Loading settings from /content/LoRA/config/config_file.toml...
/content/LoRA/config/config_file
prepare tokenizer
Downloading (…)olve/main/vocab.json: 100% 961k/961k [00:00<00:00, 1.13MB/s]
Downloading (…)olve/main/merges.txt: 100% 525k/525k [00:00<00:00, 823kB/s]
Downloading (…)cial_tokens_map.json: 100% 389/389 [00:00<00:00, 257kB/s]
Downloading (…)okenizer_config.json: 100% 905/905 [00:00<00:00, 399kB/s]
update token length: 225
Load dataset config from /content/LoRA/config/dataset_config.toml
prepare images.
found directory /content/LoRA/train_data contains 11 image files
2750 train images with repeating.
0 reg images.
no regularization images / 正則化画像が見つかりませんでした
[Dataset 0]
  batch_size: 6
  resolution: (512, 512)
  enable_bucket: True
  min_bucket_reso: 256
  max_bucket_reso: 1024
  bucket_reso_steps: 64
  bucket_no_upscale: False

  [Subset 0 of Dataset 0]
    image_dir: "/content/LoRA/train_data"
    image_count: 11
    num_repeats: 250
    shuffle_caption: True
    keep_tokens: 0
    caption_dropout_rate: 0
    caption_dropout_every_n_epoches: 0
    caption_tag_dropout_rate: 0
    color_aug: False
    flip_aug: False
    face_crop_aug_range: None
    random_crop: False
    token_warmup_min: 1,
    token_warmup_step: 0,
    is_reg: False
    class_tokens: mksks
    caption_extension: .txt


[Dataset 0]
loading image sizes.
100% 11/11 [00:00<00:00, 439.09it/s]
make buckets
number of images (including repeats) / 各bucketの画像枚数（繰り返し回数を含む）
bucket 0: resolution (384, 640), count: 250
bucket 1: resolution (512, 512), count: 1000
bucket 2: resolution (576, 448), count: 500
bucket 3: resolution (640, 384), count: 1000
mean ar error (without repeats): 0.0969292682863697
prepare accelerator
Using accelerator 0.15.0 or above.
loading model for process 0/1
load StableDiffusion checkpoint
loading u-net: <All keys matched successfully>
loading vae: <All keys matched successfully>
Downloading (…)lve/main/config.json: 100% 4.52k/4.52k [00:00<00:00, 3.13MB/s]
Downloading pytorch_model.bin: 100% 1.71G/1.71G [00:23<00:00, 73.3MB/s]
loading text encoder: <All keys matched successfully>
Replace CrossAttention.forward to use xformers
[Dataset 0]
caching latents.
100% 5/5 [00:13<00:00,  2.62s/it]
import network module: lycoris.kohya
Using rank adaptation algo: lora
Use Dropout value: 0.0
Create LyCORIS Module
create LyCORIS for Text Encoder: 72 modules.
Create LyCORIS Module
create LyCORIS for U-Net: 278 modules.
enable LyCORIS for text encoder
enable LyCORIS for U-Net
prepare optimizer, data loader etc.
Deprecated: use prepare_optimizer_params(text_encoder_lr, unet_lr, learning_rate) instead of prepare_optimizer_params(text_encoder_lr, unet_lr)
CUDA SETUP: CUDA runtime path found: /usr/local/cuda-11.8/targets/x86_64-linux/lib/libcudart.so
CUDA SETUP: Highest compute capability among GPUs detected: 7.5
CUDA SETUP: Detected CUDA version 118
CUDA SETUP: Loading binary /usr/local/lib/python3.10/dist-packages/bitsandbytes/libbitsandbytes_cuda118.so...
use 8-bit AdamW optimizer | {}
override steps. steps for 2 epochs is / 指定エポックまでのステップ数: 920
running training / 学習開始
  num train images * repeats / 学習画像の数×繰り返し回数: 2750
  num reg images / 正則化画像の数: 0
  num batches per epoch / 1epochのバッチ数: 460
  num epochs / epoch数: 2
  batch size per device / バッチサイズ: 6
  gradient accumulation steps / 勾配を合計するステップ数 = 1
  total optimization steps / 学習ステップ数: 920
steps:   0% 0/920 [00:00<?, ?it/s]epoch 1/2
steps:  50% 460/920 [11:32<11:32,  1.51s/it, loss=0.0465]saving checkpoint: /content/drive/MyDrive/LoRA/output/last-000001.safetensors
generating sample images at step / サンプル画像生成 ステップ: 460
prompt: (masterpiece, best quality, highres, illustration, perfect anatomy), 1girl, solo, 40 years old mature female, sharp mature face, lipstick, red lips, beautiful pupil, sexy blue eyes:1.2, highlights in eyes:1.4, short hair, parted bangs, blonde hair, glasses, long nose, glamorous proportion, business suit  
negative_prompt: lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry  
height: 768
width: 512
sample_steps: 28
scale: 7.0

