diff --git "a/dongman/训练日志.txt" "b/dongman/训练日志.txt" new file mode 100644--- /dev/null +++ "b/dongman/训练日志.txt" @@ -0,0 +1,371 @@ +14:19:54-924440 INFO Starting SD-Trainer Mikazuki GUI... +14:19:54-927178 INFO Base directory: /root/lora-scripts, Working directory: /root/lora-scripts +14:19:54-927884 INFO Linux Python 3.10.9 /root/.conda/envs/lora/bin/python +14:19:54-933201 INFO Starting tageditor... +14:19:54-936010 INFO Starting tensorboard... +14:19:58-704848 INFO Server started at http://127.0.0.1:28000 +TensorFlow installation not found - running with reduced feature set. + +NOTE: Using experimental fast data loading logic. To disable, pass + "--load_fast=false" and report issues on GitHub. More details: + https://github.com/tensorflow/tensorboard/issues/4784 + +TensorBoard 2.10.1 at http://127.0.0.1:6006/ (Press CTRL+C to quit) +14:20:52-995764 INFO Torch 2.3.0+cu121 +14:20:53-470315 INFO Torch backend: nVidia CUDA 12.1 cuDNN 8902 +14:20:53-876159 INFO Torch detected GPU: NVIDIA A100-SXM4-80GB VRAM 81051 Arch (8, 0) Cores 108 +14:24:14-346637 INFO Training started with config file / 训练开始，使用配置文件: /root/lora-scripts/config/autosave/20240716-142414.toml +14:24:14-349477 INFO Task 6fd14190-b173-4715-b710-7293b373447e created +The following values were not passed to `accelerate launch` and had defaults used instead: + `--num_processes` was set to a value of `1` + `--num_machines` was set to a value of `1` + `--mixed_precision` was set to a value of `'no'` + `--dynamo_backend` was set to a value of `'no'` +To avoid this warning pass in values for each of the problematic parameters or run `accelerate config`. +2024-07-16 14:24:58 INFO Loading settings from /root/lora-scripts/config/autosave/20240716-142414.toml... train_util.py:3744 + INFO /root/lora-scripts/config/autosave/20240716-142414 train_util.py:3763 +2024-07-16 14:24:58 INFO prepare tokenizer train_util.py:4227 +2024-07-16 14:24:59 INFO update token length: 255 train_util.py:4244 +2024-07-16 14:25:00 INFO prepare images. train_util.py:1572 + INFO found directory /train6/1_dongman contains 916 image files train_util.py:1519 + INFO 916 train images with repeating. train_util.py:1613 + INFO 0 reg images. train_util.py:1616 + WARNING no regularization images / 正則化画像が見つかりませんでした train_util.py:1621 + INFO [Dataset 0] config_util.py:565 + batch_size: 64 + resolution: (1024, 1024) + enable_bucket: True + network_multiplier: 1.0 + min_bucket_reso: 256 + max_bucket_reso: 2048 + bucket_reso_steps: 64 + bucket_no_upscale: False + + [Subset 0 of Dataset 0] + image_dir: "/train6/1_dongman" + image_count: 916 + num_repeats: 1 + shuffle_caption: True + keep_tokens: 0 + keep_tokens_separator: + secondary_separator: None + enable_wildcard: False + caption_dropout_rate: 0.0 + caption_dropout_every_n_epoches: 0 + caption_tag_dropout_rate: 0.0 + caption_prefix: None + caption_suffix: None + 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: dongman + caption_extension: .txt + + + INFO [Dataset 0] config_util.py:571 + INFO loading image sizes. train_util.py:853 +100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 916/916 [00:00<00:00, 85199.42it/s] + INFO make buckets train_util.py:859 + INFO number of images (including repeats) / 各bucketの画像枚数（繰り返し回数を含む） train_util.py:905 + INFO bucket 0: resolution (704, 1408), count: 29 train_util.py:910 + INFO bucket 1: resolution (768, 1280), count: 6 train_util.py:910 + INFO bucket 2: resolution (768, 1344), count: 723 train_util.py:910 + INFO bucket 3: resolution (832, 1216), count: 123 train_util.py:910 + INFO bucket 4: resolution (1216, 832), count: 