(self, x, denoiser, alpha_cumprod_sqrt, cond, uc, timestep=None, idx=None, scale=None, scale_emb=None)
| 501 | return x, s_in, alpha_cumprod_sqrt, num_sigmas, cond, uc, timesteps |
| 502 | |
| 503 | def denoise(self, x, denoiser, alpha_cumprod_sqrt, cond, uc, timestep=None, idx=None, scale=None, scale_emb=None): |
| 504 | additional_model_inputs = {} |
| 505 | |
| 506 | if isinstance(scale, torch.Tensor) == False and scale == 1: |
| 507 | additional_model_inputs["idx"] = x.new_ones([x.shape[0]]) * timestep |
| 508 | if scale_emb is not None: |
| 509 | additional_model_inputs["scale_emb"] = scale_emb |
| 510 | denoised = denoiser(x, alpha_cumprod_sqrt, cond, **additional_model_inputs).to(torch.float32) |
| 511 | else: |
| 512 | additional_model_inputs["idx"] = torch.cat([x.new_ones([x.shape[0]]) * timestep] * 2) |
| 513 | denoised = denoiser( |
| 514 | *self.guider.prepare_inputs(x, alpha_cumprod_sqrt, cond, uc), **additional_model_inputs |
| 515 | ).to(torch.float32) |
| 516 | if isinstance(self.guider, DynamicCFG): |
| 517 | denoised = self.guider( |
| 518 | denoised, (1 - alpha_cumprod_sqrt**2) ** 0.5, step_index=self.num_steps - timestep, scale=scale |
| 519 | ) |
| 520 | else: |
| 521 | denoised = self.guider(denoised, (1 - alpha_cumprod_sqrt**2) ** 0.5, scale=scale) |
| 522 | return denoised |
| 523 | |
| 524 | def sampler_step( |
| 525 | self, |
no test coverage detected