Giters
kohya-ss
/
sd-scripts
Geek Repo:
Geek Repo
Github PK Tool:
Github PK Tool
Stargazers:
4311
Watchers:
44
Issues:
787
Forks:
734
kohya-ss/sd-scripts Issues
How to set --network_train_unet_only with latest version GUI?
Updated
5 days ago
Comments count
1
AttributeError: module 'torch.optim.lr_scheduler' has no attribute 'constant_with_warmup'
Updated
6 days ago
Fused optimizerでのresume
Updated
7 days ago
prodigy optimizer dont apply changes to lora weights
Updated
7 days ago
Stop text encoder at % steps for full finetuning
Closed
9 days ago
Model diverges using deepspeed fp16 mixed-precision training
Updated
11 days ago
Comments count
3
Problem in understanding change log of 0.8.5
Updated
13 days ago
Comments count
2
Something changed. Now LR is strongly applied to SDXL. What could be change?
Updated
13 days ago
Comments count
1
Urgent help, how can I solve this problem while training DreamDance?Thank you very much!
Updated
14 days ago
Comments count
4
stable_cascade branch update or merge into the main branch?
Updated
15 days ago
UnboundLocalError occurred when not using onnx and with --max_data_loader_n_workers
Closed
17 days ago
Error during multi-CPU training
Updated
17 days ago
RuntimeError: Calculated padded input size per channel: (514 x 2). Kernel size: (3 x 3). Kernel size can't be greater than actual input size
Updated
17 days ago
Logging when using sdxl_train_network.py
Updated
19 days ago
Can this be configured in a Linux environment?
Updated
19 days ago
Comments count
6
SDXL training on single GPU fine but multi-gpu Out of memory
Closed
19 days ago
Comments count
1
How to disable lora
Updated
20 days ago
Problem in understanding the change log in v0.8.5
Updated
21 days ago
Implement exponential moving averages (EMA)
Updated
21 days ago
Comments count
3
warn or transpose images with exif rotation
Updated
23 days ago
Multichannel input (TIF files)
Updated
23 days ago
ask for help
Updated
23 days ago
Does anyone knows the usage of "block learning rate" for SDXL LoRA?
Updated
24 days ago
Comments count
3
num_samples should be a positive integer value, but got num_samples=0
Closed
a month ago
caption_separator cannot be specified in dataset config toml
Updated
a month ago
Comments count
7
error when dataset config has subsets with different flip_aug settings
Updated
a month ago
dylora broken, ValueError: Attempting to unscale FP16 gradients
Updated
a month ago
caption joined by ", " instead of caption_separator when shuffling
Updated
a month ago
Does BOFT-training switch to float32?
Updated
a month ago
Comments count
1
Added the implement IP-adapter?
Updated
a month ago
Trying to train with Masked Loss on a compound dataset containing both masked and normal datasets results in KeyError: 'conditioning_images'
Updated
a month ago
Tutorial for train_controlnet
Updated
a month ago
the generated images have a yellowish tone
Updated
a month ago
Comments count
3
How to adapt to webdataset data format?
Updated
a month ago
Don't understand why it always shows an error saying 'No data found. Please verify arguments.
Closed
a month ago
Comments count
3
cache_text_encoder_outputs.py raises AttributeError: 'Namespace' object has no attribute 'deepspeed'
Updated
a month ago
Comments count
2
Help to fix the code while running in the colab
Updated
a month ago
Stuck on step 0/xxxx, no GPU resources used
Updated
a month ago
Using Mulit-GPU in windows system with NCCL error
Updated
a month ago
Training Ultra Slow with more than 1 GPU - very likely affecting all users with more than 1 GPU
Updated
a month ago
Some RGBA images cause broken backgrounds due to incorrect RGB conversion
Updated
a month ago
Comments count
4
Issue with create_network_from_weights
Closed
a month ago
Comments count
4
How multi-gpu latent caching works?
Updated
a month ago
Comments count
4
When adaptive_noise_scale = 0 is given still tries to verify noise_offset
Closed
a month ago
Comments count
2
New training broken on Kaggle due to DistributedDataParallel and torch.distributed.elastic.multiprocessing.api
Updated
a month ago
Comments count
4
NaN showing when training LORA model
Updated
a month ago
Comments count
1
question about the how to training dreambooth-sdxl
Closed
a month ago
Comments count
2
question about the batch_size in training lora
Closed
a month ago
Comments count
4
[deepspeed bug] sd1 fine_tune.py, mixed precision with deepspeed not working
Updated
a month ago
Training stuck at step zero
Closed
a month ago
Comments count
3
Previous
Next