2 train_util.py:910 + INFO bucket 5: resolution (1344, 768), count: 33 train_util.py:910 + INFO mean ar error (without repeats): 0.011380946831128346 train_util.py:915 + INFO prepare accelerator train_db.py:106 +wandb: Currently logged in as: cn42083120024 (renwu). Use `wandb login --relogin` to force relogin +wandb: Appending key for api.wandb.ai to your netrc file: /root/.netrc +accelerator device: cuda +2024-07-16 14:25:24 INFO loading model for process 0/1 train_util.py:4385 + INFO load StableDiffusion checkpoint: ./sd-models/model.safetensors train_util.py:4341 +2024-07-16 14:25:29 INFO UNet2DConditionModel: 64, 8, 768, False, False original_unet.py:1387 +2024-07-16 14:25:56 INFO loading u-net: model_util.py:1009 +2024-07-16 14:26:01 INFO loading vae: model_util.py:1017 +2024-07-16 14:26:09 INFO loading text encoder: model_util.py:1074 + INFO Enable xformers for U-Net train_util.py:2660 + INFO [Dataset 0] train_util.py:2079 + INFO caching latents. train_util.py:974 + INFO checking cache validity... train_util.py:984 +100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████���█████████████████| 916/916 [00:00<00:00, 4252.76it/s] +2024-07-16 14:26:10 INFO caching latents... train_util.py:1021 +100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 916/916 [08:53<00:00, 1.72it/s] +2024-07-16 14:35:03 INFO CrossAttnDownBlock2D False -> True original_unet.py:1521 + INFO CrossAttnDownBlock2D False -> True original_unet.py:1521 + INFO CrossAttnDownBlock2D False -> True original_unet.py:1521 + INFO DownBlock2D False -> True original_unet.py:1521 + INFO UNetMidBlock2DCrossAttn False -> True original_unet.py:1521 + INFO UpBlock2D False -> True original_unet.py:1521 + INFO CrossAttnUpBlock2D False -> True original_unet.py:1521 + INFO CrossAttnUpBlock2D False -> True original_unet.py:1521 + INFO CrossAttnUpBlock2D False -> True original_unet.py:1521 +prepare optimizer, data loader etc. +2024-07-16 14:35:04 INFO use 8-bit AdamW optimizer | {} train_util.py:3889 +override steps. steps for 30 epochs is / 指定エポックまでのステップ数: 540 +running training / 学習開始 + num train images * repeats / 学習画像の数×繰り返し回数: 916 + num reg images / 正則化画像の数: 0 + num batches per epoch / 1epochのバッチ数: 18 + num epochs / epoch数: 30 + batch size per device / バッチサイズ: 64 + total train batch size (with parallel & distributed & accumulation) / 総バッチサイズ（並列学習、勾配合計含む）: 64 + gradient ccumulation steps / 勾配を合計するステップ数 = 1 + total optimization steps / 学習ステップ数: 540 +steps: 0%| | 0/540 [00:00 True original_unet.py:1521 + INFO CrossAttnDownBlock2D False -> True original_unet.py:1521 + INFO CrossAttnDownBlock2D False -> True original_unet.py:1521 + INFO DownBlock2D False -> True original_unet.py:1521 + INFO UNetMidBlock2DCrossAttn False -> True original_unet.py:1521 + INFO UpBlock2D False -> True original_unet.py:1521 + INFO CrossAttnUpBlock2D False -> True original_unet.py:1521 + INFO CrossAttnUpBlock2D False -> True original_unet.py:1521 + INFO CrossAttnUpBlock2D False -> True original_unet.py:1521 +prepare optimizer, data loader etc. +2024-07-16 14:35:04 INFO use 8-bit AdamW optimizer | {} train_util.py:3889 +override steps. steps for 30 epochs is / 指定エポックまでのステップ数: 540 +running training / 学習開始 + num train images * repeats / 学習画像の数×繰り返し回数: 916 + num reg images / 正則化画像の数: 0 + num batches per epoch / 1epochのバッチ数: 18 + num epochs / epoch数: 30 + batch size per device / バッチサイズ: 64 + total train batch size (with parallel & distributed & accumulation) / 総バッチサイズ（並列学習、勾配合計含む）: 64 + gradient ccumulation steps / 勾配を合計するステップ数 = 1 + total optimization steps / 学習ステップ数: 540 +steps: 0%| | 0/540 [00:00