Error in Inpaint Pipeline #19

by relno - opened

Thanks for your great work. I'm not sure if this model was intended to be used in this pipeline, but the below error occurred when trying to use pipe = StableDiffusionInpaintPipeline.from_pretrained(...model...)

ValueError: Incorrect configuration settings! The config of `pipeline.unet`: FrozenDict([('sample_size', 32), ('in_channels', 4), ('out_channels', 4), ('center_input_sample', False), ('flip_sin_to_cos', True), ('freq_shift', 0), ('down_block_types', ['CrossAttnDownBlock2D', 'CrossAttnDownBlock2D', 'CrossAttnDownBlock2D', 'DownBlock2D']), ('up_block_types', ['UpBlock2D', 'CrossAttnUpBlock2D', 'CrossAttnUpBlock2D', 'CrossAttnUpBlock2D']), ('block_out_channels', [320, 640, 1280, 1280]), ('layers_per_block', 2), ('downsample_padding', 1), ('mid_block_scale_factor', 1), ('act_fn', 'silu'), ('norm_num_groups', 32), ('norm_eps', 1e-05), ('cross_attention_dim', 768), ('attention_head_dim', 8), ('_class_name', 'UNet2DConditionModel'), ('_diffusers_version', '0.8.0.dev0'), ('_name_or_path', '/root/.cache/huggingface/diffusers/models--prompthero--openjourney/snapshots/c187919d42cc45f7035ca0232a357d2dffc02882/unet')]) expects 4 but received `num_channels_latents`: 4 + `num_channels_mask`: 1 + `num_channels_masked_image`: 4 = 9. Please verify the config of `pipeline.unet` or your `mask_image` or `image` input.

i'm getting this error too. were you able to find a solution?

It is because you need to use an inpainting model such as "stabilityai/stable-diffusion-2-inpainting". If you use a model meant for Text to Image & Image to Image, it won't work.

How do I fine tune an inpainting model using Dreambooth? Same as fine tuning a regular Stable Diffusion model? I initially fine tuned a Stable Diffusion model on my photos using Dreambooth. I want to use inpainting with this model to edit and improve the generations. Thanks!

relno changed discussion status to closed

Sign up or log in to comment