When I set the batch size to 2, an error message is printed: Error Occurred!: (512, 512, x, x)

[benzhangdragonplus]

I am training on imageslides, and when I set the batch_size=2 in prompts.yaml， will print error message: Error Occurred!: (512, 512, x, x)，but it won't stop training， and when I set the batch size to 1, it's normal again，why？

Here is my configuration file：

prompts_file: "trainscripts/imagesliders/data/light/prompts.yaml" pretrained_model: name_or_path: "D:\zealot\sdxl_v10VAEFix.safetensors" # you can also use .ckpt or .safetensors models v2: false # true if model is v2.x v_pred: false # true if model uses v-prediction network: type: "c3lier" # or "c3lier" or "lierla" rank: 4 # #################### alpha: 1.0 training_method: "full" # full, selfattn, xattn, noxattn, or innoxattn train: precision: "bfloat16" noise_scheduler: "ddim" # or "ddpm", "lms", "euler_a" # #################### iterations: 1600 lr: 0.0002 # #################### optimizer: "AdamW" # #################### lr_scheduler: "constant" max_denoising_steps: 50 save: name: "light_temp" path: "models" per_steps: 200 precision: "bfloat16" logging: use_wandb: false verbose: false other: use_xformers: true

• target: "" # what word for erasing the positive concept from positive: "" # concept to erase unconditional: "" # word to take the difference from the positive concept neutral: "" # starting point for conditioning the target action: "erase" # erase or enhance guidance_scale: -1 # ###################### resolution: 1024 dynamic_resolution: false batch_size: 2

[rohitgandikota]

Hey, thanks for the question! We did not configure batch size for image training. Earlier in our experiments using higher batch size was creating memory issues.