fusing

/

latent-diffusion-text2im-large

like 6

Model version: 030d95399e64c1ff7be798e62b9bc6ef6ec7bf3b
diffusers version: 36b459f6e644426a5e70639d2338d8dac4df367c https://github.com/huggingface/diffusers/tree/36b459f6e644426a5e70639d2338d8dac4df367c

AttributeError: 'FrozenDict' object has no attribute 'timesteps'

File ~/.cache/.../lib/python3.10/site-packages/diffusers/pipelines/latent_diffusion/pipeline_latent_diffusion.py:584, in LatentDiffusionPipeline.__call__(self, prompt, batch_size, generator, torch_device, eta, guidance_scale, num_inference_steps)
    581 text_input = self.tokenizer(prompt, padding="max_length", max_length=77, return_tensors="pt").to(torch_device)
    582 text_embedding = self.bert(text_input.input_ids)
--> 584 num_trained_timesteps = self.scheduler.config.timesteps
    585 inference_step_times = range(0, num_trained_timesteps, num_trained_timesteps // num_inference_steps)
    587 image = torch.randn(
    588     (batch_size, self.unet.in_channels, self.unet.image_size, self.unet.image_size),
    589     generator=generator,
    590 ).to(torch_device)