for ScheduleFree.

#1631 waomodder closed 2 months ago
4
Issue Reminder

#1630 lynrayy opened 2 months ago
0
Extract Flux Lora not working after 93/304

#1629 DarkViewAI closed 2 months ago
0
Make timesteps work in the standard way when Huber loss is used

#1628 recris closed 2 months ago
6
When I train with multiple GPUs, this error will be reported. This has been bothering me for a long time. How can I solve it

#1627 jinwei1660 closed 2 months ago
1
how to use flux full-finetune safetensors in diffusers?

#1626 neuron-party opened 2 months ago
2
Strange behavior with --network_args parameter

#1624 codecomputation opened 2 months ago
0
neither 'setup.py' nor 'pyproject.toml' found.

#1623 wzgrx closed 2 months ago
0
Inverted extraction for FLUX.

#1622 DarkAlchy closed 2 months ago
4
lora to original flux lora

#1621 LixDemon closed 2 months ago
1
fix: backward compatibility for text_encoder_lr

#1620 Akegarasu closed 2 months ago
1
Retain alpha in `pil_resize` for `--alpha_mask`

#1619 emcmanus closed 2 months ago
4
TypeError: get_cosine_schedule_with_warmup() got an unexpected keyword argument 'num_decay_steps'

#1618 AndreR closed 1 month ago
6
AttributeError: 'LycorisNetworkKohya' object has no attribute 'train_t5xxl'

#1617 sd2530615 closed 2 months ago
4
ValueError: invalid literal for int() with base 10: 'still' when using facecrop

#1616 Seedmanc closed 2 months ago
1
Bug fix: alpha_mask load

#1615 Maru-mee closed 2 months ago
5
Issue for convert_flux_lora.py: RuntimeError: Some tensors share memory...

#1614 terrificdm closed 1 month ago
5
[Feature Request] Decouple `linear1` and `linear2` Flux layers in network_args

#1613 EricBCoding opened 2 months ago
8
Flux merge tool does not properly handle LoRA keys from AI-Toolkit when merging w/ Flux.dev model - All keys "Unused" / Unchanged model file outputted

#1612 CCpt5 opened 2 months ago
9
Does the FLUX/SD3 branch not work with SDXL yet?

#1611 DarkAlchy closed 2 months ago
0
FLUX LoRA Extraction Resemblance Issue

#1610 FurkanGozukara closed 2 months ago
1
Either "cuDNN SDPA backward got grad_output.strides() != output.strides()" or OutOfMemory

#1609 Tablaski closed 2 months ago
5
Flux does not support Prodigy?

#1608 DarkAlchy closed 2 months ago
2
Flux fine-tune text encoder rate is too high; need to be able to set separately

#1607 BelieveDiffusion opened 2 months ago
2
Update submodules as well since MPS changes were needed

#1606 JoeyOverby opened 2 months ago
0
sd3 schedule free opt

#1605 kohya-ss closed 2 months ago
6
flux lora training does not start,error: returned non-zero exit status 1

#1603 zxf5843029 closed 2 months ago
4
unexpected keyword argument 'num_decay_steps' when using cosine or linear scheduler (sd3 branch)

#1602 Enyakk closed 2 months ago
1
poolFD failed error when caching latents for flux finetuning

#1601 huxian0402 opened 2 months ago
4
New ScheduleFree support for Flux

#1600 sdbds closed 2 months ago
2
上手く学習が完了しません

#1599 bell123405 closed 2 months ago
4
flux with flow matching schedule

#1598 Jimzhou82sub closed 2 months ago
0
Error loading fbgemm.dll

#1597 AndreR opened 2 months ago
1
How to convert a safetensors(finetuned sdxl model) to diffusers dir format?

#1596 laolongboy opened 2 months ago
0
FLUX training for 8GB VRAM?

#1595 cocktailpeanut opened 2 months ago
12
What does split_mode actually do?

#1594 WainWong closed 2 months ago
2

Previous Next