remove scale latent and no-crop options from hires fix

support copy-pasting new parameters for hires fix

modules/processing.py

@@ -506,14 +506,12 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
     firstphase_width_truncated = 0
     firstphase_height_truncated = 0
 
-    def __init__(self, enable_hr=False, scale_latent=True, denoising_strength=0.75, firstphase_width=512, firstphase_height=512, crop_scale=False, **kwargs):
+    def __init__(self, enable_hr=False, denoising_strength=0.75, firstphase_width=512, firstphase_height=512, **kwargs):
         super().__init__(**kwargs)
         self.enable_hr = enable_hr
-        self.scale_latent = scale_latent
         self.denoising_strength = denoising_strength
         self.firstphase_width = firstphase_width
         self.firstphase_height = firstphase_height
-        self.crop_scale = crop_scale
 
     def init(self, all_prompts, all_seeds, all_subseeds):
         if self.enable_hr:
@@ -530,6 +528,8 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
             samples = self.sampler.sample(self, x, conditioning, unconditional_conditioning)
             return samples
 
+        self.extra_generation_params["First pass size"] = f"{self.firstphase_width}x{self.firstphase_height}"
+
         x = create_random_tensors([opt_C, self.firstphase_height // opt_f, self.firstphase_width // opt_f], seeds=seeds, subseeds=subseeds, subseed_strength=self.subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self)
         samples = self.sampler.sample(self, x, conditioning, unconditional_conditioning)
 
@@ -538,46 +538,36 @@ class StableDiffusionProcessingTxt2Img(StableDiffusionProcessing):
         width_ratio = self.width/self.firstphase_width
         height_ratio = self.height/self.firstphase_height
 
-        if self.crop_scale:
+        if width_ratio > height_ratio:
-            if width_ratio > height_ratio:
+            truncate_y = int((self.width - self.firstphase_width) / width_ratio / height_ratio / opt_f)
-                #Crop to landscape
-                truncate_y = int((self.width - self.firstphase_width) / width_ratio / height_ratio / opt_f)
 
-            elif width_ratio < height_ratio:
+        elif width_ratio < height_ratio:
-                #Crop to portrait
+            truncate_x = int((self.height - self.firstphase_height) / width_ratio / height_ratio / opt_f)
-                truncate_x = int((self.height - self.firstphase_height) / width_ratio / height_ratio / opt_f)
                 
-            samples = samples[:, :, truncate_y//2:samples.shape[2]-truncate_y//2, truncate_x//2:samples.shape[3]-truncate_x//2]
+        samples = samples[:, :, truncate_y//2:samples.shape[2]-truncate_y//2, truncate_x//2:samples.shape[3]-truncate_x//2]
 
-        
+        decoded_samples = decode_first_stage(self.sd_model, samples)
 
-        
+        if opts.upscaler_for_img2img is None or opts.upscaler_for_img2img == "None":
-
+            decoded_samples = torch.nn.functional.interpolate(decoded_samples, size=(self.height, self.width), mode="bilinear")
-        if self.scale_latent:
-            samples = torch.nn.functional.interpolate(samples, size=(self.height // opt_f, self.width // opt_f), mode="bilinear")
         else:
-            decoded_samples = decode_first_stage(self.sd_model, samples)
+            lowres_samples = torch.clamp((decoded_samples + 1.0) / 2.0, min=0.0, max=1.0)
 
-            if opts.upscaler_for_img2img is None or opts.upscaler_for_img2img == "None":
+            batch_images = []
-                decoded_samples = torch.nn.functional.interpolate(decoded_samples, size=(self.height, self.width), mode="bilinear")
+            for i, x_sample in enumerate(lowres_samples):
-            else:
+                x_sample = 255. * np.moveaxis(x_sample.cpu().numpy(), 0, 2)
-                lowres_samples = torch.clamp((decoded_samples + 1.0) / 2.0, min=0.0, max=1.0)
+                x_sample = x_sample.astype(np.uint8)
+                image = Image.fromarray(x_sample)
+                image = images.resize_image(0, image, self.width, self.height)
+                image = np.array(image).astype(np.float32) / 255.0
+                image = np.moveaxis(image, 2, 0)
+                batch_images.append(image)
 
-                batch_images = []
+            decoded_samples = torch.from_numpy(np.array(batch_images))
-                for i, x_sample in enumerate(lowres_samples):
+            decoded_samples = decoded_samples.to(shared.device)
-                    x_sample = 255. * np.moveaxis(x_sample.cpu().numpy(), 0, 2)
+            decoded_samples = 2. * decoded_samples - 1.
-                    x_sample = x_sample.astype(np.uint8)
-                    image = Image.fromarray(x_sample)
-                    image = images.resize_image(0, image, self.width, self.height)
-                    image = np.array(image).astype(np.float32) / 255.0
-                    image = np.moveaxis(image, 2, 0)
-                    batch_images.append(image)
 
