Loading modules/script_callbacks.py +7 −1 Original line number Original line Diff line number Diff line Loading @@ -29,7 +29,7 @@ class ImageSaveParams: class CFGDenoiserParams: class CFGDenoiserParams: def __init__(self, x, image_cond, sigma, sampling_step, total_sampling_steps): def __init__(self, x, image_cond, sigma, sampling_step, total_sampling_steps, tensor, uncond): self.x = x self.x = x """Latent image representation in the process of being denoised""" """Latent image representation in the process of being denoised""" Loading @@ -45,6 +45,12 @@ class CFGDenoiserParams: self.total_sampling_steps = total_sampling_steps self.total_sampling_steps = total_sampling_steps """Total number of sampling steps planned""" """Total number of sampling steps planned""" self.tensor = tensor """ Encoder hidden states of conditioning""" self.uncond = uncond """ Encoder hidden states of unconditioning""" class CFGDenoisedParams: class CFGDenoisedParams: def __init__(self, x, sampling_step, total_sampling_steps): def __init__(self, x, sampling_step, total_sampling_steps): Loading modules/sd_samplers_kdiffusion.py +3 −1 Original line number Original line Diff line number Diff line Loading @@ -101,11 +101,13 @@ class CFGDenoiser(torch.nn.Module): sigma_in = torch.cat([torch.stack([sigma[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [sigma] + [sigma]) sigma_in = torch.cat([torch.stack([sigma[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [sigma] + [sigma]) image_cond_in = torch.cat([torch.stack([image_cond[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [image_cond] + [torch.zeros_like(self.init_latent)]) image_cond_in = torch.cat([torch.stack([image_cond[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [image_cond] + [torch.zeros_like(self.init_latent)]) denoiser_params = CFGDenoiserParams(x_in, image_cond_in, sigma_in, state.sampling_step, state.sampling_steps) denoiser_params = CFGDenoiserParams(x_in, image_cond_in, sigma_in, state.sampling_step, state.sampling_steps, tensor, uncond) cfg_denoiser_callback(denoiser_params) cfg_denoiser_callback(denoiser_params) x_in = denoiser_params.x x_in = denoiser_params.x image_cond_in = denoiser_params.image_cond image_cond_in = denoiser_params.image_cond sigma_in = denoiser_params.sigma sigma_in = denoiser_params.sigma tensor = denoiser_params.tensor uncond = denoiser_params.uncond if tensor.shape[1] == uncond.shape[1]: if tensor.shape[1] == uncond.shape[1]: if not is_edit_model: if not is_edit_model: Loading Loading
modules/script_callbacks.py +7 −1 Original line number Original line Diff line number Diff line Loading @@ -29,7 +29,7 @@ class ImageSaveParams: class CFGDenoiserParams: class CFGDenoiserParams: def __init__(self, x, image_cond, sigma, sampling_step, total_sampling_steps): def __init__(self, x, image_cond, sigma, sampling_step, total_sampling_steps, tensor, uncond): self.x = x self.x = x """Latent image representation in the process of being denoised""" """Latent image representation in the process of being denoised""" Loading @@ -45,6 +45,12 @@ class CFGDenoiserParams: self.total_sampling_steps = total_sampling_steps self.total_sampling_steps = total_sampling_steps """Total number of sampling steps planned""" """Total number of sampling steps planned""" self.tensor = tensor """ Encoder hidden states of conditioning""" self.uncond = uncond """ Encoder hidden states of unconditioning""" class CFGDenoisedParams: class CFGDenoisedParams: def __init__(self, x, sampling_step, total_sampling_steps): def __init__(self, x, sampling_step, total_sampling_steps): Loading
modules/sd_samplers_kdiffusion.py +3 −1 Original line number Original line Diff line number Diff line Loading @@ -101,11 +101,13 @@ class CFGDenoiser(torch.nn.Module): sigma_in = torch.cat([torch.stack([sigma[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [sigma] + [sigma]) sigma_in = torch.cat([torch.stack([sigma[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [sigma] + [sigma]) image_cond_in = torch.cat([torch.stack([image_cond[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [image_cond] + [torch.zeros_like(self.init_latent)]) image_cond_in = torch.cat([torch.stack([image_cond[i] for _ in range(n)]) for i, n in enumerate(repeats)] + [image_cond] + [torch.zeros_like(self.init_latent)]) denoiser_params = CFGDenoiserParams(x_in, image_cond_in, sigma_in, state.sampling_step, state.sampling_steps) denoiser_params = CFGDenoiserParams(x_in, image_cond_in, sigma_in, state.sampling_step, state.sampling_steps, tensor, uncond) cfg_denoiser_callback(denoiser_params) cfg_denoiser_callback(denoiser_params) x_in = denoiser_params.x x_in = denoiser_params.x image_cond_in = denoiser_params.image_cond image_cond_in = denoiser_params.image_cond sigma_in = denoiser_params.sigma sigma_in = denoiser_params.sigma tensor = denoiser_params.tensor uncond = denoiser_params.uncond if tensor.shape[1] == uncond.shape[1]: if tensor.shape[1] == uncond.shape[1]: if not is_edit_model: if not is_edit_model: Loading