  0% 0/28 [00:00<?, ?it/s]
  4% 1/28 [00:00<00:15,  1.73it/s]
  7% 2/28 [00:00<00:11,  2.35it/s]
 11% 3/28 [00:01<00:09,  2.63it/s]
 14% 4/28 [00:01<00:08,  2.77it/s]
 18% 5/28 [00:01<00:07,  2.88it/s]
 21% 6/28 [00:02<00:07,  2.95it/s]
 25% 7/28 [00:02<00:07,  2.97it/s]
 29% 8/28 [00:02<00:06,  3.00it/s]
 32% 9/28 [00:03<00:06,  3.02it/s]
 36% 10/28 [00:03<00:05,  3.03it/s]
 39% 11/28 [00:03<00:05,  3.04it/s]
 43% 12/28 [00:04<00:05,  3.05it/s]
 46% 13/28 [00:04<00:04,  3.05it/s]
 50% 14/28 [00:04<00:04,  3.05it/s]
 54% 15/28 [00:05<00:04,  3.05it/s]
 57% 16/28 [00:05<00:03,  3.05it/s]
 61% 17/28 [00:05<00:03,  3.05it/s]
 64% 18/28 [00:06<00:03,  3.06it/s]
 68% 19/28 [00:06<00:02,  3.05it/s]
 71% 20/28 [00:06<00:02,  3.05it/s]
 75% 21/28 [00:07<00:02,  3.05it/s]
 79% 22/28 [00:07<00:01,  3.05it/s]
 82% 23/28 [00:07<00:01,  3.05it/s]
 86% 24/28 [00:08<00:01,  3.05it/s]
 89% 25/28 [00:08<00:00,  3.05it/s]
 93% 26/28 [00:08<00:00,  3.06it/s]
 96% 27/28 [00:09<00:00,  3.06it/s]
100% 28/28 [00:09<00:00,  2.98it/s]
epoch 2/2
steps: 100% 920/920 [23:20<00:00,  1.52s/it, loss=0.0274]generating sample images at step / サンプル画像生成 ステップ: 920
prompt: (masterpiece, best quality, highres, illustration, perfect anatomy), 1girl, solo, 40 years old mature female, sharp mature face, lipstick, red lips, beautiful pupil, sexy blue eyes:1.2, highlights in eyes:1.4, short hair, parted bangs, blonde hair, glasses, long nose, glamorous proportion, business suit  
negative_prompt: lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry  
height: 768
width: 512
sample_steps: 28
scale: 7.0

  0% 0/28 [00:00<?, ?it/s]
  4% 1/28 [00:00<00:08,  3.11it/s]
  7% 2/28 [00:00<00:08,  3.10it/s]
 11% 3/28 [00:00<00:08,  3.01it/s]
 14% 4/28 [00:01<00:07,  3.05it/s]
 18% 5/28 [00:01<00:07,  3.07it/s]
 21% 6/28 [00:01<00:07,  3.04it/s]
 25% 7/28 [00:02<00:06,  3.04it/s]
 29% 8/28 [00:02<00:06,  3.06it/s]
 32% 9/28 [00:02<00:06,  3.05it/s]
 36% 10/28 [00:03<00:05,  3.05it/s]
 39% 11/28 [00:03<00:05,  3.06it/s]
 43% 12/28 [00:03<00:05,  3.06it/s]
 46% 13/28 [00:04<00:04,  3.06it/s]
 50% 14/28 [00:04<00:04,  3.07it/s]
 54% 15/28 [00:04<00:04,  3.06it/s]
 57% 16/28 [00:05<00:03,  3.06it/s]
 61% 17/28 [00:05<00:03,  3.06it/s]
 64% 18/28 [00:05<00:03,  3.06it/s]
 68% 19/28 [00:06<00:02,  3.05it/s]
 71% 20/28 [00:06<00:02,  3.06it/s]
 75% 21/28 [00:06<00:02,  3.06it/s]
 79% 22/28 [00:07<00:01,  3.05it/s]
 82% 23/28 [00:07<00:01,  3.05it/s]
 86% 24/28 [00:07<00:01,  3.05it/s]
 89% 25/28 [00:08<00:00,  3.05it/s]
 93% 26/28 [00:08<00:00,  3.05it/s]
 96% 27/28 [00:08<00:00,  3.05it/s]
100% 28/28 [00:09<00:00,  3.05it/s]
save trained model to /content/drive/MyDrive/LoRA/output/last.safetensors
model saved.
steps: 100% 920/920 [23:31<00:00,  1.53s/it, loss=0.0274]