-                decoded_samples = torch.from_numpy(np.array(batch_images))
+        samples = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(decoded_samples))
-                decoded_samples = decoded_samples.to(shared.device)
-                decoded_samples = 2. * decoded_samples - 1.
-
-            samples = self.sd_model.get_first_stage_encoding(self.sd_model.encode_first_stage(decoded_samples))
 
         shared.state.nextjob()
 

modules/txt2img.py

@@ -6,7 +6,7 @@ import modules.processing as processing
 from modules.ui import plaintext_to_html
 
 
-def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, scale_latent: bool, denoising_strength: float, firstphase_width: int, firstphase_height: int, crop_scale: bool, *args):
+def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2: str, steps: int, sampler_index: int, restore_faces: bool, tiling: bool, n_iter: int, batch_size: int, cfg_scale: float, seed: int, subseed: int, subseed_strength: float, seed_resize_from_h: int, seed_resize_from_w: int, seed_enable_extras: bool, height: int, width: int, enable_hr: bool, denoising_strength: float, firstphase_width: int, firstphase_height: int, *args):
     p = StableDiffusionProcessingTxt2Img(
         sd_model=shared.sd_model,
         outpath_samples=opts.outdir_samples or opts.outdir_txt2img_samples,
@@ -30,12 +30,9 @@ def txt2img(prompt: str, negative_prompt: str, prompt_style: str, prompt_style2:
         restore_faces=restore_faces,
         tiling=tiling,
         enable_hr=enable_hr,
-        scale_latent=scale_latent if enable_hr else None,
         denoising_strength=denoising_strength if enable_hr else None,
-		firstphase_width=firstphase_width if enable_hr else None,
+        firstphase_width=firstphase_width if enable_hr else None,
-		firstphase_height=firstphase_height if enable_hr else None,
+        firstphase_height=firstphase_height if enable_hr else None,
-        crop_scale=crop_scale if enable_hr else None,
-
     )
 
     if cmd_opts.enable_console_prompts:

modules/ui.py

@@ -540,16 +540,9 @@ def create_ui(wrap_gradio_gpu_call):
                     enable_hr = gr.Checkbox(label='Highres. fix', value=False)
 
                 with gr.Row(visible=False) as hr_options:
-                    with gr.Column(scale=1.0):
+                    firstphase_width = gr.Slider(minimum=64, maximum=1024, step=64, label="First pass width", value=512)
-                        firstphase_width = gr.Slider(minimum=64, maximum=1024, step=64, label="First pass width", value=512)
+                    firstphase_height = gr.Slider(minimum=64, maximum=1024, step=64, label="First pass height", value=512)
-                        firstphase_height = gr.Slider(minimum=64, maximum=1024, step=64, label="First pass height", value=512)
+                    denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7)
-                        
-                    with gr.Column(scale=1.0):
-                        with gr.Row():
-                            crop_scale = gr.Checkbox(label='Crop when scaling', value=False)
-                            scale_latent = gr.Checkbox(label='Scale latent', value=False)
-                        with gr.Row():
-                            denoising_strength = gr.Slider(minimum=0.0, maximum=1.0, step=0.01, label='Denoising strength', value=0.7)
 
                 with gr.Row(equal_height=True):
                     batch_count = gr.Slider(minimum=1, step=1, label='Batch count', value=1)
@@ -610,11 +603,9 @@ def create_ui(wrap_gradio_gpu_call):
                     height,
                     width,
                     enable_hr,
-                    scale_latent,
                     denoising_strength,
                     firstphase_width,
                     firstphase_height,
-                    crop_scale,
                 ] + custom_inputs,
                 outputs=[
                     txt2img_gallery,
@@ -679,8 +670,8 @@ def create_ui(wrap_gradio_gpu_call):
                 (denoising_strength, "Denoising strength"),
                 (enable_hr, lambda d: "Denoising strength" in d),
                 (hr_options, lambda d: gr.Row.update(visible="Denoising strength" in d)),
-                (firstphase_width, "First pass width"),
+                (firstphase_width, "First pass size-1"),
-                (firstphase_height, "First pass height"),
+                (firstphase_height, "First pass size-2"),
             ]
             modules.generation_parameters_copypaste.connect_paste(paste, txt2img_paste_fields, txt2img_prompt)
             token_button.click(fn=update_token_counter, inputs=[txt2img_prompt, steps], outputs=[token_counter])