Function process_decode

modules/processing_diffusers.py:470–527

(p: processing.StableDiffusionProcessing, output)

def process_decode(p: processing.StableDiffusionProcessing, output):
    shared.sd_model = sd_models.apply_balanced_offload(shared.sd_model, exclude=['vae'])
    if output is not None:
        if hasattr(output, 'bytes') and output.bytes is not None:
            log.debug(f'Generated: bytes={len(output.bytes)}')
            return output
        audio = getattr(output, 'audio', None)
        if not hasattr(output, 'images') and hasattr(output, 'frames'):
            log.debug(f'Generated: frames={len(output.frames[0])}')
            output.images = output.frames[0]
        if output.images is not None and len(output.images) > 0 and isinstance(output.images[0], Image.Image):
            return attach_audio(output.images, audio)
        model = shared.sd_model if not is_refiner_enabled(p) else shared.sd_refiner
        if not hasattr(model, 'vae'):
            if hasattr(model, 'pipe') and hasattr(model.pipe, 'vae'):
                model = model.pipe
        if (hasattr(model, "vae") or hasattr(model, "vqgan")) and (output.images is not None) and (len(output.images) > 0):
            if p.hr_resize_mode > 0 and (p.hr_upscaler != 'None' or p.hr_resize_mode == 5):
                width = max(getattr(p, 'width', 0), getattr(p, 'hr_upscale_to_x', 0))
                height = max(getattr(p, 'height', 0), getattr(p, 'hr_upscale_to_y', 0))
            else:
                width = getattr(p, 'width', 0)
                height = getattr(p, 'height', 0)
            frames = p.task_args.get('num_frames', None) or getattr(p, 'frames', None)
            if isinstance(output.images, list):
                results = []
                for i in range(len(output.images)):
                    result_batch = processing_vae.vae_decode(
                        latents = output.images[i],
                        model = model,
                        vae_type = p.vae_type,
                        width = width,
                        height = height,
                        frames = frames,
                    )
                    for result in list(result_batch):
                        results.append(result)
            else:
                results = processing_vae.vae_decode(
                    latents = output.images,
                    model = model,
                    vae_type = p.vae_type,
                    width = width,
                    height = height,
                    frames = frames,
                )
                if not isinstance(results, list):
                    results = list(results)
        elif hasattr(output, 'images'):
            results = output.images
        else:
            log.warning('Processing: no results')
            results = []
    else:
        log.warning('Processing: no results')
        audio = None
        results = []
    return attach_audio(results, audio)