The following values were not passed to `accelerate launch` and had defaults used instead: `--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`. /workspace/sd-scripts/library/lpw_stable_diffusion.py:70: SyntaxWarning: invalid escape sequence '\(' \( - literal character '(' /workspace/sd-scripts/library/strategy_base.py:94: SyntaxWarning: invalid escape sequence '\(' \( - literal character '(' /workspace/sd-scripts/library/custom_train_functions.py:174: SyntaxWarning: invalid escape sequence '\(' \( - literal character '(' 2026-07-24 13:34:00 INFO Loading dataset config from train_network.py:926 /workspace/dataset.toml INFO prepare images. dreambooth_dataset.py:274 INFO get image size from name dreambooth_dataset.py:149 of cache files 0%| | 0/268 [00:00 INFO Loaded Qwen3 text encoder. anima_utils.py:267 Parameters: 596,049,920 INFO Loading Anima VAE... anima_train_network.py:107 INFO Using image-only anima_train_utils.py:205 Qwen-Image 2D VAE INFO Initializing qwen_image_autoencoder_kl_2d.py:535 image-only QwenImage 2D VAE INFO Loading VAE qwen_image_autoencoder_kl_2d.py:550 from /workspace/mode ls/anima/split_ files/vae/qwen_ image_vae.safet ensors INFO Converted ComfyUI qwen_image_autoencoder_kl.py:1538 AutoencoderKL state dict keys to official format INFO Skipped 8 qwen_image_autoencoder_kl_2d.py:454 temporal-only VAE weights for QwenImage 2D VAE INFO Loaded qwen_image_autoencoder_kl_2d.py:555 image-only QwenImage 2D VAE: INFO [Dataset 0] dataset.py:1311 INFO caching latents with caching dataset.py:752 strategy. INFO caching latents... dataset.py:802 accelerator device: cuda 0%| | 0/268 [00:00