AiAF's picture
PixelNinjaArt_LoRA_Flux1 Repo update. Uploading the following: Sample images of training steps 0-2000 session. LoRA file. README.md, _latent_cache archive, aptimizer.pt archive, config..yaml file. Style accuracy is Noticeably better thayt previous versions I trained on civitai. Something I also did differently Regarding training in this session is that I removed a few images from the old dataset as well as made sure that the number of images was an even number. According to my own research, as well as my own experience, Having an odd number of images causes issues with memory management (think of the baking an unevern number of muffins analogy and why doing that would be ineffecient.). Due to the sheer amount of V Ram I had available for this, Doing this probably wasn`t totally necessary, but I feel it`s a good practice and habit to do. The original data set had 93 images, but I trimmed it down to eighty six this time. Anatomical degredation seem to be way less Prevalent when using this version. This current version is trained to 2000 steps, but I will likely train it up to 4000 in the future to see if I can get even better accuracy without having any anatomical degradation. In past training sessions on the on site Civitai trainer, having images that were even Remotely NSFW caused generating images to have the same style as the data said, but have severe anatomical degradation. extra limbs. bubbly looking lambs. missing limbs in general. etc etc. Well, it did happen a little bit with this version, it was nowhere near as bad as previous versions Perhaps I had different settings, or perhaps there is there`s something different going on when using the ostris ai toolkit trainer (https://github.com/ostris/ai-toolkit & https://github.com/AiArtFactory/ai-toolkit ). To make this very long commit message short, it looks promising
94722d4
raw
history blame
3.68 kB
job: extension
config:
name: D-ART-Flux1
process:
- type: sd_trainer
training_folder: /content/output
device: cuda:0
network:
type: lora
linear: 16
linear_alpha: 16
save:
dtype: float16
save_every: 250
max_step_saves_to_keep: 10
datasets:
- folder_path: /content/dataset
caption_ext: txt
caption_dropout_rate: 0.05
shuffle_tokens: false
cache_latents_to_disk: true
resolution:
- 512
- 768
- 1024
train:
batch_size: 1
steps: 2000
gradient_accumulation_steps: 1
train_unet: true
train_text_encoder: false
content_or_style: content
gradient_checkpointing: true
noise_scheduler: flowmatch
optimizer: adamw8bit
lr: 0.0004
ema_config:
use_ema: true
ema_decay: 0.99
dtype: bf16
model:
name_or_path: black-forest-labs/FLUX.1-dev
is_flux: true
quantize: true
sample:
sampler: flowmatch
sample_every: 250
width: 1024
height: 1024
prompts:
- D-ART \(Artist\), @18dart5, @18dart3, @18dart2, @18dart1, illustration of
a woman with blonde hair, wearing a white off-the-shoulder bodysuit, She
has a neutral expression and is looking directly at the camera. The background
is a plain, light grey wall. The image has a high-quality, professional feel.
- D-ART \(Artist\), @18dart5, @18dart3, @18dart2, @18dart1, illustration of
a woman with blonde hair, wearing a black lace bra, She has a small silver
heart-shaped pendant necklace around her neck. She is looking directly at
the camera with a neutral expression. The background is a plain, light-colored
wall. The image is high quality .
- 'D-ART \(Artist\), @18dart5, @18dart3, @18dart2, @18dart1, illustration of
a woman with long blonds hair, wearing a white shirt with a plunging neckline, She
has a serious expression and is looking directly at the camera. The background
is a plain, textured grey wall. The lighting is natural, highlighting her
features. The image has a high-quality, professional feel. The woman is the
main subject, positioned in the center of the frame. '
- 'D-ART \(Artist\), @18dart5, @18dart3, @18dart2, @18dart1, illustration of
a woman with blonde hair wearing a black turtleneck sweater and a gold necklace
with a large pendant. She is sitting in the passenger seat of a car, smiling
at the camera with her right hand resting on her ear. The background shows
a clear blue sky and some greenery outside the car window. The image is high
quality and has a professional feel. '
- D-ART \(Artist\), @18dart5, @18dart3, @18dart2, @18dart1, illustration of
a woman with blonde hair, wearing a blue and black Starfleet uniform with
a gold Star Fleet com-badge on the chest. She is standing in a well lit room. The
woman has a serious expression and is looking directly at the camera. The
image has a high-quality, professional feel.
- D-ART \(Artist\), @18dart5, @18dart3, @18dart2, @18dart1, illustration of
a woman with blonde hair,wearing a blue t-shirt with white polka dots, standing
in a snowy forest with tall pine trees in the background. She has a neutral
expression and is looking directly at the camera. The image is high quality
and has a professional feel.
neg: ''
seed: 9000
walk_seed: true
guidance_scale: 4
sample_steps: 20
meta:
name: D-ART-Flux1
version: '3.0'